Cover page

M87’s Central Black Hole in Polarized Light (Image source : https://apod.nasa.gov/apod/ap210331.html )

M87 has a supermassive blackhole at its centre that accrete matter that spin around in relativistic velocities

before falling into the blackhole. The Event Horizon Telescope (EHT) is a virtual radio telescope that has a

size equivalent to the size of the earth. This allows astronomers to see the ﬁner details as well as the bulk

behaviour of the object that they are observing. To know more, see https://eventhorizontelescope.org/blog/

wobbling-shadow-m87-black-hole

Managing Editor Chief Editor Editorial Board Correspondence

Ninan Sajeeth Philip Abraham Mulamootil K Babu Joseph The Chief Editor

Ajit K Kembhavi airis4D

Geetha Paul Thelliyoor - 689544

Arun Kumar Aniyan India

Jorunal Publisher Details

Publisher : airis4D, Thelliyoor 689544, India

Website : www.airis4d.com

Email : nsp@airis4d.com

Phone : +919497552476

Editorial

by Fr Dr Abraham Mulamoottil

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Airis4D’s latest issue features a diverse set of topics that cater to readers interested in machine learning,

astronomy, biosciences, and climate. The articles provide valuable insights into the parameters and hyperpa-

rameters of machine learning algorithms, language models, black holes, radio waves, stem cells, and cloud

formation.

The article on machine learning emphasizes the importance of selecting the optimal algorithm, using the

right amount and quality of data, and evaluating the model’s performance using various measures. Parameters

and hyperparameters are two critical variables in machine learning models that can signiﬁcantly aﬀect their

performance and generalization ability. The text also provides examples of hyperparameters in diﬀerent machine

learning models and techniques for hyperparameter tuning.

The article on language models delves into the two types of language models and their statistical techniques

to predict the likelihood of the next word or phrase based on context and previous words. The text highlights

the importance of choosing the appropriate value of ’n’ to capture more context without increasing sparsity and

techniques like smoothing and backing oﬀ to address the sparsity issue.

The articles on astronomy and biosciences provide valuable insights into black holes, radio waves, and

stem cells’ properties and applications in healthcare and regenerative medicine. The article on cloud formation

discusses the diﬀerent types of clouds and their impact on the Earth’s climate.

The issue’s ﬁnal section oﬀers a ﬁction piece called Cranky the Scientist, providing a unique twist to the

conventional format of academic journals.

Overall, Airis4D’s latest issue oﬀers informative and engaging articles that cater to a diverse set of readers

interested in various ﬁelds. The articles provide valuable insights and knowledge while keeping the content

accessible and engaging. In conclusion, Airis4D Journal provides valuable insights into various scientiﬁc ﬁelds,

making it a great source of information for researchers, students, and enthusiasts alike. The articles provide a

good starting point for exploring diﬀerent topics and should be further revised and edited to improve readability

and coherence.

An End Note: ChatGPT has been widely accepted today as one of the most advanced and capable language

models for natural language processing. It has the ability to understand human language, generate responses, and

learn from its interactions with users. Its advanced AI capabilities have made it a valuable tool for a wide range

of applications, including chatbots, virtual assistants, language translation, and more. Its compatibility with

various programming languages and frameworks, along with its open-source nature, has made it accessible to

developers and businesses of all sizes, leading to its widespread adoption. ChatGPT’s acceptance and adoption

are expected to continue to grow as AI technology advances and more use cases for language processing emerge.

Contents

Editorial ii

I Artiﬁcial Intelligence and Machine Learning 1

1 Nuts and Bolts of Machine Learning - Part 2 2

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 Parameters and Hyperparameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Language Models: Enabling Machines to Understand and Generate Human Language 7

2.1 What is a language model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.2 Statistical Language Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.3 Neural Language Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

II Astronomy and Astrophysics 12

1 The Mass of Black Holes 13

1.1 A Black Hole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.2 The Properties of a Black Hole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.3 The Black Hole Zoo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1.4 Supermassive Black Holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

1.5 The Supermassive Object at the Centre of our Galaxy . . . . . . . . . . . . . . . . . . . . . . 19

1.6 The Orbit of Star S2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

1.7 Historical Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

1.8 Is the Compact Object a Supermassive Black Hole? . . . . . . . . . . . . . . . . . . . . . . . 25

2 The Little Green Men in Space 26

2.1 What are radio waves? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.2 Why did radio waves have to be discovered? . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.3 What is electromagnetic radiation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.4 What are the mechanisms that create EM waves? . . . . . . . . . . . . . . . . . . . . . . . . 27

2.5 A radio universe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.6 The story of the Little Green Men . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.7 Why do we need radio astronomy? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.8 The challenges in radio astronomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3 Pulsating Variable Stars 32

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.2 Why do they pulsate? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3.3 Diﬀerent types of pulsating variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

CONTENTS

III Biosciences 40

1 The Avatars of Stem Cells and its applications in healthcare 41

1.1 What is the procedure behind stem cell therapy? . . . . . . . . . . . . . . . . . . . . . . . . . 44

1.2 Applications of stem cells in modern research . . . . . . . . . . . . . . . . . . . . . . . . . . 45

1.3 Promises of Stem Cells In Regenerative Medicine . . . . . . . . . . . . . . . . . . . . . . . . 47

IV General 49

1 Beyond the Fluﬀy: The Enchanting World of Clouds 50

1.1 Cloud Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

1.2 Types of Clouds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

1.3 What Causes Rain? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

1.4 How Clouds Aﬀect the Earth’s Climate? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

V Fiction 54

1 Cranky the Scientist 55

Part I

Artiﬁcial Intelligence and Machine Learning

Nuts and Bolts of Machine Learning - Part 2

by Blesson George

airis4D, Vol.1, No.4, 2023

www.airis4d.com

1.1 Introduction

Machine learning is deﬁned as the usage and development of computer systems that can learn and adapt

without being explicitly programmed. Although while ML learns without explicit instructions, it is not true

that the system will generate a model without human intervention. There are misunderstandings that, given

the required data, a system would automatically develop a model. In this article, we demonstrate the role of

users/humans in model construction.

The initial task of users is to determine the optimal algorithm for the programme. Consider the classiﬁcation

job as an illustration. There are a variety of methods, such as SVM, logistic regression, Decision tree, Random

forest, and Naive Bayes, for performing the task. The choice of algorithm for a speciﬁc problem is left to the

discretion of the user. User must make a decision based on a variety of considerations. For instance, if the user

has the option to choose between speed and accuracy, he has several options. Kernel SVM, Random Forest,

and Neurel network are methods that provide greater accuracy than other methods, however the user will select

Linear SVM, logistic regression, or the Naive Bayes approach based on speed. Again, he must evaluate whether

or not he need explanability. Logistic regression and decision tree are approaches with superior explainability

compared to others.

People with domain experience can give useful insights on the sorts of data and patterns that are likely to

be there. This information can assist in narrowing down the list of candidate algorithms to those that are most

suited to solving the problem. Furthermore, Users may evaluate the eﬀectiveness of various machine learning

(ML) algorithms using measures such as accuracy, precision, recall, and F1 score. This evaluation can assist in

determining which algorithm is the most eﬀective solution for the given problem.

User participation to ML model creation is not limited to algorithm ﬁnding. The user is responsible for

ensuring enough amount and quality of data for training. Users may test and assess machine learning models

to oﬀer feedback on their performance. This input may be utilised to enhance the model and increase its utility

for the intended application. Users can discover possible use cases for machine learning (ML) models and oﬀer

comments on how the models might be enhanced to better meet their needs. Users may also assist in identifying

the most essential traits and parameters to include in machine learning models. This can guarantee that the

models are centred on the most pertinent and valuable data.

1.2 Parameters and Hyperparameters

Assume that we have identiﬁed the optimal method for our problem. The work is not yet complete. The

user must provide some parameters of the algorithm. Before proceeding, let’s deﬁne the two sorts of variables

in the ML algorithm: parameters and hyperparameters.

1.2.1 Parameters

The internal variables of a model that are learned from training data are referred to as parameters in machine

learning. These factors are applied to fresh data to generate forecasts. The objective of machine learning is to

discover the model parameters that reliably predict the output given the input.

For instance, the parameters of a linear regression model are the slope and intercept of the line that best

ﬁts the data. The parameters of a neural network are the weights and biases of the network’s nodes. During

training, the algorithm modiﬁes the values of these parameters to reduce the disparity between the expected and

actual output.

Once the model has been trained, the learnt parameters are applied to new data to generate predictions.

Notably, the quality of the predictions is dependent on the quality and amount of the training data, as well as

the model’s complexity and its ability to generalise to new data.

1.2.2 Hyperparameters

In machine learning, hyperparameters are parameters that cannot be directly learnt from training data and

are instead deﬁned before to the training process. These options govern the learning algorithm’s behaviour

and dictate how the model is taught. As they inﬂuence the model’s bias and variance, hyperparameters play a

signiﬁcant role in deﬁning a model’s performance.

The term ”hyper” is included in the name because these parameters control the training process and

determine the model’s parameters. The values of the hyperparameters are ﬁxed prior to the start of training and

are not altered throughout the training process.

1.2.3 A few illustraions of ML parameters

1.2.3.1 Linear regression

Imagine a linear regression model that predicts an output based on an input value (s). The model of

machine learning is represented by a line equation, given as

x) = θ

+ θ

x (1.1)

The θ

and θ

represents the model’s parameters. The linear regression model does not have explicit

hyperparameters. However linear models like Ridge, Lasso etc have the loss function, dropuout rate, learning

rate etc as the hyperparameters. For nonlinear regressions like SGD regression, the degree of polynomial

features can be a user-speciﬁed hyperparameter value.

1.2.3.2 Support Vector Machine (SVM)

A Support Vector Machine is a supervised machine learning approach that is applicable to both classiﬁcation

and regression tasks. It uses a technique known as the kernel trick to alter the data, and based on these

transformations, it identiﬁes the optimal boundary between the potential outputs. SVM operates by identifying

1.2 Parameters and Hyperparameters

Figure 1: Figure showing the linear regression plot between population in a city vs the proﬁt. The values of θ

and θ

are -3.62 and 1.17 in this case. Image courtesy: https://gtraskas.github.io/post/ex1/

the ideal separating hyperplane that optimises the training data margin. SVM parameters include support

vector indexes, intercept, and the number of support vectors for each class, among others. User-supplied

hyperparameters include the kind of kernel used to generate the hyperplane, regularisation parameter, degree of

the polynomial kernel function, kernel coeﬃcient, etc.

Figure 2: Figure showing the eﬀect of Regularization parameter on the eﬀect of deciding the support vectors.

Image courtesy: https://www.codingninjas.com/codestudio/library/svm-hyperparameter-tuning-using-gridsearchcv

1.2.3.3 Convolutional Neural Networks

Convolutional Neural Networks (CNNs) are a specialized type of neural network that is commonly used for

image classiﬁcation and recognition tasks. Like all neural networks, CNNs have parameters and hyperparameters

that are critical to their performance. Optimizing Convolutional Neural Network (CNN) hyperparameters is a

diﬃcult task for many academics and practitioners. To obtain hyperparameters with improved performance,

specialists must manually conﬁgure a set of hyperparameter options.

The parameters optimized during the training process includes Filters or kernels (which are the feature

extractors in CNNs), Weights (parameters that determine the strength of the connections between neurons in

1.3 Conclusion

the network and Biases (parameters that are added to the output of each neuron)

Hyperparameters are the parameters determined by the user. There are two types of hyperparameters in

case of CNN. They are

Hyperparameter that determines the network structure: This includes the size of the ﬁlter, Kernel

Type(e.g., edge detection, sharpen), the rate at which the kernel pass over the input,padding, layers

between input and output layers and the type of Activation functions.

Hyperparameter that determines the network trained such as: Learning rate, Momentum, number of

epochs, Batch size etc

The parameters and hyperparameters of a few of the networks have been presented in detail. The user

must conﬁgure certain network settings for each network. Setting up hyperparameters requires expert domain

expertise. Hyperparameter tuning is a method for determining the appropriate value of hyperaparameters.

1.2.4 Hyperparameter Tuning

Hyperparameter tuning is the process of selecting the optimal set of hyperparameters for a machine learning

algorithm. Hyperparameters are parameters that are not learned by the algorithm itself, but set by the user

before the training process begins. Examples of hyperparameters include the learning rate, batch size, number

of hidden layers, and number of neurons in each hidden layer of a neural network.

The performance of a machine learning model is highly dependent on the choice of hyperparameters.

Therefore, hyperparameter tuning is an important step in the machine learning workﬂow. The goal of hyperpa-

rameter tuning is to ﬁnd the set of hyperparameters that results in the best performance of the model on a given

task.

