Skip to main content
main-content
Top

About this book

This book constitutes the proceedings of the First International Conference on Modeling, Machine Learning and Astronomy, MMLA 2019, held in Bangalore, India, in November 2019.

The 11 full papers and 3 short papers presented in this volume were carefully reviewed and selected from 63 submissions. They are organized in topical sections on ​modeling and foundations; machine learning applications; astronomy and astroinformatics.

Table of Contents

Frontmatter

Modeling and Foundations

Frontmatter

Optimizing Inter-nationality of Journals: A Classical Gradient Approach Revisited via Swarm Intelligence

Abstract
With the growth of a vast number of new journals, the de facto definitions of Internationality has raised debate across researchers. A robust set of metrics, not prone to manipulation, is paramount for evaluating influence when journals claim “International” status. The ScientoBASE project defines internationality in terms of publication quality and spread of influence beyond geographical boundaries. This is acheived through quantified metrics, like the NLIQ, OCQ, SNIP and ICR, passed into the Cobb Douglas Production Function to estimate the range of influence a journal has over its audience. The global optima of this range is the maximum projected internationality score, or the internationality index of the journal. The optimization, however, being multivariate and constrained presents several challenges to classical techniques, such as curvature variation, premature convergence and parameter scaling. This study approaches these issues by optimizing through the Swarm Intelligence meta-heuristic. Particle Swarm Optimization makes no assumptions on the function being optimized and does away with the need to calculate a gradient. These advantages circumvent the aforementioned issues and highlight the need for traction on machine learning in optimization. The model presented here observes that each journal has an associated globally optimal internationality score that fluctuates proportionally to input metrics, thereby describing a robust confluence of key influence indicators that pave way for investigating alternative criteria for attributing credits to publications.
Luckyson Khaidem, Rahul Yedida, Abhijit J. Theophilus

Stirling Numbers via Combinatorial Sums

Abstract
In this paper, we have derived a formula to find combinatorial sums of the type \( \displaystyle {\sum _{r=0}^{n} {r^{k}} {n \atopwithdelims ()r}}\) where \(k \in \mathbb {N}\). The formula is conveniently expressed as a sum of terms multiplied by certain co-efficients. These co-efficients satisfy a recurrence relation, which is also derived in the process of finding the above sum. Upon solving the recurrence, these numbers turn out to be the Stirling Numbers of the first and second kind. Here on, it is trivial to prove the mutual inverse property of both these sequences of numbers due to linear algebra.
Anwesh Bhattacharya, Bivas Bhattacharya

Random Subspace Combined LDA Based Machine Learning Model for OSCC Classifier

Abstract
Oral squamous cell carcinoma (OSCC) remains a major death causing oral cancer in developing countries. In recent years, tremendous development in medical imaging devices made microscopic colour images of biopsy samples available to the researchers. Image processing and machine learning techniques can be used to develop automatic cancer grading mechanism. In this work, automatic OSCC classifier using Linear Discriminant Analysis combined with Random Subspace is developed and analyzed. The proposed classifiers automatically classifies the input image in one of the four categories, namely: Normal, Grade-I, II or III. Total 83 colour and texture features are computed from the 100 Haemotoxylin and Eosin (H&E) stained images of oral mucosa. The overall accuracy of the proposed classifier is 93.5% with sensitivity and specificity of 0.89 and 0.95 respectively.
Archana Nawandhar, Navin Kumar, Lakshmi Yamujala

Automatic Annotation of Deceptive Online Reviews Using Topic Modelling

Abstract
The increasing popularity of e-commerce websites and online review platforms has unfortunately led to the advent of review spammers. This has, in turn, led to many problems, both in business and in academia. One of the major challenges in this field is the annotation of deceptive reviews. To date, different approaches have been employed in the creation of a labelled dataset for classification tasks. Many of these works follow a general approach and do not focus on any particular property of deceptive reviews. We believe that a fine-grained approach would be more suitable for such a complex problem. This paper focuses on a single property of deceptive reviews; the out-of-context property. We first find the minimum length of review required for obtaining coherent topics. We then propose a methodology for scoring and labelling the reviews and evaluate it by training different classifiers. We obtain an F-measure of 93.64 using labelled reviews obtained through the proposed methodology.
R. N. Pramukha, P. S. Venugopala