There are several techniques that can be used for hyperparameter tuning, including grid search, random

search, and Bayesian optimization. Grid search involves trying every combination of hyperparameters in a

pre-deﬁned search space. Random search involves sampling random combinations of hyperparameters from

the search space. Bayesian optimization involves modeling the performance of the model as a function of the

hyperparameters and using this model to guide the search for the optimal set of hyperparameters.

Hyperparameter tuning can be time-consuming and computationally expensive, but it is essential for

achieving the best performance of a machine learning model.

1.3 Conclusion

Understanding parameters and hyperparameters is vital for the development of eﬀective machine learning

models. Parameters are internal variables learned by the model during training, whereas hyperparameters are

settings that aﬀect the learning algorithm’s behaviour. Each of these variables have a substantial eﬀect on how

well the model performs, and making the right choice between them may result in a model that is more accurate,

eﬃcient, and easily interpretable, as well as one that generalize well to new data.

References

1. Hyperparameter tuning for machine learning models

2. Simple Linear Regression — Parameter Estimates Explained.

3. InDepth: Parameter tuning for Decision Tree

4. Parameters and Hyperparameters in Machine Learning and Deep Learning

1.3 Conclusion

About the Author

Blesson George is currently working as Assistant Professor of Physics at CMS College Kottayam,

Kerala. His research interests include developing machine learning algorithms and application of machine

learning techniques in protein studies.

Language Models: Enabling Machines to

Understand and Generate Human Language

by Jinsu Ann Mathew

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Have you ever noticed the ‘text prediction feature’ on mobile phones which suggests the next word or

phrase that the user might type based on the context and the words they have already entered (Figure 1). This

is one of the various use-cases of language models used in Natural Language Processing (NLP). But what is

a language model? To answer that question, we ﬁrst need to clarify the concept of a model and its usage in

machine learning.

Models are used to represent complex and confusing real-world systems in a simpliﬁed way within a

speciﬁc ﬁeld of interest, also known as a domain. For example, consider ﬁnancial models. They are used to

represent complex ﬁnancial systems and forecast future ﬁnancial outcomes based on historical data and current

trends. These models may be used for a variety of purposes, such as predicting stock prices, estimating the

future value of an investment, or analyzing the ﬁnancial impact of a business decision.

Machine learning models are much the same. It can be thought of as programs that are trained to detect

patterns in new data and make predictions. Essentially, they are algorithms that can recognize patterns or make

predictions on previously unseen datasets. Unlike rule-based programs, these models do not require explicit

coding and can evolve over time as new data is incorporated into the system. Initially, these models are trained

https://www.samsung.com/ie/support/mobile-devices/how-can-i-personalise-and-turn-predictive-text-on-and-oﬀ-on-my-samsung-galaxy-device/

Figure 1: Example suggestions provided by ‘text prediction feature’ for the typed query preﬁx ”Th”.

2.1 What is a language model

on a speciﬁc set of data and provided with an algorithm to analyze and extract patterns from the input data.

Once the models have been trained, they can then be used to predict outcomes on new, unseen datasets.

2.1 What is a language model

A language model is a probability distribution over words or word sequences. It’s a a type of statistical

model that is designed to analyze and understand the patterns of human language. It helps to predict the

likelihood of the next word or sequence of words in a given sentence, based on the context and previous words.

Generally, a set of example sentences serves as input to the language model, which generates a probability

distribution over sequences of words.

The modeling of language through probabilistic approaches can take on various forms depending on the

intended purpose of the language model. From a technical perspective various types diﬀer in terms of the quantity

of text data analyzed and the mathematical techniques employed for analysis. For example, A language model

designed to provide autocomplete suggestions for an email client may use diﬀerent mathematical techniques

and analyze text data diﬀerently than a language model designed for sentiment analysis of social media posts.

The former may focus on predicting the next word in a sentence based on the user’s writing history, while the

latter may be more concerned with identifying emotional tone and sentiment expressed There are primarily two

types of language models:

Statistical language model

Neural language model

2.2 Statistical Language Model

Statistical language modeling (SLM) aims to capture the underlying patterns and structure of natural

language in order to enhance the performance of various language-based applications. SLM relies on established

statistical techniques such as N-grams, Hidden Markov Models (HMM), and linguistic rules to learn the

probability distribution of words in textual data.

One of the most important features of statistical language models is their ability to predict the next word

in a sequence based on the words that came before it. This is done by using the preceding words as context

to compute the probability of each word appearing in the sequence. To accomplish this, statistical language

models often represent each word in the lexicon as a unique numerical value. These values are derived from the

training data and capture the underlying patterns and relationships between words. In this way, SLM enables

language-based applications to accurately predict and generate text based on the patterns and relationships found

in the training data.

Statistical Language Models are further classiﬁed into:

2.2.1 N-Gram model

A relatively simple type of language model is the N-gram model. The n-gram model calculates the

probability of a given sequence of words by estimating the probability of each word based on its previous n-1

words. For example, in a bigram model (n=2), the probability of a word depends only on the preceding word.

In a trigram model (n=3), the probability of a word depends on the two preceding words (Figure 2).

The choice of n in the n-gram model depends on the task and the size of the corpus. A larger value of n

captures more context and leads to a more accurate model, but it also increases the sparsity of the data, making

2.2 Statistical Language Model

(image courtesy:https://tejinderpal-singh17.medium.com/language-models-a-brief-overview-91631f8f1e13

Figure 2: Example of unigram, bigram and trigram

Figure 3: Bidirectional language model

it more diﬃcult to estimate probabilities. Conversely, a smaller value of n leads to a simpler model with fewer

parameters, but it may not capture enough context to be useful for some tasks.

2.2.2 Bidirectional model

A bidirectional language model is a type of language model that takes into account both the left and right

context of a word or sentence when predicting the next word in a sequence (Figure 3).

In a traditional language model, the probability of the next word depends only on the preceding words in

the sequence. In a bidirectional language model, however, the probability of the next word depends on both the

preceding words and the following words. This allows the model to better capture the context and meaning of

the text, as it can take into account both the preceding and following words to make a prediction.

One advantage of bidirectional language models is that they can capture long-range dependencies in the

text, as they consider both the preceding and following words. However, they are typically more computationally

expensive than traditional language models, as they require processing the text in both directions.

2.2.3 Exponential model

Exponential model is a type of statistical language model that uses the principle of maximum entropy to

estimate the probability distribution of words in a language. This is also called maximum entropy model.The

principle of maximum entropy, also known as the maximum entropy principle, is a statistical method used to

make predictions based on incomplete information.

The basic idea behind this principle is to choose the probability distribution that maximizes the entropy,

subject to constraints that are known or assumed. In the case of a language model, the constraints are the

observed frequencies of words in a training corpus, and the goal is to estimate the probability distribution of

words that best ﬁts these constraints. The maximum entropy language model achieves this by choosing the

2.3 Neural Language Model

probability distribution that maximizes the entropy subject to the observed frequencies of words.

One advantage of the maximum entropy language model is that it can handle a wide range of features and

constraints, such as n-gram frequencies, syntactic structures, and semantic information. This ﬂexibility makes

it useful in many natural language processing tasks, such as machine translation, speech recognition, and text

classiﬁcation.

2.3 Neural Language Model

A neural network language model (NNLM) is a type of language model that uses neural networks to learn

the probability distribution of words in a language. It is a type of neural language model that is trained on large

amounts of text data to predict the probability distribution of the next word in a sequence.

The basic idea behind the neural network language model is to use a neural network to map a sequence of

words to a probability distribution over the next word in the sequence. The neural network typically consists of

an input layer that receives the words of the sequence, one-hot encoded or as word embeddings, and one or more

hidden layers that perform nonlinear transformations on the input. The output layer produces the probability

distribution over the vocabulary of possible next words.

During training, the neural network is optimized to maximize the probability of the next word in the

sequence given the previous words. This is typically done using backpropagation and gradient descent to adjust

the weights of the neural network. The trained neural network can then be used to generate new text or to predict

the next word in a sequence.

One advantage of the neural network language model is that it can capture complex relationships between

words, such as syntactic and semantic dependencies, and can generalize to unseen sequences of words. It has

shown to outperform traditional n-gram language models in many natural language processing tasks, such as

language modeling, machine translation, and text generation.

Summary

In conclusion, a language model is a fundamental concept in natural language processing that enables

machines to understand and generate human language. It is a statistical model that estimates the probability

distribution of words in a language, and it can be used for a variety of tasks, such as speech recognition, machine

translation, and text generation. N-gram language models are the simplest and most widely used type of language

models, which estimate the probability of a word given its preceding n-1 words. However, they suﬀer from

the sparsity problem when dealing with large vocabularies, and they do not capture long-range dependencies

between words. To address these limitations, more advanced language models have been developed, such as the

maximum entropy language model and the neural network language model. The maximum entropy language

model uses the principle of maximum entropy to estimate the probability distribution of words, while the neural

network language model uses neural networks to learn the probability distribution of words. Neural network

language models have shown to outperform traditional n-gram language models in many natural language

processing tasks, due to their ability to capture complex relationships between words and generalize to unseen

sequences of words. Overall, language models play a critical role in natural language processing, and their

development and reﬁnement will continue to improve the ability of machines to understand and generate human

language.

2.3 Neural Language Model

References

Machine Learning Models, Javatpoint

What Is a Language Model?, TUANA C¸ ELIK, deepset, July 2022

Building Language Models in NLP, Koushiki Dasgupta Chaudhuri, AnalyticsVidhya, January 2022

A Beginner’s Guide to Language Models, M

or Kapronczay, builtin, Decenber 2022

Language Models- A Brief Overview, Tejinderpal Singh, medium, june 2021

Neural net language models, Yoshua Bengio, Scholarpedia

What are Language Models in NLP?, daﬀodil, july 2020

Understanding Language Models in NLP, Jammie sandy, medium, february 2023

Language Models in AI, Dennis Ash, medium, february 2021

About the Author

Jinsu Ann Mathew is a research scholar in Natural Language Processing and Chemical Informatics.

Her interests include applying basic scientiﬁc research on computational linguistics, practical applications of

human language technology, and interdisciplinary work in computational physics.

Part II

Astronomy and Astrophysics

The Mass of Black Holes

by Ajit Kembhavi

airis4D, Vol.1, No.4, 2023

www.airis4d.com

1.1 A Black Hole

The term Black Hole is now well established in the public mind. It is known as an object from which no

light escapes, making it all black, and therefore invisible. The real nature of a black hole is rather more bizarre

than the simple understanding of it. It is an object which has mass, which can be very large, and yet it has zero

size. Since the mass is packed into zero volume, the density of the object is inﬁnitely large, which is not usually

acceptable as a possible physical situation. The inﬁnite density is not just the one problem, there is an even

greater one.

In Newton’s theory of gravity there can in principle be objects from which light cannot escape. But these

objects inevitably involve very strong gravitational ﬁelds which can be correctly described not by Newton’s

theory, but by Albert Einstein’s general theory of relativity. In Einstein’s theory, the force of gravity is related

to the structure of space-time, which consists of the three dimensions of space and time combined into a single

entity. Unlike the space and time of our usual experience, this space-time is dynamic in nature, which means

that its properties are inﬂuenced by the presence of matter and energy, in a manner which is determined by

Einstein’s equations of gravitation. The greater the mass of an object is, the greater is the curvature produced

by it in space-time, which manifests itself as a stronger gravitational force. Because a black hole has non-zero

mass but zero size and therefore inﬁnitely large density, the curvature it produces is inﬁnitely large too. This

leads to a complete breakdown of the space-time structure as we understand it. At the black hole, space and

time cease to have the meaning we intuitively associate with them. And yet black holes are real objects whose

behaviour is predictable and observable using current telescopes and instruments.

1.2 The Properties of a Black Hole

A black hole is a very pristine object with just two properties, mass and spin. The spin is similar to the

rotation of the Earth round its own axis, and provides angular momentum to the black hole. Given the extreme

conditions, the spin produces interesting eﬀects not seen around normal rotating bodies. Theoretically, a black

hole can also have electric charge, but that is not expected in real black holes because the matter from which a

black hole is generally formed has zero total electric charge.

A black hole with zero spin is the simplest as it has just one property, mass. The mathematical form for

such a black hole was discovered by Karl Schwarzschild in 1916, just a year after Einstein announced his general

theory of relativity in 1915. But the name black hole was ﬁrst used in the 1960s, and was made famous by

1.3 The Black Hole Zoo

the great physicist John Wheeler. The Schwarzschild black hole has an imaginary spherical surface associated

with it known as the event horizon. This has a radius which is proportional to the mass and is known as the

Schwarzschild radius R

In the equation, M is the mass of the black hole, G is Newton’s constant of gravitation and cis the speed of

light. For a black hole of the mass of the Sun, the Schwarzschild radius is about 1.5 km, while for a supermassive

black hole with a mass 10

times the mass of the Sun, the radius would be about 150 million km. A black hole

is shown schematically in Figure 1

Figure 1: A black hole with mass, but no spin. The spherical event horizon, which is not a real surface, is

indicated.

No particles, light rays or signals can emerge to the outside world from within the event horizon. Any

matter or light ray or radiation of any kind entering the event horizon is lost forever. The name black hole is

thus fully justiﬁed. But then how does one detect such a black hole at all? The gravitational ﬁeld of the black

hole can be felt by the stars and any other matter surrounding it, which can lead to subtle or dramatic eﬀects

which reveal the presence of the black hole.

The mathematical form of a black hole with mass as well as spin was ﬁrst discovered by Roy Kerr in

1963. This too has an event horizon, which is more complicated in shape than the simple spherical shape for

the Schwarzschild black hole. There is also another surface associated with the Kerr black hole known as the

ergosphere, which is outside the event horizon. Any object entering the ergosphere is dragged into rotation by

the eﬀect of the spinning black hole.

1.3 The Black Hole Zoo

The black hole properties described above are valid for all black hole masses, the theory does not change

with the mass. In practice, some ranges of black hole mass are important because various astrophysical processes

lead to the formation of such black holes. For example, the evolution of stars with large mass can lead to the

1.3 The Black Hole Zoo

formation of a class of black holes, known as stellar mass black holes, with mass in the range of a few times the

Solar mass to about one hundred times the Solar mass. A Solar mass is the mass of our Sun, which is 2x10

Kg.

It is convenient to use this mass as the unit for astronomical objects like stars and galaxies. Black holes which

power active galaxies are much more massive, with mass in the range of a few million Solar masses to more than

a billion Solar masses. These are known as supermassive black holes. Another class is intermediate mass black

holes with mas in the range of about one hundred Solar masses to about a hundred thousand Solar masses. These

can be associated with low power active galaxies, globular clusters, ultra-luminous X-ray sources etc. A quite

diﬀerent class consists of primordial black holes which were ﬁrst hypothesized by Stephen Hawking. These

have mass as small as 10

-5

Solar mass and can lose their mass through quantum processes. In the following, we

will describe stellar mass and supermassive black holes, but not the other kinds.

Stellar Mass Black Holes: Black holes in this group have a mass in the range of a few times the Solar

mass to about 85 times the Solar mass. How such black holes are formed is well understood, especially for

black holes with mass less than about 20 Solar masses. Stars generate energy by nuclear fusion in their central

regions. When a star with a mass of greater than about 25 Solar masses exhausts all its nuclear fuel, the inner

region of the star implodes, while the outer parts are thrown out in a gigantic explosion known as a supernova.

The imploding inner part collapses to form a black hole, whose mass depends on how much was the mass of the

initial star (the implosion of stars with mass less than 25 Solar masses leads to the formation of a white dwarf

or a neutron star, depending on the initial mass). Such black holes are expected to be common in our Galaxy

and are known as stellar mass black holes.

Figure 2: Schematic of a black hole X-ray binary. The oval object on the right is a star which has expanded due

to evolution, with its original spherical shape distorted by gravitational and centrifugal forces. Matter ﬂows from

the star to the black hole and an accretion disc is formed around the black hole due to the angular momentum

of the inﬂowing matter. The disc and a surrounding cloud of energetic electrons are responsible for the X-ray

emission. Twin jets emanating from a region close to the black hole are seen.

We cannot easily detect solitary stellar mass black holes, but we can observe their eﬀects when they are in

a binary system. Such a system consists of two astronomical objects revolving around each other, each attracted

by the other’s gravitational ﬁeld. It is possible in such a system for one object to be a star, while the other is a

black hole. When the right conditions are met, matter ﬂows from the star onto the black hole, forms an accretion

disc around it and gradually enters the event horizon. The disc is at high temperature and releases X-rays and

other forms of high energy electromagnetic radiation. The radiation can be boosted to even higher energy

1.3 The Black Hole Zoo

through interaction with high energy electrons which are present around the accretion disc. All these processes

occur outside the event horizon, so much of the radiation does not fall into the black hole and escapes to great

distances. The radiation can be detected using suitable telescopes and instruments carried by satellites in orbit

around the Earth. About 20 such X-ray emitting binary systems are known, where there is high conﬁdence that

one component of the binary is a black hole. These black holes have a mass which ranges from about ﬁve Solar

masses to about 20 Solar masses.

There are X-ray binary systems in which the companion to the star is a neutron star rather than a black hole.

There are many similarities between neutron star and black hole binaries so it not always easy to discriminate

between the two. There are some diﬀerences too, for example a neutron star binary can show pulsations in its

X-ray emission, while a black hole binary cannot have such pulsations. The most important discriminant is of

course the mass. Neutron stars cannot have a mass greater than about three Solar masses. So, if the compact

object in X-ray binary has a mass much signiﬁcantly greater than three Solar masses, then it has to be a black

hole. A diﬃculty in using the argument is that often it is diﬃcult to accurately estimate the mass. The main

cause of the diﬃculty is that the mass measurement depends on the inclination of the plane of the binary orbit

to the line of sight, which is generally not known. Given the measurements, a range of mass is possible for the

compact object in a binary, and it is not always possible to say with certainty whether the object is a neutron star

or a black hole. But when the mass is much larger than the neutron star mass limit, it is extremely likely that the

object is indeed a black hole. So, it is safe to say that black holes have been detected in X-ray binary systems.

Any remaining doubt about the existence of black holes was set aside by the detection of gravitational

waves in 2015. An important prediction made by Albert Einstein in 1917 from his new theory is the existence of

gravitational waves. These waves are emitted when large masses undergo large accelerations, i.e. their already

high velocities change very rapidly. Such a circumstance is obtained when two black holes form a binary system

and go around each other very rapidly; the binary then becomes a strong emitter of gravitational waves. The

emitted waves carry away energy from the system, so that it shrinks in size, the black holes move closer and

go around each other even more rapidly, leading to even greater emission of gravitational waves. The result is

that the black holes rapidly spiral towards each other and ﬁnally merge together forming a single black hole,

which settles down to its pristine form through a short phase known as a ringdown. This was ﬁrst predicted by

Professor C. V. Vishveshwara in 1970 (the ringdown phase is named after the ringing down of a bell, which

happens when a bell is struck and allowed to vibrate freely).

Such a spiral-in and merger of two black holes was observed by the two LIGO gravitational wave detectors

in the USA in 1915, after a heroic eﬀort spanning more than two decades. From the pattern of the waveform

observed it can be unambiguously concluded that the two objects which spiraled in, as well as the object formed

from the merger are black holes. It is also possible to accurately measure the masses of the black holes, which

are 36 and 29 Solar masses respectively for the two black holes before the merger and 62 Solar masses for the

resultant black hole after the merger. The diﬀerence of three Solar masses between the merged mass and the

pre-merger masses was radiated away as energy in the form gravitational waves. The observation led to the

award of the Nobel Prize for 2017 to Professors Barry Barish, Kip Thorne and Rainer Weiss. The existence of

stellar mass black holes is now ﬁrmly established due to the many merging black holes which have been found

with gravitational wave detectors. The range of stellar mass black holes currently detected is shown in Figure

2. The large black hole masses detected through merging events cannot easily be explained by the usual stellar

evolution theory and establishing how they formed is a very active area of research.

1.4 Supermassive Black Holes

Figure 3: Black hole masses determined from X-ray binary related observations and from mergers detected

by the LIGO and Virgo gravitational wave detectors. The blue circles indicate the masses of black holes in

binaries before the merger and after the merger. The purple circles indicate black hole masses estimated from

electromagnetic (EM) observations at X-ray wavelengths. The yellow circles are masses of neutron stars in

X-ray binaries, again measured through EM means, and the orange circles indicate masses of neutron stars

which were either one both components in a merging binary which was gravitationally detected. Finally, the

white circles are the black hole masses which arose out of mergers involving at least one neutron star. Image

courtesy LIGO-Virgo/ Frank Elavsky & Aaron Geller (Northwestern)

1.4 Supermassive Black Holes

The other main kind of black holes are those known as supermassive black holes which are believed to be

located at the centres of galaxies. A galaxy consists of a collection of a large number of stars, gas and dust.

The most familiar kind are the spiral galaxies, whose images showing great spiral arms are familiar to everyone.

But a signiﬁcant number of galaxies appear to have elliptical shapes (these are projects on the sky of ellipsoidal

objects) while others have irregular shapes. A normal galaxy, like our own Milky Way which is a spiral galaxy,

has about a hundred billion to a thousand billion stars, and is about a hundred thousand light years in diametre

(a light year is the distance travelled by light in one year, travelling at a speed of 300,000 km per second, which

is about a nine trillion km). The light coming to us from a galaxy is the sum total of the emission from the

stars and gas in it, with the colour of the light altered when dust is present in the galaxy. It is believed that

galaxies were ﬁrst formed sometime after the Big Bang in which the Universe was created, and have undergone

signiﬁcant evolution in the billions of years after they were formed.

About 10 percent of all galaxies are active galaxies: they emit radio, X-ray and other radiations which

cannot be ascribed to the stars, and which must have a diﬀerent source. The energy emitted in this form by some

of the brighter active galaxies can be as high 10

erg/sec or more, which is far greater than the energy emitted

by all the stars in a galaxy like the Milky Way, which is about 10

erg/sec. Such radiation often varies rapidly

with time, which tells astronomers that much of it must be coming from a very compact source at the centre of

the galaxy. Some of the radio radiation can come from magniﬁcent large structures which are much bigger than

1.4 Supermassive Black Holes

the galaxy, and are fed energy by narrow jets emanating from the centre. Most intriguingly, small parts of the

radio structure at the centre of the galaxy are observed to be moving away each other with speeds exceeding the

speed of light. This seems to be contradicting Einstein’s special theory of relativity, according to which nothing

can move faster than light. This was a great puzzle in the 1960s when the high speeds were ﬁrst measured, until

a young research student in Cambridge, now Lord Martin Rees, explained that the observed high speeds were

an illusion due to radio waves being emitted by objects moving close to the line of sight, at speeds approaching,

but not exceeding, the speed of light.

Figure 4: The radio galaxy 3C348, also known as Hercules A. At the centre of the image an elliptical galaxy

is seen. The image at optical wavelengths was taken with the Hubble Space Telescope. The coloured extended

structure, much larger than the host elliptical galaxy is from observation made with the Very large Array at

radio wavelengths. The colour has been introduced for illustrative purposes only, it traces the emission at radio

wavelengths. Two narrow radio jets emerge from the galaxy to power the large radio lobes. The source of the

energy is a believed to be a supermassive black hole at the centre of the galaxy. Image Credit: NASA, ESA,

STScI.

All the above observations point to the existence, at the centre of active galaxies, a very massive and

compact object, with mass in the range of millions to billions of Solar masses. There are reasons to believe

that these objects are black holes, which are known as supermassive black holes. What is interesting is that

supermassive objects, which could very well be supermassive black holes, are also being found in the centre of

normal galaxies. These objects are discovered using a variety of eﬀects that their gravitational ﬁeld produces

on stars and gas in the vicinity of the object. It seems possible that supermassive black holes exist at the centre

of all galaxies. In a small fraction, these black holes receive a regular supply of matter in the form of stars and

gas from their vicinity, resulting in the formation of an accretion disc around the black holes from which matter

gradually falls into the black hole, releasing energy which makes the galaxy active.

We saw above that stellar mass black holes are formed in the evolution of large mass stars. How are

supermassive black holes formed? That is still an open question, but the understanding is that massive black

holes were formed in the early epochs of the Universe,

or in the early stages in the formation of galaxies, which merged to form even greater mass black holes.

The mass can also increase when there is a regular supply of stars and gas to be swallowed by the black hole.

Black holes with mass intermediate between the stellar mass black holes and supermassive black holes are also

1.5 The Supermassive Object at the Centre of our Galaxy

known.

The mass of supermassive black holes in distant galaxies can be estimated using a variety of techniques.

One technique is to use the motion of stars or gas around the centre of a galaxy, from which the mass can be

inferred using gravitation theory and dynamics. This has been done in the most direct way for our own Milky

Way galaxy, which is described below. There are more indirect techniques like reverberation mapping and

measuring the shapes of emission lines in the X-ray part of the spectrum. These methods provide estimates of

the supermassive black which can be used in further considerations. We will now describe how the mass of the

supermassive black hole at the centre of our Galaxy, the Milky Way or Akashganga, is determined from detailed

observation of stars in the central region of the Galaxy.