Machine Learning Applications

Frontmatter

Machine Learning Technique for Analyzing the Behavior of Fish in an Aquarium

Abstract
Over the years lots of marine biologists and scientists have made efforts to study the various problem related to the ocean and its animals. One such problem faced by the fisheries is the change in the physical and mental state of the fishes when they are affected by natural calamities, external agents, change in environment, epidemic outbreak or many such factors. Under such circumstances it’s probable that the fishes in the region may get affected and due to a small amount of bad fish the whole population could suffer which is a huge loss both biologically and economically. There are numerous research and studies back on the fact that fishes tend to change their behaviour in the situation of distress. The behavioural change is just in response to the distress caused to them either physically or psychologically. The motion of the fish over a period of time interval can help us learn about the behaviour it is showing. And a significant change in its pattern of motion can alarm us that there is some issue with the fish. The main idea is to identify and establish the relationship between the movements and behaviour of the fish. The video footage of a Tilapia Genus fish in a standard size aquarium set up with the help of surveillance cameras, tracked the motion of the fish in 2-D space is collected and the movement of the fish over a period of time is categorised. Based on the output the losses can be reduced.
Rishabh Bhaskaran, Rajesh Kanna Baskaran, C. Vijayalakshmi

Faster Convergence to N-Queens Problem Using Reinforcement Learning

Abstract
Algorithmic complexity has been a constraint to solving problems efficiently. Wide use of an algorithm is dependent on its space and time complexity for large inputs. Exploiting an inherent pattern to solve a problem could be easy compared to an algorithm-based approach. Such patterns are quite necessary at cracking games with a vast number of possibilities as an algorithm-based approach would be computationally expensive and time-consuming. The N-Queens problem is one such problem with many possible configurations and realizing a solution to this is hard as the value of N increases. Reinforcement Learning has proven to be good at building an agent that can learn these hidden patterns over time to converge to a solution faster. This study shows how reinforcement learning can outperform traditional algorithms in solving the N-Queens problem.
Patnala Prudhvi Raj, Preet Shah, Pragnya Suresh

Classification of Corpus Callosum Layer in Mid-saggital MRI Images Using Machine Learning Techniques for Autism Disorder

Abstract
Autism is a neuro developmental disorder that affects the social interaction and communication skills of the children. It is characterized by repetitive behavior, lack of eye contact and unusual facial expressions. Corpus Callosum (CC) is the largest white matter area in the central nervous system that helps in transmission of information between both the hemispheres of brain. In autism kids, CC in the brain region shrinks and shape variations occur, making it as the region of interest with respect to diagnosis of autism disorder. Though there are many methods to segment and classify CC, there is still a need for accurate segmentation and automatic classification of CC. Since CC shares similar intensity and close proximity to other parts of the brain, segmentation of only CC region becomes challenging. To address this challenge, in the proposed work level set segmentation technique is used to segment Corpus callosum and the segmented images are validated against the ground truth using jaccard and dice index. From the segmented images geometric, texture and statistical features are extracted. Feature reduction methods such as Principal Component Analysis (PCA) and Independent Component Analysis (ICA) are incorporated for selecting the most significant set of features. Machine learning algorithms such as Support vector machine (SVM) and Extreme learning machine (ELM) are proposed to classify the image as normal and abnormal. The proposed algorithm demonstrates the classification accuracy of 97% and 96.5% using SVM and ELM respectively.
A. Ramanathan, T. Christy Bobby

Dynamic Systems Simulation and Control Using Consecutive Recurrent Neural Networks

Abstract
In this paper, we introduce a novel architecture to connecting adaptive learning and neural networks into an arbitrary machine’s control system paradigm. Two consecutive Recurrent Neural Networks (RNNs) are used together to accurately model the dynamic characteristics of electromechanical systems that include controllers, actuators and motors. The age-old method of achieving control with the use of the – Proportional, Integral and Derivative constants is well understood as a simplified method that does not capture the complexities of the inherent nonlinearities of complex control systems. In the context of controlling and simulating electromechanical systems, we propose an alternative to PID Controllers, employing a sequence of two Recurrent Neural Networks. The first RNN emulates the behavior of the controller, and the second the actuator/motor. The second RNN, when used in isolation potentially serves as an advantageous alternative to extant testing methods of electro-mechanical systems.
Srikanth Chandar, Harsha Sunder

Enhanced Hybrid Segmentation with Non Local Block and Deep Residual Networks

Abstract
The use of convolutional neural networks (CNNs) has increased in the edge devices due to its successful performance. Various such applications includes semantic segmentation which is one of the most challenging tasks due to the involvement of tremendous model size and parameters. In this paper, an enhanced hybrid segmentation with non-local block and deep residual networks is introduced for pixel level semantic segmentation. A light weight model is developed to facilitate deployment on edge devices. Skip connection is applied to fire layer in encoder and decoder block of segmentation model and non-local block is inserted in between encoder and decoder. Due to such amendments, the proposed network has optimized the number of parameters to only 2 Million whereas SegNet-Basic architecture required 5 Million. The performance is validated on the Camvid dataset and 87.5% accuracy is achieved.
Yuvaram Singh, Guda Ramachandra Kaladhara Sarma, Kameshwar Rao