The work leading to the identiﬁcation of a supermassive object was done independently by Professor

Reinhard Genzel and his group in Germany, and Professor Andrea Ghez and her group in the USA. Genzel and

Ghez were awarded half the Nobel Prize for Physics in 2020 for their eﬀorts.The other half of the prize was

awarded to Professor Roger Penrose, for his theoretical work on black holes.

1.5 The Supermassive Object at the Centre of our Galaxy

The Milky Way is a spiral galaxy with more than a hundred billion (10

) stars. A large fraction of these

stars is distributed in a relatively thin disc, with the stars rotating around the galactic centre while remaining in

the plane of the disc. The disc has prominent spiral arms which give spiral galaxies their name. The diametre

of the disc is about a hundred thousand light years.

The second prominent structure of the galaxy is known as the bulge. It is nearly spherical in shape with a

diametre smaller than the diametre of the disc. Each star in the bulge moves around the centre in a plane, but

the orbits are not all coplanar. The orbital planes are distributed around the centre in such way that the bulge

stars appear to be distributed nearly spherically around the centre. In addition to the bulge and the disc, there is

a much larger, nearly spherical structure known as the Galactic Halo. This consists of old stars and globular

clusters. In addition the halo contains invisible dark matter, which in fact constitutes about 95 percent of the

total mass of the galaxy. The nature of the dark matter is presently not known.

Figure 5: This is a panoramic view of our galaxy in near-infrared light obtained by mosaicking images from

the 2MASS survey. The disc and the bulge are seen, though not the spiral arms, because we have only an

edge-on view of our own galaxy. The halo is not bright enough to be seen in the image (Image Courtesy

IPAC/CALTECH).

1.5 The Supermassive Object at the Centre of our Galaxy

A panoramic view of our Galaxy is shown in Figure 5. We cannot see spiral arms in Figure 5 because the

image shown is a sideway representation of our Galaxy. The spiral arms are shown in in the sketch in Figure

6. The shape of the arms is determined through optical, near-infrared and radio observations. Some prominent

arms have been named.

Figure 6: The arms of the Milky Way Galaxy are shown in this diagram. The structure of the arms has been

determined using optical, radio and near-infrared observations. The thick linear structure at the centre to which

the spiral arms are attached is known as a bar. Image Courtesy: NASA/JPL-Caltech/R. Hurt (SSC/Caltech)

For the purposes of determining the mass of a possible compact object at the centre, we will be concerned

only with stars very close to the centre. If the motion of such stars could be observed, it would be possible to

determine the shape of the orbit of the stars. Then using Newton’s law of gravitation and mechanics, the mass

of the object can be determined.

The bulge of our Galaxy, which includes the central region, is located in the constellation of Sagittarius

and can be seen as a large, faint nebulosity with dark patches in it. The central region is a complex environment

with a high density of stars in a cluster around the centre, stars which are being formed, exploding stars, gas

and dust. A source emitting radio waves has for long been known in the central region. In 1974 Bruce Balick

and Robert Brown discovered a very compact component of the radio source, which was later named Sgr A*.

Observations have shown that its size is smaller than the distance between the Sun and the Earth, which is about

150 million km, a size which is very tiny compared to the scale of the galaxy. Sgr A* shows no motion, which

suggests it is very massive, and is located at the centre of the galaxy.

As mentioned above, the mass of the central object can be measured by determining the orbits of the stars

in the galaxy. Observing individual stars in this region is very diﬃcult for two reasons: First, the centre is

about 25,000 light years away from the Sun, which makes the stars very faint. This dimming is made worse by

the presence of dust in the centre, which obscures the stars, so they appear to be much fainter than would be

warranted by their distance from us. Second, stars in the centre are crowded together, so they seemingly blend

1.6 The Orbit of Star S2

with each other, and are diﬃcult to distinguish as individual objects. To get over these problems, it is necessary

to use the largest telescopes and to observe in the near-infrared region of the spectrum, where the obscurations

due to the dust is low. It is also necessary to use specialised techniques like adaptive optics and speckle

interferometry to increase the resolution of the observations, so that individual stars can be distinguished. All

this high technology makes it possible for motion of individual stars close to the centre of the galaxy to be

followed over a period of time, establishing the shape of their orbit. Stars in the central region of the Galaxy are

shown in Figure 7.

Figure 7: The image on the left shows stars in the central region of our Galaxy. It was made with the Keck

10 metre telescope, and is a composite of three images in diﬀerent bands of the near-infrared region of the

spectrum. Speckle interferometry and adaptive optics were used to improve image resolution so that individual

stars in the crowded ﬁeld could be seen separately. The central region is marked by a rectangle. The bar in the

upper left-hand corner of the image shows that the size of the box is a few light months. This is a tiny fraction

of the 100,000 light years diametre of our Galaxy. The image on the right shows an expanded version of the

box. The position of Sgr A* is indicated by a star. Image courtesy W. M. Keck Observatory.

1.6 The Orbit of Star S2

Determining the path of a star requires its position to be known at diﬀerent times over a long period.

Imagine determining the path of a planet like Mars or Jupiter in the sky. The position of the planet at given times

can be determined in comparison with the position of neighbouring stars in the sky. The stars can themselves

to be taken to be stationery, since they are much further away from us than the planets, and their motion is

negligible compared to the motion of the planets. In practice, over centuries various coordinate systems have

been developed relative to which the motion of planets, and even the tiny motions of stars can be determined

with extreme accuracy.

Such measurements are much more diﬃcult in the galactic centre, as the stars there are in incessant motion

and it is extremely hard to determine the position of any one star over a period which stretches to years. Several

reference stars used for the purpose are known to be radio emitters. Their position at a given time can be

determined very accurately using an array of radio telescopes spread over the Earth. With this method, the

position of a star can be determined to an accuracy of about ten millionth of a degree.

1.6 The Orbit of Star S2

Reinhardt Genzel’s group used the Very Large Telescopes (VLT, there are four of these, each with a primary

mirror of 8.2 metres in diametre) and the New Technology Telescope (3.58 metres primary) of the European

Southern Observatory (ESO) located in Chile for their observations. Andrea Ghez and her group used the Keck

Telescopes (there are two of these, each with primary of 10 m diametre) located in Hawaii. Both the groups

followed the motion of many stars in the cluster, with the best results obtained for a star designated as S2, which

during its motion very closely approaches the central object. We will describe the results obtained for this star

over the years, beginning with results based on about a decade of observations during 1992-2002 and published

by the two groups in 2002-03. Shown in Figure 8 are the results on the motion of star S2 obtained by Genzel’s

group.

Figure 8: The red image on the left shows a number of stars from the around Sgr A*. The position of S2 at

various times from 1992 to 2002 are on the right of the ﬁgure, along with the position of Sgr A*. The short

bars on each point indicate the extent of the possible errors in the position. The path of the star over the years

is indicated by the ellipse, which is obtained as the best representation of the data. It is found that Sgr A* is

located at one focus of the rather elongated ellipse (every ellipse has two foci). Figure courtesy ESO.

From Newton’s theory of gravitation, it follows that a large mass located at the focus, coincident with Sgr

A*, should be exerting a gravitational force on star S2, which makes it follow the elliptical orbit. Calculations

showed that the mass of the object is close to 3.3 million times the mass of the Sun. The time taken by S2 to

complete one orbit, which is known as the period of the orbit, was about 15.6 years and the closest distance of

approach of S2 to Sgr A* was about 0.73 light days, that is about 19 billion kilometres. The estimated mass has

increased somewhat as further data has accumulated, it is now 4.3 million times the mass of the Sun and the

period is 16.02 years. of observations

The two groups of astronomers continued to monitor S2 beyond 2003. The later observations provided

data points which covered the whole ellipse, as they were taken over a span of time longer than one orbital

period. A complete orbit is shown in Figure 9.

1.6 The Orbit of Star S2

Figure 9: The ﬁgure shows data points independently determined by the group of Reinhardt Genzel (in blue)

and Andrea Ghez (in red). There is close agreement between the results of the two groups. An ellipse is seen

to provide a very good representation of the data points, except for a small deviation seen at the top, which is

discussed further in the text. Figure source: Gillessen et. al. Astrophysical Journal, 707, L114, 2009.

A close look at Figure 9 shows that the ellipse is not quite perfect. At the top of the ﬁgure, it is seen that the

ellipse does not quite close on itself, there is a small gap. The gap indicates that there is a slight rotation in the

ellipse from orbit to orbit. We know that Sgr A* is a very compact source, so it should appear as a point in the

diagram. But the source appears to be slightly extended, as can be seen near the bottom of the ﬁgure. This is

interpreted as a slight motion of the gravitating mass. Such a motion would produce a deviation from a perfect

elliptical orbit, as is observed, and seemed to provide the right explanation. But later observations showed that

the deviation is due to a quite diﬀerent eﬀect.

Figure 10: The trajectory of S2, including new points from the Gravity experiment. The ﬁgure is explained in

the text. Figure source Gravity Collaboration, Astronomy and Astrophysics, 636, L5, 2020.

1.7 Historical Note

In Einstein’s theory of gravitation, the orbit of a body around another gravitating body is nearly elliptical,

but not exactly. Over successive orbits, the eﬀect of the theory is to produce a slight rotation of the ellipse. This

is known as the precession of the perihelion, because the shift in the ellipse means that the point at which the

two bodies are closest shifts from orbit to orbit. A result of the shifting ellipse would be the kind of gap which

is observed.

Recently, the accuracy of the measurements of the position of S2 over time has improved more than

tenfold, because of the availability of a new instrument called GRAVITY, which combines the optical beams

from the four VLTs. In Figure 10 are shown (Figure Courtesy ESO) the results obtained by the GRAVITY team,

combined with earlier observations.

Figure 11: The trajectory of star S2 over many years. Each orbit around the supermassive object takes about

16 years and is a nearly perfect ellipse. Over each orbit the ellipse swings through about 0.2 degrees. The time

taken by the ellipse to complete a full circle is about 28,800 years, as is explained in the text. Image courtesy

(Credit: ESO/L. Calc¸ada).

The grey curve, which is a near-perfect ellipse except for the small deviation near the top as before, has been

obtained by calculations which use Einstein’s theory. The observed points are very well aligned with the curve,

which shows how precise the observations are. The black cross towards the bottom of the ellipse indicates the

position of the massive compact object. The red crosses indicate infra-red ﬂashes observed around Sgr A*. In

this new data set, there is no evidence that the object is moving, it is stationary. So, the deviation of the ellipse

for a perfect form is wholly explained as being due to the eﬀect of precession predicted by Einstein’s theory.

This is a remarkable result, and is the ﬁrst observed precession in the context of a supermassive object. Such

precession has previously been observed in the orbit of Mercury around the Sun and in binary pulsar systems.

Because of the precession, the trajectory of S2 over many years looks like an ellipse which keeps rotating in

space. This is shown schematically in Figure 11. The precession is about 0.2 degrees per orbit. Since the orbital

period S2 is about 16 years, the ellipse will complete one full rotation through 360 degrees in about 28,800

years.

1.7 Historical Note

The notion of the presence of a supermassive black hole at the centre of galaxies became popular after the

discovery of quasars, active galactic nuclei, superluminal motion etc. pointed to the existence such objects at

the centres of galaxies. In the mid-1960s, before many of these observations were made, Fred Hoyle and Jayant

1.8 Is the Compact Object a Supermassive Black Hole?

Narlikar had considered the existence of objects, with mass of about a billion times of the mass of the Sun,

around which matter would condense to form a galaxy. They predicted from their model that there would be

highly concentrated points of light in the centres of elliptical galaxies. This is probably the ﬁrst mention of a

supermassive object at the centre of galaxies.

1.8 Is the Compact Object a Supermassive Black Hole?

What is the nature of the compact object? We have seen that the closest distance of approach of star S2

to the gravitating object is about 19 billion kilometres, so the size of the object should be much smaller. If the

compact mass is a black hole, then it would have zero radius. But if it is not, it would have a non-zero small

radius. What could such an object be? The object is not emitting much light, so it must be made up of a large

number of small dark objects, like stars of very low mass, or neutron stars, or even stellar mass black holes

which are the remnants of stars. But calculations show that any such aggregate would have very high density at

the centre, and would collapse under its own gravity, or simply dissipate, in less than a million years. This time

is far, far shorter than the known age of the galaxy, which is more than 13 billion years. More exotic possibilities

like a dark mass made up of exotic particles like neutrinos, gravitinos, axinos etc., or a ball of bosons are

also possible. But it is presently not possible to produce robust models of such objects consistent with known

theory and observations. It appears therefore that most of the mass of the compact object at the centre of our

galaxy is the mass of a supermassive black hole, with the remaining mass being due to a star cluster around it.