Astronomy and AstroInformatics

Frontmatter

Chaotic Quantum Behaved Particle Swarm Optimization for Multiobjective Optimization in Habitability Studies

Abstract
In this paper, based on the Quantum-behaved Particle Swarm Optimization algorithm in [13], we evolve the algorithm to optimize a multiobjective optimization problem, namely the Cobb Douglas Habitability function which is based on “CES production functions” in Economics. We also propose some changes to the Quantum-behaved Particle Swarm Optimization algorithm to mitigate the problem of the algorithm prematurely converging and show the results of the proposed changes to the Quantum-behaved Particle Swarm Optimization.
Arun John, Anish Murthy

The System of Open Star Clusters Revisited

Abstract
The system of open clusters is an excellent probe of the structure and evolution of the galactic disk. Their spatial, size, age and mass distributions provide valuable information on the cluster formation process. Present day astronomy is rich in data, and hence in this work, we attempt to build up a comprehensive statistical study of star clusters. This study is based on available catalogues, both homogeneous and inhomogeneous, to provide some useful insights on the evolutionary history of the system of open clusters and consequently, the galaxy. We find that the optimum size of a cluster for its survival is 3–4 pc. We also find that there exists a simple linear relationship between the age and the mean linear diameters of clusters and also with normalised reddening. Using the catalogues based on Gaia DR2 and other catalogs, we find, that the system of open clusters provides valuable clues to our understanding of the evolution of the galaxy. This system can be partitioned by k-means to get clusters in a statistical sense, which indicates possible cluster formation in the galaxy at different galactocentric distances and with different parameters. These suggests a combination of the scenarios of overall halo collapse and accretion to explain the formation of the disk of the galaxy. This method is proposed to be used for the study of external galaxies using catalogues of extragalactic clusters as it works well with the clusters of the Milky Way.
Priya Hasan, S. N. Hasan

Genetic Bi-objective Optimization Approach to Habitability Score

Abstract
The search for life outside the Solar System is an endeavour of astronomers all around the world. With hundreds of exoplanets being discovered due to advances in astronomy, there is a need to classify the habitability of these exoplanets. This is typically done using various metrics such as the Earth Similarity Index or the Planetary Habitability Index. In this paper, Genetic Algorithms are used to evaluate the best possible habitability scores using the Cobb-Douglas Habitability Score. Genetic Algorithm is a classic evolutionary algorithm used for solving optimization problems. The working of the algorithm is established through comparison with various benchmark functions and its functionality is extended to Multi-Objective optimization. The Cobb-Douglas Habitability Function is formulated as a bi-objective as well as a single objective optimization problem to find the optimal values to maximize the Cobb-Douglas Habitability Score for a set of promising exoplanets.
Sriram Krishna, Niharika Pentapati

Machine Learning Based Analysis of Gravitational Waves

Abstract
Gravitational waves has been a serious subject of study in the modern day astrophysics. Where on one end the strain produced by gravitational waves on matter could be practically studied by Laser Interferometers such as LIGO, the strain generated by celestial bodies on the other end a priori obtained by numerical relativity in the form of waveforms. It is often the case that these waveforms are only used to study the properties of black holes. This article tries to extrapolate such methodologies to weaker celestial bodies for the primary purpose of adding a new dimensionality in the prudent realm of possibilities. There is a necessity to approach such studies from a statistical perspective. Utilizing the combination of Statistical and Machine Learning tools not only assist in analyzing data effectively but also aid in creating a generalized computational model.
Surbhi Agrawal, Rahul Aedula, D. S. Rahul Surya

Thermal Suitability Scheme: Habitability Classification of Exoplanets

Abstract
In this paper, a metric for estimating the potential habitability of exoplanets, called thermal suitability score (TSS) is developed based on machine learning (ML). As compared to prior literature, the TSS ascertains habitability by using a sign – positive for potentially habitable, and negative for non-habitable – and a number indicating the extent to which an exoplanet is habitable or non-habitable. The TSS is used on the data provided in the University of Puerto Rico’s Planetary Habitability Laboratory’s Exoplanets Catalog (PHL-EC).
Suryoday Basak

Backmatter

Additional information

Premium Partner

    Image Credits