The GRAVITY instrument and other future instruments will enable us to penetrate deeper towards the object,

possibly revealing eﬀects which could be unambiguously attributed to a supermassive black hole.

Acknowledgment

Much of the material in this article ﬁrst appeared in The Edition (inthedition.wordpress.com). I thank Mr.

Kaushik Bhowmik for permission to use it. This article will serve as the basis for further articles on black holes

in airis4D

About the Author

Professor Ajit Kembhavi is an emeritus Professor at Inter University Centre for Astronomy and

Astrophysics and is also the Principal Investigator of the Pune Knowledge Cluster. He was the former director

of Inter University Centre for Astronomy and Astrophysics (IUCAA), Pune, and the International Astronomical

Union vice president. In collaboration with IUCAA, he pioneered astronomy outreach activities from the late

80s to promote astronomy research in Indian universities. The Speak with an Astronomer monthly interactive

program to answer questions based on his article will allow young enthusiasts to gain profound knowledge about

the topic.

The Little Green Men in Space

by Linn Abraham

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Radios, apart from being a fun and useful piece of technology, are critical to almost all communication

in our daily lives. Did you know that cellphones use radio waves for communicating with each other? Or that

radio waves are used in WiFi or Bluetooth communications? If you have ever used a traditional radio receiver

for tuning into radio stations you have seen that between the designated stations which broadcast programs all

you get is noise. If so why do we use radio waves in astronomy? Do we expect radio stations to be run by the

denizens of galaxies near or far? Turns out that we did ﬁnd such a broadcast when we listened very carefully into

the depths of space using powerful radio receivers. Lets ﬁnd out what that was all about and more interesting

things about radio astronomy in this article.

2.1 What are radio waves?

Radios were all the rage in the not so distant past. Radio waves revolutionized communication and enter-

tainment in the early 20th century. During world war II radios played a critical role in military communication.

In the 1960s and 1970s, FM radio gained popularity for its superior sound quality. Although today radio has

mostly fallen out of fashion in the entertainment industry they are critical to most of our modern communication.

Television broadcasting, satellite communications and even the GPS system used for navigation all make use

of radio waves. Modern methods of wireless communication such as cellphone communication, WiFi and

Bluetooth would be unthinkable without radio waves. Radars that make use of radio waves are used for various

applications. Medical imaging is another area which makes use of radio waves. Radio Frequency Identiﬁcation

(RFID) tags are used in a variety of places and are fast becoming common place.

2.2 Why did radio waves have to be discovered?

Life that evolved on earth developed mechanisms to detect visible light, the things that we call eyes,

probably because of the interplay of various factors. The visible spectrum corresponds to wavelengths that

are easily absorbed and reﬂected by the materials found in nature, such as water, plants, and animals. This

allows organisms to perceive and distinguish objects in their environment, such as food, predators, and potential

mates. Furthermore, the Sun emits a large amount of radiation in the visible portion of the spectrum, which

provides the energy for photosynthesis and drives the Earth’s climate and weather patterns. The other parts

of the electromagnetic spectrum remained largely hidden from us. After Maxwell predicted the existence

of electromagnetic waves in 1864, Heinrich Hertz produced radio waves in the late 1880s to experimentally

2.3 What is electromagnetic radiation?

demonstrate their existence. The discovery of other kinds of EM waves such as X-rays soon followed. The

discovery of radio waves sparked a revolution in science and technology. The telegraph which was the primary

means of long distance communication now had a wireless alternative. Other uses of these newly discovered

waves soon followed. Before we learn more about radio waves it is essential to have a good understanding of

what electromagnetic radiation is and how it is produced both naturally and by artiﬁcial means.

2.3 What is electromagnetic radiation?

Stationary electric charges produce electric ﬁelds, whereas moving electric charges produce both electric

and magnetic ﬁelds. When a charge accelerates or undergoes a change in its magnitude (charge or discharge) a

disturbance in the electromagnetic ﬁeld is produced. This is often a pulse of electromagnetic radiation that can

self propagate outwards at the speed of light. This is very much like how you create a ripple on the surface of a

still pond by throwing a stone into it. In order to create a steady wave as opposed to just a pulse you need to have

a source that oscillates regularly. The frequency of such a wave is decided by the frequency of the source of the

disturbance and it is this frequency that decides the properties of the radiation. The frequency also decides the

energy of the radiation and often our ability to detect the radiation. Thus electromagnetic radiation falls on a

broad spectrum depending on the frequency from low frequency radio waves to the highest frequency gamma

rays.

2.4 What are the mechanisms that create EM waves?

The diﬀerent possible ways in which you can accelerate charges leads to the diﬀerent source of electromag-

netic radiation. In general you have thermal and non-thermal sources of electromagnetic waves. Examples of

thermal radiation include Continuous spectrum emissions related to the temperature of the object or material and

Speciﬁc frequency emissions from neutral hydrogen and other atoms and molecules. Examples of non-thermal

mechanisms include emissions due to synchrotron radiation and Ampliﬁed emissions due to astrophysical

masers. In a hot object be it solid, liquid or gas the constituent particles are always in accelerated motion either

due to vibration or collisions. As a result electromagnetic radiation is emitted at all possible frequencies in the

spectrum but the amount of energy in each frequency depends on the temperature of the object.

2.5 A radio universe

We already saw several man-made source of radio waves. What are some of the natural sources of radio

waves? On Earth, the discharge of electricity in lightnings is a source of radio waves. The bigger source of radio

waves is the Sun. Solar ﬂares and coronal mass ejections can produce intense bursts of radio waves that can

be a threat to our communication and navigational systems. These radio waves are produced through a process

called plasma emission. Plasma, which is a highly ionized gas, can produce radio waves when it is disturbed

by energetic charged particles. In the case of solar ﬂares and CMEs, the charged particles accelerated by the

events can create disturbances in the plasma that lead to the emission of radio waves.

In 1932, Karl Jansky a radio engineer at the Bell Telephone Laboratories was assigned to study radio inter-

ference from thunderstorms. During his investigations he found static that he concluded to be of extraterrestrial

origin and not related to the static from thunderstorms. Upon further investigations he discovered the source of

the static to be the Milky way galaxy and opened up radio astronomy as a ﬁeld of study. The typical astrophysical

sources of radio waves are dark dust clouds (refer Fig 4). Visible stars radiate a lot of electromagnetic energy by

2.6 The story of the Little Green Men

Figure 1: Full-size replica of the ﬁrst radio telescope, Jansky’s dipole array of 1932, preserved at the US Green

Bank Observatory in Green Bank, West Virginia. Image Credit: Wikipedia

Figure 2: The ﬁrst parabolic ”dish” radio telescope by Grote Reber, Wheaton, Illinois, 1937. Image Credit:

Wikipedia

deﬁnition in the visible region of the spectrum. Part of the energy has to be in the microwave (short wave radio)

part of the spectrum, and that is the part astronomers study using radio telescopes. Despite their temperatures

however all visible stars do not make for good radio frequency sources. We can detect stars at radio frequencies

only if they emit by non-thermal mechanisms, or if they are in our solar system (that is, our sun), or if there is

gas beyond the star which is emitting (for example, a stellar wind).

2.6 The story of the Little Green Men

In 1967, Jocelyn Bell, a postgraduate student at Cambridge University, was working with her supervisor,

Antony Hewish, on a radio telescope designed to study quasars. While analyzing the data from the telescope,

she noticed a strange signal that was pulsing regularly. The signal diﬀered from the generally chaotic nature of

most cosmic phenomenon. Additionally the signal was of a very speciﬁc radio frequency, whereas most natural

sources typically radiate across a wider range. These things led the research team to contemplate the possibility

2.6 The story of the Little Green Men

Figure 3: The LOFAR core (”superterp”) near Exloo, Netherlands. The bridges give an idea of the scale.

Image Credit: Wikipedia

Figure 4: Table showing diﬀerent astrophysical sources of radio waves

2.7 Why do we need radio astronomy?

Figure 5: Graphical depiction of Pulsars. Image Credit: Space.com

of having found an artiﬁcially created signal from an extraterrestrial intelligent species. They even dubbed the

source as LGM 1 or Little Green Men 1. However their discovery later turned out to be the ﬁrst detection of a

pulsar. Pulsars spin rapidly, while simultaneously radiating opposing beams of radio waves out into space. The

setup is similar to a lighthouse that spins around one up-and-down axis and radiates two beams of light from

a second axis. Antony Hewish and Martin Ryle received the Nobel Prize in physics in 1974 for their work in

radio astronomy and pulsars.

2.7 Why do we need radio astronomy?

Radio astronomy allows us to detect and study phenomena that emit little or no visible light, such as

pulsars and black holes. And in this way they act as complimentary source of information. Radio waves can

pass through clouds of dust and gas, allowing astronomers to observe objects that are obscured from view

in optical telescopes. They can travel vast distances without being absorbed, making it possible to study

objects at the farthest reaches of the universe. Finally radio telescopes can collect data continuously, day and

night, providing astronomers with a wealth of information about the universe around the clock Apart from the

discovery of pulsar radio astronomy has resulted in several remarkable astronomical discoveries. The discovery

of the cosmic microwave background radiation was made using the radio telescope in New Jersey. It provided

irrefutable evidence for the Big Bang model of the origin of the universe. More recently in 2019, the Event

Horizon Telescope project produced the ﬁrst-ever image of a black hole. The image was created using data from

a network of radio telescopes around the world.

2.8 The challenges in radio astronomy

Radio astronomy is also a very technically demanding ﬁeld. The building of large radio telescope involves

complex engineering and construction challenges, and maintaining and operating these instruments requires

highly trained professionals. Radio waves are relatively weak and can be easily overwhelmed by noise from

various sources, including terrestrial sources such as cell phone towers and other electronic devices. Thus

radio telescopes must be carefully designed and located in remote areas to minimize interference. Scattering

2.9 References

Figure 6: The proposed Lunar Crater Radio Telescope (LCRT) on the Far-Side of the Moon. Image

Credit:NASA

and distortion of radio signal by the atmosphere is another area of concern. Radio astronomers must carefully

calibrate their instruments and develop sophisticated data analysis techniques to separate the desired signal from

the background noise. These and other eﬀorts in radio astronomy often involves analyzing vast amounts of data,

which requires sophisticated computer algorithms and data processing techniques.

2.9 References

1. Little Green Men? Pulsars Presented a Mystery 50 Years Ago

2. Cosmic microwave background

3. PSR B1919+21

4. Basics of Radio Astronomy by Diane Fisher Miller

About the Author

Linn Abraham is a researcher in Physics, specializing in A.I. applications to astronomy. He is

currently involved in the development of CNN based Computer Vision tools for classiﬁcations of astronomical

sources from PanSTARRS optical images. He has used data from a several large astronomical surveys including

SDSS, CRTS, ZTF and PanSTARRS for his research.

Pulsating Variable Stars

by Sindhu G

airis4D, Vol.1, No.4, 2023

www.airis4d.com

3.1 Introduction

A variable star is a star whose brightness as seen from Earth varies with time. As we have observed, the

brightness ﬂuctuations exhibited by variable stars range from a thousandth of a magnitude to twenty magnitudes

and occur over a time frame that can vary from a fraction of a second to several years, depending on the type of

star. From the previous article “Variable Stars” we get an idea about what variable stars are. This article will

explore the fascinating world of pulsating variable stars, a type of intrinsic variable stars. Intrinsic variables

are characterized by changes in the physical properties of the star, resulting in variations in brightness. The

brightness of pulsating variable stars varies due to a physical change within the star. The surface layers of

pulsating variables expand and contract periodically. This means the star periodically grows and shrinks in size.

The luminosity of pulsating stars ﬂuctuates as they expand and contract, altering their luminosity and spectral

characteristics. Pulsations can be divided into two categories: radial pulsations, in which the entire star expands

and contracts as a unit, and non-radial pulsations, where certain parts of the star expand and others contract. A

star that pulsates radially will stay in a spherical shape, however, a star that experiences non-radial pulsations

may deviate from a sphere periodically.

Astronomers ﬁnd the study of pulsating variables to be of great importance. By analyzing the light curves

of these variables, they are able to gain insight into the interior processes in stars. The most beneﬁcial aspect of

many types of pulsating variables is their period-luminosity relationship, which allows astronomers to measure

the distance to such stars.

3.2 Why do they pulsate?

If the outward pressure of a star exceeds the inward-acting gravitational force, its outer layers will expand

outward. As the star expands, its gravitational force weakens while its outward pressure drops at an even higher

rate. This leads to a point where gravity and pressure are in equilibrium. However, the outward-moving layers

have momentum that resists a change in motion. This momentum carries the layers past the equilibrium position.

As the gravitational force acts on the layer it slows down. At a certain point, the outward gas and radiation

pressure is no longer strong enough to counter the inward-acting gravitational force, causing it to stop. Due to

an imbalance of forces, the outer layers of the star start to collapse inwards. This causes gravity to increase, but

the pressure increases at a higher rate. With the outward pressure being greater than the inward gravitational

force, the collapsing layer slows down and eventually stops. This brings us back to the start, where the outward

3.3 Diﬀerent types of pulsating variables

Figure 1: Basic Hertzsprung - Russell diagram (Image Courtesy: NASA)

pressure is stronger than the gravitational force, thus beginning the pulsation cycle once more.

3.3 Diﬀerent types of pulsating variables

The diﬀerent types of pulsating variables can often be distinguished by their pulsation period, mass,

evolutionary status, and the characteristics of their pulsations. Cepheids, RR Lyraes, and Long Period

Variables are pulsating variable stars and occupy regions on the H-R diagram called instability strips. The

Hertzsprung–Russell(H-R) diagram is a scatter plot of stars showing the relationship between the star’s absolute

magnitudes or luminosities versus their stellar classiﬁcations or eﬀective temperatures (Figure 1) and (Figure

2).

The term instability strip (Figure 3) usually refers to a region of the Hertzsprung–Russell diagram largely

occupied by several related classes of pulsating variable stars. The instability strip intersects the main sequence

in the region of A and F stars and extends to G and early K bright supergiants. The vast majority of stars located

above the main sequence in the instability strip are variable. At the point where the instability strip meets the

main sequence, the majority of stars are usually stable, but certain variable stars, such as roAp stars, may exist

there as well. Most stars are classiﬁed according to their temperatures (spectral type) from the hottest to the

coolest: O, B, A, F, G, K, and M. These categories are further divided into subclasses from the hottest (0) to

the coolest (9) (Figure 4). For example, the hottest B stars are B0 and the coolest are B9, followed by A0. Each

major spectral type has its own unique spectra. The most common stellar classiﬁcations seen on H-R diagrams

are O, B, A, F, G, K, and M, but there are also additional and extended spectral classes like Wolf-Rayet stars

(W), cool dwarfs (L), brown dwarfs (T), carbon stars (C), and stars with zirconium oxide lines that are between

M and C stars (S). Some examples of type of pulsating variables are given below:

3.3.1 Cepheids(CEP)

Cepheids (Figure 5)are pulsating variables which are very luminous and have large masses, and they exhibit

periodic variations in brightness over a period of time ranging from 1 to 70 days. They take their name from

3.3 Diﬀerent types of pulsating variables

Figure 2: H-R diagram of luminosity vs temperature (Image Courtesy: The European Southern Observatory)

Figure 3: An HR diagram with the instability strip (Image Courtesy: ATNF)

3.3 Diﬀerent types of pulsating variables

Figure 4: Some examples of Harvard classiﬁcation stellar spectra (Image Courtesy : NASA)

Delta Cephei, the ﬁrst-such pulsating variable star discovered by John Goodricke in 1784. Cepheid light curves

are typically characterized by a sharp increase in brightness followed by a gradual decline, resembling a shark

ﬁn. The amplitude range of these light curves usually ranges from 0.1 to 2. Cepheids have a high luminosity and

are spectral class F at maximum, and G to K at minimum. The later the spectral class of a Cepheid, the longer its

period. Cepheids have a ﬁxed relationship between period and absolute magnitude, as well as a relation between

period and mean density of the star. Henrietta Leavitt (Figure 6) was the ﬁrst to discover the period-luminosity

relationship of Delta Cepheids, which made them very eﬀective for measuring the distance to galaxies within

the Local Group and even outside of it. Edwin Hubble then used this method to conﬁrm that the spiral nebulae

were actually distant galaxies. There are two types of Cepheids: Type I or Classical Cepheids and Type II, or W

Virginis Cepheids. Both of these types are located in the Instability Strip of the Hertzsprung-Russell Diagram.

3.3.1.1 Classical Cepheid variables

Classical Cepheids, also known as Delta Cephei variables, are yellow supergiants belonging to population

I. They are young, massive size, and high luminosity stars. These stars exhibit regular pulsations that occur

over a period of several days to months. The majority of classical Cepheids display a period that falls within

the range of 5 to 10 days, and their amplitudes in visible light range between 0.5 to 2.0 magnitudes. These

ﬂuctuations are less noticeable when observed at infrared wavebands. These stars are known to be 1.5 to 2

magnitudes more luminous than Type II Cepheids. A well- deﬁned relationship between period and luminosity

exists among Classical Cepheids. The longer the Cepheid’s period, the more luminous it is intrinsically. This is

a signiﬁcant ﬁnding, as it enables Classical Cepheids to be utilized as standard candles to determine distances.

Type I Cepheids are located on the Instability Strip of the Hertzsprung-Russell (HR) diagram.

3.3.1.2 Type II Cepheids

Type II Cepheids, historically known as W Virginis stars, exhibit extremely regular light pulsations , and

their luminosity relationship is comparable to that of Delta Cephei. These stars belong to population II stars

and are older than Type I Cepheids. Type II Cepheids typically have lower metallicity, lower mass, and slightly

lower luminosity than Type I Cepheids. Their period versus luminosity relationship is slightly oﬀset from that

of Type I Cepheids.

3.3 Diﬀerent types of pulsating variables

Figure 5: RS Puppis, one of the brightest known Cepheid variable stars in the Milky Way galaxy (Image Courtesy

: Hubble Space Telescope)

3.3.2 RR Lyrae variables

RR Lyrae variables (Figure 7) are older pulsating white giants that exhibit low metallicity. They are common

in globular clusters. RR Lyrae stars belong to Population II and are older than Type I Cepheids. However,

they have lower mass than Type II Cepheids. They also exhibit a well-established relationship between their

period and luminosity, making them useful for determining distances as well. RR lyrae stars are characterized

by their brief pulsation cycles, typically ranging from 1.5 hours to a day, and exhibit a brightness range of 0.3

to 2 magnitudes. These stars typically belong to spectral classes A7 to F5. Sub-types of RR Lyrae variables are

classiﬁed according to the shape of their light curves. RR Lyraes ﬁt on the Instability Strip of the HR diagram.

3.3.3 RV Tauri variables

RV Tauri variables are yellow supergiant stars. Their light curves show alternating deep and shallow

minima. The double-peaked variation commonly observed in RV Tauri variables is typically characterized by

periods ranging between 30 to 100 days and amplitude ranges of 3 to 4 magnitudes. At maximum brightness,

the spectra of RV Tauri variables are categorized as type F or G, while at minimum brightness, they are type K

or M. These stars are situated close to the instability strip, displaying cooler temperatures than Type I Cepheids

but higher luminosity than Type II Cepheids.

3.3.4 Long period variables

LPVs (long-period variables) refer to red giants or supergiants that pulsate with periods ranging between

30 to 1000 days. Typically, these stars have spectral types of M, R, C, or N. There are two categories of LPVs:

Mira and Semiregular.

3.3 Diﬀerent types of pulsating variables

Figure 6: Henrietta Swan Leavitt (Image Courtesy : Popular Astronomy)

Figure 7: RR Lyrae (Image Courtesy : Digitized Sky Survey - STScI/NASA, Colored & Healpixed by CDS)

3.3 Diﬀerent types of pulsating variables

Figure 8: Pulsating mira variable star (Chi Cyg) (Image Courtesy : SAO/NASA)

3.3.4.1 Mira variables

Mira variables (Figure 8) are red giants that exhibit periodic pulsations. Mira, in particular, displays a

cycle lasting 331 days, during which its brightness varies by nearly 6 magnitudes in the visible waveband. Its

eﬀective temperature ranges from 1,900 K to 2,600 K. Mira is also a visual binary, with its companion star being

a variable as well. The Mira-type stars are characterized by long periods that range between approximately 80

to 1,000 days, and their visual brightness varies by 2.5 to 10 magnitudes.

3.3.4.2 Semiregular variables

Semiregular variables are giants and supergiants that display periodicity along with intervals of semi-

regular or irregular light changes. Typically, these stars exhibit periods ranging from 30 to 1000 days, with

amplitude variations generally less than 2.5 magnitudes.

3.3.5 Delta Scuti variables

Delta Scuti variables share similarities with Cepheids, but they are notably dimmer and exhibit much

shorter periods. These stars typically display an amplitude range of 0.003 to 0.9 magnitudes and a period

ranging from 0.01 to 0.2 days. Their spectral type generally falls between A0 and F5.

3.3.6 Rapidly oscillating Ap variables

These stars, which belong to the spectral type A, and occasionally F0, are a subclass of Delta Scuti variables

situated on the main sequence. They exhibit remarkably rapid variations, with periods ranging a few minutes

and amplitudes of a few thousandths of a magnitude.

3.3 Diﬀerent types of pulsating variables

3.3.7 Beta Cephei variables

Beta Cephei variables undergo pulsations of short periods, typically ranging from 0.1 to 0.6 days. These

pulsations have an amplitude of 0.01 to 0.3 magnitudes. Many stars of this kind exhibit multiple pulsation

periods.

References

Understanding Variable Stars, John R Percy, Cambridge University Press.

Pulsating Variable Stars

GCVS variability types, Samus N.N., Kazarovets E.V., Durlevich O.V., Kireeva N.N., Pastukhova

E.N.,General Catalogue of Variable Stars: Version GCVS 5.1,Astronomy Reports, 2017, vol. 61, No. 1,

pp. 80-88 2017ARep...61...80S

Pulsating Variable Stars and The Hertzsprung-Russell Diagram

Types of Variable Stars: A Guide for Beginners

Variable star

About the Author

Sindhu G is a research scholar in Physics doing research in Astronomy & Astrophysics. Her research

mainly focuses on classiﬁcation of variable stars using diﬀerent machine learning algorithms. She is also doing

the period prediction of diﬀerent types of variable stars, especially eclipsing binaries and on the study of optical

counterparts of x-ray binaries.

Part III

Biosciences

The Avatars of Stem Cells and its applications

in healthcare

by Geetha Paul

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Stem cells are undiﬀerentiated cells that can develop into various types (avatars) of specialised cells

in the body. They have the unique ability to divide and self-renew, making them essential in developing,

growing, and repairing tissues in the body. Self-renewal is the ability of cells to proliferate without the loss of

diﬀerentiation potential and without undergoing senescence (biological ageing). By creating patient-speciﬁc

stem cell derivatives, researchers can study the eﬀects of drugs on a speciﬁc individual. Using stem cells to

diﬀerentiate into any cell type in the body, they can also ﬁnd applications for generating diﬀerent types of cells

and tissues for research purposes. Thus stem cell avatars have attracted signiﬁcant attention in modern research

in view of their potential applications in medicine and biotechnology.

Figure 1: Origin of Embryonic stem cells from the zygote (oocyte+sperm) to totipotent cells forming the

Blastocyst. Cells from the inner lining of the blastocyst (blastomeres) are separated and cultured as embryonic

stem cells. [Image courtesy:https://www.dreamstime.com]

There are two main types of stem cells: embryonic stem cells (ESCs) and adult stem cells (ASCs).

Embryonic stem cells (ESCs) are derived from the inner cell mass of a blastocyst, which is a very early

stage (3 to 5 days old) of embryonic development. At this stage, an embryo is called a blastocyst and has about

150 cells.

These are pluripotent stem cells, meaning they can divide into more stem cells or can become any type of

cell in the body. This versatility allows embryonic stem cells to regenerate or repair diseased tissue and organs.

Adult stem cells ( ASCs) There are two types of adult stem cells. One type comes from fully developed

tissues such as the brain, skin, and bone marrow. They are found in small numbers in most adult tissues

throughout the body. The second type is induced pluripotent stem cells (iPSCs). These adult stem cells are

derived from somatic cells isolated from various tissues and cultured in vitro. The cells are exposed to a combi-

nation of transcription factors called ’reprogramming’ to achieve the pluripotent state. Once reprogramming is

completed and the cells express speciﬁc pluripotency markers, the resulting cells are induced pluripotent stem

cells – iPSCs. Compared with embryonic stem cells, adult stem cells have a limited ability to give rise to various

body cells.

Stem cells are classiﬁed according to their origin, possibly from the embryo, umbilical cords or adult

tissue. Cord and ES cells are generally said to be equivalent to pluripotent blastomeres. At the same time, adult

stem cells (also found in children and adults) possess a more comprehensive range of plasticity ranging from

pluripotent to multipotent (Table 1).

Level of plasticity Type of cells Cell Types Obtainable

Totipotent Zygote All cell types develop into embryos.

Pluripotent Embryonic Stem cells (ES cells) All cell types do not develop into embryos.

Pluripotent Umbilical stem cells Same as ES cells

Pluripotent to multipo-

tent

Adult stem cells (AS cells)

a.Haematopoietic stem cells

b. Mesenchymal stem cells

c.Others (undeﬁned collection

method)d.Transdiﬀerentiation

(under special conditions)

d.Transdiﬀerentiation (under

special conditions)

a.Red, white blood cells, immune cells, others

b.Lipocytes, cartilage, bone, tendon and liga-

ments, myocytes, skin cells, neurons etc.

c.Others (collection methods still not very well-

characterised)

are Neurons, hepatocytes, and pneumocytes.

d.Transdiﬀerentiation (occurs under special con-

ditions)

Crossing of germ layers.

Table 1: A Comparison of the diﬀerent types of Stem Cells in Terms of Plasticity:

Totipotent stem cells are capable of giving rise to all other type of cells. This includes the fertilised egg

and the subsequent new cells formed until approximately day 4 of development.

Pluripotent stem cells refer to embryonic stem cells (ESCs) and induced pluripotent stem cells. These

cells can give rise to all types except those that form the amniotic sac and the placenta. ESCs (embryonic

stem cells) come from the inner cell mass in the blastocyst, while iPSCs (induced pluripotent stem cells) are

generated from somatic cells.

Multipotent stem cells are also called adult stem cells, these are cells that can develop into closely related

cell types.

Figure 2: Shows the culture of Adult stem cells. (iPSCs)The adult somatic cells are cultured, reprogrammed

into induced pluripotent stem cells, and ﬁnally, they are diﬀerentiated to speciﬁed organ cells.

[Image courtesy: https://images.app.goo.gl]

1.1 What is the procedure behind stem cell therapy?

Figure 3: Shows the origin of Totipotent, Pluripotent, and Multipotent types of Stem Cells .

[ Image courtesy: https://theory.labster.com/stem-cell-types/]

1.1 What is the procedure behind stem cell therapy?

The procedure behind stem cell technology depends on the type of stem cells used and the speciﬁc

application. However, some general steps are often involved in the process:

Source of Stem Cells: The ﬁrst step is to obtain a source of stem cells. This could be embryonic stem

cells derived from a blastocyst or adult stem cells obtained from bone marrow or adipose tissue.

Isolation and Culturing: Once the stem cells are obtained, they are isolated and cultured in the laboratory.

This involves growing the cells in a special culture medium that provides the necessary nutrients and growth

factors for the cells to divide and diﬀerentiate.

Diﬀerentiation: Depending on the application, the stem cells may need to be induced to diﬀerentiate into

speciﬁc cell types. This can be done by exposing the cells to speciﬁc growth factors or chemicals that stimulate

diﬀerentiation.

Transplantation: If the goal is to use stem cells for regenerative medicine, the diﬀerentiated cells can be

transplanted into the patient’s body. This can be done through various methods, such as injection or implanting

the cells into the damaged tissue.

Monitoring and Follow-up: After transplantation, the patient’s progress is monitored to ensure that the

cells function as intended and that there are no adverse eﬀects.

1.2 Applications of stem cells in modern research

In some cases, additional treatments or follow-up procedures may be required.

Figure 4: Steps involved in stem cell therapy.

[Image courtesy: https://www.drkraghu.in/ home]

1.2 Applications of stem cells in modern research

Stem cells have wide applicability in medical science, drug development and patient care. In drug discovery,

it allows patient speciﬁc drugs based on their stemcells. These stemcells can be developed and tested even

before administering it to the concerned.

Researchers can test how new drugs interact with those cells by growing stem cells in the laboratory and

inducing them to diﬀerentiate into speciﬁc cell types.

Figure 5: A schematic diagram of using stem cells based on Drug discovery.

1.3 Promises of Stem Cells In Regenerative Medicine

Stem cell therapy also plays a signiﬁcant role in Disease Modelling. Stem cells can be used to create

models of human diseases, which can help researchers study the underlying mechanisms of these diseases and

develop new treatments.

Figure 6: Somatic cells from the patient generate diseased iPSCs. These diseased iPSCs may be repaired by

Gene Therapy and used to create healthy somatic cells to be transplanted into the patient. They may also ﬁnd

applications to produce unrepaired somatic cells for disease modelling or drug screening.

[ Image courtesy:https://www.frontiersin.org/ articles/ 10.3389/ fcell.2015.00002/full]

The use of iPSCs (Induced Pluripotent Stem Cells) for disease modelling is based on the fact that they are

capable of self-renewing and that they can diﬀerentiate into all types of cells of the human body, which can

be utilised for the preparation of diﬀerent disease models to study those diseases. Moreover, a patient-speciﬁc

iPSC could be widely used in developing speciﬁc therapeutics regimens/drugs.

Regeneration of damaged or diseased tissues is a major application of stem cell therapy. Its potential in

treating conditions such as spinal cord injuries, heart mussile injuries due to heart diseases or diabetes are

various avenues for further research.

1.3 Promises of Stem Cells In Regenerative Medicine

Figure 7: Promises of stem cells in regenerative medicine: the six classes of stem cells, that is, embryonic stem

cells (ESCs), tissue-speciﬁc progenitor stem cells (TSPSCs), mesenchymal stem cells (MSCs), umbilical cord

stem cells (UCSCs), bone marrow stem cells (BMSCs), and induced pluripotent stem cells (iPSCs), have many

promises in regenerative medicine and disease therapeutics.

[ Image courtesy: https://www.ncbi.nlm.nih.gov/pmc/ articles/ PMC4969512/ ]

Tissue Engineering is a popular method using stem cells open new methods for growing tissues for organ

transplantation. This could eliminate the need for organ donors and reduce the risk of implant rejection.

Overall, stem cells have the potential to revolutionise medicine and signiﬁcantly improve our ability to treat

a wide range of diseases and conditions. It’s important to note that stem cell technology is still in the early stages

of development, and there is still much research to understand its potential and limitations fully. Additionally,

there are ethical concerns surrounding the use of embryonic stem cells, which has led to the development of

alternative sources of stem cells, such as induced pluripotent stem cells (iPSCs).

References:

1.Agius C.M, and R.Blundell (2007), Medwell Journals, International Journal of Molecular Medicine

and Advance Sciences 3 (4): 145-150, 2007

2.https://stemcellres.biomedcentral.com/articles/10.1186/s13287-019-1165-5

3.https://www.stanfordchildrens.org/en/topic/default?id=what-are-stem-cells-160-38

4.https://www.researchgate.net/publication/261041118

5.https://www.nature.com/articles/s41392-022-01134-4

6.https://www.dreamstime.com/%D0%BE%D1%81%D0%BD%D0%BE%D0%B2%D0%BD%D1%8B%

D0%B5-rgb-image163320490

1.3 Promises of Stem Cells In Regenerative Medicine

7.https://images.app.goo.gl/tMDcDzvo5Sbo9CVt6

8.https://theory.labster.com/ips

9.https://theory.labster.com/stem-cell-types/

10.https://www.drkraghu.in/home

11.https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4969512/

12.https://www.frontiersin.org/articles/10.3389/fcell.2015.00002/full

13.https://stemcellres.biomedcentral.com/articles/10.1186/s13287-019-1455-y

14.https://www.nature.com/articles/s41392-022-01134-4.

About the Author

Geetha Paul is one of the directors of airis4D. She leads the Biosciences Division. Her research

interests extends from Cell & Molecular Biology to Environmental Sciences, Odonatology, and Aquatic Biology.

Part IV

General

Beyond the Fluﬀy: The Enchanting World of

Clouds

by Robin Jacob Roy

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Clouds in the sky are one of the most beautiful and fascinating natural phenomena that we can witness in

our everyday lives. These masses of condensed water vapor ﬂoat in the sky, taking on a variety of shapes and

sizes that can be both enchanting and awe-inspiring.

Clouds are formed when water vapor in the atmosphere condenses into tiny water droplets or ice crystals,

and the type of cloud that forms depends on the altitude, temperature, and humidity of the air. Cloud formation

starts when the air contains as much water vapor (gas) as it can hold. This is called the saturation point. This

saturation point can be reached through two methods. Firstly, the air accumulates moisture until it reaches its

maximum capacity. Alternatively, the temperature of the moisture-ﬁlled air can be reduced, causing the amount

of moisture it can hold to decrease. Saturation can be achieved through evaporation or condensation, resulting

in the visible formation of water droplets in the form of fog or clouds.

Clouds play a crucial role in the Earth’s climate system, regulating the amount of incoming solar radiation

that reaches the Earth’s surface and controlling the planet’s energy balance. Clouds can indicate the likelihood

of precipitation and help regulate the temperature of the Earth’s surface. They can also create stunning displays

of color during sunrise and sunset, as the light interacts with the water droplets in the clouds.

1.1 Cloud Formation

Cloud formation can occur in several ways. One of the most common ways is through the process of

adiabatic cooling. When warm, moist air rises, it expands and cools due to the lower atmospheric pressure at

higher altitudes. As the air cools, it can no longer hold as much moisture, so water vapor begins to condense

into tiny water droplets or ice crystals, forming clouds.

Another way clouds can form is through the process of frontal lifting. When two air masses of diﬀerent

temperatures and humidities meet, the warmer, moister air is forced to rise over the cooler, drier air, causing the

moisture to condense into clouds.

1.2 Types of Clouds

Clouds are named based on two criteria. One way is by their location in the sky, as some clouds are high

up, while others are closer to the Earth’s surface. Low clouds can even touch the ground and are known as fog,

1.3 What Causes Rain?

Figure 1: Major Cloud Types occurring at various altitudes.Source: Extraction, classiﬁcation and visualization

of 3-dimensional clouds simulated by cloud-resolving atmospheric model, Daisuke Matsuoka, International

Journal of Modeling Simulation and Scientiﬁc Computing, April 2017.

while middle clouds are found between low and high clouds.

Another way of naming clouds is based on their shape. High clouds are called cirrus clouds, which

resemble feathers. Middle clouds are referred to as cumulus clouds, which look like large cotton balls in the

sky. Stratus clouds, on the other hand, are low clouds that cover the sky like bed sheets.

Here are some of the most common types of clouds which are distinguished based on their altitude, shape,

and composition:

1. Cumulus clouds: These clouds are white, ﬂuﬀy, and usually have a ﬂat base with a dome-like top. They

form when warm, moist air rises and cools, causing water vapor to condense into visible clouds.

2. Stratus clouds: Stratus clouds are low-lying and usually cover the entire sky in a uniform layer. They

can be gray or white and can form in both warm and cold air masses.

3. Cirrus clouds: These clouds are high up in the atmosphere and are thin and wispy in appearance. They

are made up of ice crystals and are often associated with fair weather.

4. Cumulonimbus clouds: These are large, towering clouds that are often associated with thunderstorms.

They have a ﬂat base and a dome-like top and can reach up to 60,000 feet in height.

5. Altostratus clouds: These clouds are mid-level clouds that form in the middle of the troposphere. They

are often gray or blue-gray in appearance and can produce light precipitation.

6. Stratocumulus clouds: These clouds are low-lying and often appear as rounded masses or rolls. They

can be white, gray, or dark in appearance.

7. Nimbostratus clouds: These clouds are dark and usually cover the entire sky. They are often associated

with heavy precipitation and can form at any altitude.

Understanding the various types of clouds can help in predicting weather patterns and making weather forecasts.

Major cloud types are shown in ﬁgure 1.

1.3 What Causes Rain?

Rain and snow are both forms of precipitation that form when water vapor in the atmosphere condenses

into liquid or solid water particles and falls back to the Earth’s surface.

Here’s the process of how rain and snow form from clouds:

1. Water vapor: The process begins with the presence of water vapor in the atmosphere. Water vapor is an

invisible gas that is created when water evaporates from the Earth’s surface and rises into the atmosphere.

2. Cloud formation: As water vapor rises, it cools and condenses into tiny water droplets or ice crystals,

forming clouds. Clouds can form at diﬀerent altitudes, depending on the temperature and humidity of the

air.

3. Precipitation formation: As the water droplets or ice crystals in the clouds grow larger and heavier, they

eventually fall back to the ground as precipitation. The type of precipitation that forms depends on the

temperature of the atmosphere at diﬀerent levels.

4. Rain formation: Rain forms when the temperature in the lower atmosphere is above freezing, causing the

water droplets to remain in a liquid state as they fall to the ground.

5. Snow formation: Snow forms when the temperature in the lower atmosphere is below freezing, causing

the water droplets to freeze into ice crystals. The ice crystals may then continue to grow as more water

vapor freezes onto them. The snowﬂakes become heavier and eventually fall to the ground.

6. Other forms of precipitation: Other forms of precipitation can also form, such as sleet or freezing rain,

when the temperature in the lower atmosphere is near or just above freezing, causing the precipitation to

partially melt before it reaches the ground.

Overall, the process of rain and snow formation from clouds is a complex interaction between temperature,

humidity, and air pressure in the atmosphere.

1.4 How Clouds Aﬀect the Earth’s Climate?

Clouds have a signiﬁcant impact on the Earth’s climate. They reﬂect incoming solar radiation back into

space, which helps to keep the planet cool. This is known as the albedo eﬀect. Clouds also absorb and re-radiate

heat energy, which can trap heat in the Earth’s atmosphere and contribute to the greenhouse eﬀect. The net eﬀect

of clouds on the Earth’s climate depends on a number of factors, including their type, altitude, and coverage.

Low-level clouds, such as stratus clouds, tend to have a cooling eﬀect on the Earth’s climate because they

reﬂect more sunlight than they absorb. This is particularly true for bright, white clouds, which have a high

albedo. Cumulus clouds, on the other hand, can have a warming eﬀect on the Earth’s climate because they

absorb more sunlight than they reﬂect. This is because cumulus clouds are often darker than stratus clouds, and

thus have a lower albedo.

High-level clouds, such as cirrus clouds, also have a warming eﬀect on the Earth’s climate. This is because

they absorb and re-radiate heat energy more eﬀectively than they reﬂect sunlight. This can lead to a net warming

eﬀect, particularly at night when the clouds trap heat in the Earth’s atmosphere.

In addition to their impact on the Earth’s climate, clouds also play an important role in the water cycle.

They are responsible for distributing moisture around the planet, which is essential for the growth of plants

and the survival of animals. When clouds release their moisture as precipitation, it can recharge groundwater

reserves and replenish rivers, lakes, and other bodies of our planet.

Overall, clouds are an important factor in the Earth’s climate system, and their properties and behavior are

a subject of ongoing research and study by climate scientists.

1.4 How Clouds Aﬀect the Earth’s Climate?

References:

Cloud Formation

Cloud Development

Cloud Introduction

Cloud Types

Extraction, classiﬁcation and visualization of 3-dimensional clouds simulated by cloud-resolving atmo-

spheric model, Daisuke Matsuoka, International Journal of Modeling Simulation and Scientiﬁc Comput-

ing, April 2017.

About the Author

Robin is a researcher in Physics specializing in the applications of machine learning for remote

sensing. He is particularly interested in using computer vision to address challenges in the ﬁelds of biodiversity,

protein studies, and astronomy. He is currently working on classifying satellite images with Landsat and Sentinel

data.

Part V

Fiction

Cranky the Scientist

by Ninan Sajeeth Philip

airis4D, Vol.1, No.4, 2023

www.airis4d.com

Cranky is not his real name. Nobody knows his real name. People call him Cranky because he comes up

with crazy ideas now and then for global problems. He is a wanderer with no domestic issues, which has helped

him focus on global problems like climate change, earthquakes, etc.

But this time, Cranky is not the subject of laughter anymore. His name has been nominated for the world

distinguished scientist award. The headlines in the newspaper surprised everyone. Many even turned over the

newspaper to verify the date to conﬁrm that it was not April ﬁrst! How can he, with so many well-known

scientists, be nominated for the world distinguished scientist award?

Whether good or bad news, all newspapers follow the same policy of creating a captive story. It was not

diﬀerent for the case of Cranky either. Starting from his ancestral history, his childhood stories and even some

of the things everyone used to joke about him have appeared with charm and courage sprinkled like salt and

pepper on the morning toast. Overnight transformation of everything he was ridiculed for.

What is it that he has done? Everyone heard that question from every corner of the room where the paper

was read.

Yes, the newspapers sold out like hot pancakes. Curious eyes swam across the lines as a ﬁsh dropped into

a new water tank. Oh my God was the common word that almost every mouth uttered.

All the newspapers said that Mr Crank had developed a permanent earthquake solution, which he is

honoured for.

Scientists know that earthquakes are caused by platelets that move beneath the surface of the earth’s outer

crust. When they move and push over others, tension grows with time. How much this tension can grow depends

on the strength of each platelet that gets pressed. It is like the ﬁght between superpowers where the cold war

continues until, at one point, it bursts to do unexpected damage to everyone. Tiny platelets easily break, and no

one notices the hundreds that cause minor earthquakes that only sensitive instruments can detect.

The scientiﬁc model of plate tectonics, or the movement of the tectonic plates in the earth’s lithosphere, is

globally approved as the cause of earthquakes. It is believed that continents drift from each other and become

separate due to the movement of these tectonic plates on which they rest. Thus, for example, South America ﬁts

perfectly to the left lower part of Africa, indicating that they were once part of the same landscape. When these

moving plates detach and push into other ﬂoating continents, they sometimes cause the plates to uplift the soil

above and slide over one another. Mount Everest is said to be the consequence of the Indian continent pushing

over the tectonic plates of the Asian continent, causing it to rise so high above the surface. The ﬁnding of sea

species fossils on the cliﬀs of Everest further substantiates the claim that this was really what happened.

What if the plates do not slip over each other and result in tectonic uplift like in the case of mount Everest?

It might have caused hundreds of thousands of years to push the plates to the height of Everest. A simple

calculation of the potential energy from high school science will tell us the enormous amount of energy involved

in lifting so much soil to the height of Everest.

Imagine this much energy being built up in the plates without the so-called tectonic uplift. There are two

possibilities. One is that the plates slip over one another and continue to move without pushing each other.

The other possibility is catastrophic. The two plates compete with each other, demonstrating their strengths

over hundreds of thousands of years, storing the entire energy to lift Everest as potential energy in the form of

internal stress.

How long can these plates withstand this stress? It is like a wrestling show. When both parties are of

comparable strength, the ﬁghts continue until both falls! Likewise, the plates will grow their stress until they

reach the breaking point and release their energy into the land around them.

Imagine someone lifting and dropping Everest into the sea! It will cause the sea to rise in all directions

and ﬂood all over. This is called a Tsunami. Tsunami is caused when an eruption happens under the sea. If it is

on land, it shakes the ground so heavily that it will behave like water! That’s why buildings fall, and trees get

uprooted in no time.

This is all known for ages, but what is so special about Cranky? He developed the technology for eﬃciently

releasing the energy stored in the platelets without dispersing it to the surroundings. That’s clever, isn’t it? Yes,

it is, and that’s what made him so famous.

What did Cranky do? Well, that is what we call genius. With no formal education or even a family to

support him, Cranky has been an orphan who took shelter under the shop roofs when they closed down and left.

He has been an eyewitness to several events. Many had no time or interest to bother them. Though was poor

due to obvious circumstances, he would save every penny he could to spend on anything that fascinated him.

He would be a silent listener to every conversation and every news that’s played on public television. He won’t

comment publicly as he has become an introvert, one of the hard lessons that life has taught him to remain safe.

On one such day, the TV said that a building built somewhere on the coast in a protected zone would be

demolished on a day next week. Cranky pulled out all his pockets, picked up every penny he could gather and

was determined to witness the excellent demolishing technique the TV described in detail. It said it would all

be done with perfect timing, making the building meltdown as a pile of dust within the blink of an eye.

Cranky was on time to the building site, which was already ﬂooded with onlookers like in a festival show,

waiting for the crackers to blast into the sky. The siren went one by one while all could hear their hearts beat.

Lo, the ﬁrst cracker burst, followed by the second, third and fourth until it was just sound alone that none could

count. The building burst without a shake; as was said, it was powder thin when it rolled from top to down,

making no shake or sound other than the people’s cry atop.

When it ran down tears to all, it burst sparks in Cranky’s brain. He rushed to the nearest shade and started

recapturing all he saw. It was not just like a bomb that bursts, shaking everything around, but a silent fall of

a tower without even letting anyone feel it. How could that be? That was his thought, and he broke his shell

and came out with this wonderful thought of breaking down those tectonic plates. He had no formal training in

science but had the brain, only genius’s own. He could relate it to someone who once said that the best way to

destroy an enemy is to let him destroy himself! He was sure the crackers were not like bombs, though they were

strong and loud. Indeed, their timing and strategic placement worked to unbind the bonds that held the building

high and convert its potential energy to dust it down.

Ureka, it struck his brain. Here is what would save the world, and let’s not let anyone cry anymore from

the wrath nature bestows without warning. That was just his turning point when he was heard by a passing guy,

who was just there on his holiday trip, but by himself, a famous physicist. No one knew where Cranky went,

Figure 1: Usually, when energy is released through plate tectonics, it is transmitted to all parts of the globe

through surrounding media and other platelets. The strategically guided triggering of platelet breaks made this

impossible, and the total energy could be conﬁned to the platelets without escaping. This energy was spent to

break down the platelets, turning them into a pile of sand conﬁned to the same volume.

nor anyone cared much where he went. However, they often had him in their casual talks.

Cranky had gone to Japan, where the gentleman took him from shore to work on a project for nature care.

The newspapers read the news from Tokyo, quoting the famous scientist, Cranky, a genius in line with Einstein

and Newton. He has given them a clue to release the hundred thousand years of stress and to break down

the plates beneath, powdering them in their own strength without being noticed by the land around! They all

remarked that the logic is simple, but Cranky has stoned Goliath on his forehead, saving the land we all live in.

Had he not done it so well as he did, the growing power those platelets had would have drowned all life in

Japan on the day it might burst. They had indeed succeeded in crunching all of it to dust and grain without even

the sensitive seismometers knowing which is but to the credit of Cranky, the scientist. The scientiﬁc contribution

of Cranky is the following. Usually, when energy is released through plate tectonics, it is transmitted to all

parts of the globe through surrounding media and other platelets. The strategically guided triggering of platelet

breaks made this impossible, and the total energy could be conﬁned to the platelets without escaping. This

energy was spent to break down the platelets, turning them into a pile of sand conﬁned to the same volume.

This is how he managed to release the tension that prevailed for centuries beneath the surface.

About the Author

Professor Ninan Sajeeth Philip is a Visiting Professor at the Inter-University Centre for Astronomy

and Astrophysics (IUCAA), Pune. He is also an Adjunct Professor of AI in Applied Medical Sciences [BCMCH,

Thiruvalla] and a Senior Advisor for the Pune Knowledge Cluster (PKC). He is the Dean and Director of airis4D

and has a teaching experience of 33+ years in Physics. His area of specialisation is AI and ML.

About airis4D

Artiﬁcial Intelligence Research and Intelligent Systems (airis4D) is an AI and Bio-sciences Research Centre.

The Centre aims to create new knowledge in the ﬁeld of Space Science, Astronomy, Robotics, Agri Science,

Industry, and Biodiversity to bring Progress and Plenitude to the People and the Planet.

Vision

Humanity is in the 4th Industrial Revolution era, which operates on a cyber-physical production system. Cutting-

edge research and development in science and technology to create new knowledge and skills become the key to

the new world economy. Most of the resources for this goal can be harnessed by integrating biological systems

with intelligent computing systems oﬀered by AI. The future survival of humans, animals, and the ecosystem

depends on how eﬃciently the realities and resources are responsibly used for abundance and wellness. Artiﬁcial

intelligence Research and Intelligent Systems pursue this vision and look for the best actions that ensure an

abundant environment and ecosystem for the planet and the people.

Mission Statement

The 4D in airis4D represents the mission to Dream, Design, Develop, and Deploy Knowledge with the ﬁre of

commitment and dedication towards humanity and the ecosystem.

Dream

To promote the unlimited human potential to dream the impossible.

Design

To nurture the human capacity to articulate a dream and logically realise it.

Develop

To assist the talents to materialise a design into a product, a service, a knowledge that beneﬁts the community

and the planet.

Deploy

To realise and educate humanity that a knowledge that is not deployed makes no diﬀerence by its absence.

Campus

Situated in a lush green village campus in Thelliyoor, Kerala, India, airis4D was established under the auspicious

of SEED Foundation (Susthiratha, Environment, Education Development Foundation) a not-for-proﬁt company

for promoting Education, Research. Engineering, Biology, Development, etc.

The whole campus is powered by Solar power and has a rain harvesting facility to provide suﬃcient water supply

for up to three months of drought. The computing facility in the campus is accessible from anywhere through a

dedicated optical ﬁbre internet connectivity 24×7.

There is a freshwater stream that originates from the nearby hills and ﬂows through the middle of the campus.

The campus is a noted habitat for the biodiversity of tropical Fauna and Flora. airis4D carry out periodic and

systematic water quality and species diversity surveys in the region to ensure its richness. It is our pride that

the site has consistently been environment-friendly and rich in biodiversity. airis4D is also growing fruit plants

that can feed birds and provide water bodies to survive the drought.