Showing 1–50 of 116 results for author: Roychowdhury, S

A Compass for Navigating the World of Sentence Embeddings for the Telecom Domain

Authors: Sujoy Roychowdhury, Sumit Soman, H. G. Ranjani, Vansh Chhabra, Neeraj Gunda, Subhadip Bandyopadhyay, Sai Krishna Bala

Abstract: A plethora of sentence embedding models makes it challenging to choose one, especially for domains such as telecom, rich with specialized vocabulary. We evaluate multiple embeddings obtained from publicly available models and their domain-adapted variants, on both point retrieval accuracies as well as their (95\%) confidence intervals. We establish a systematic method to obtain thresholds for simi… ▽ More A plethora of sentence embedding models makes it challenging to choose one, especially for domains such as telecom, rich with specialized vocabulary. We evaluate multiple embeddings obtained from publicly available models and their domain-adapted variants, on both point retrieval accuracies as well as their (95\%) confidence intervals. We establish a systematic method to obtain thresholds for similarity scores for different embeddings. We observe that fine-tuning improves mean bootstrapped accuracies as well as tightens confidence intervals. The pre-training combined with fine-tuning makes confidence intervals even tighter. To understand these variations, we analyse and report significant correlations between the distributional overlap between top-$K$, correct and random sentence similarities with retrieval accuracies and similarity thresholds. Following current literature, we analyze if retrieval accuracy variations can be attributed to isotropy of embeddings. Our conclusions are that isotropy of embeddings (as measured by two independent state-of-the-art isotropy metric definitions) cannot be attributed to better retrieval performance. However, domain adaptation which improves retrieval accuracies also improves isotropy. We establish that domain adaptation moves domain specific embeddings further away from general domain embeddings. △ Less

Submitted 18 June, 2024; originally announced June 2024.

Comments: 10 pages, 3 figures, 4 tables

MSC Class: 68T50 ACM Class: I.2.7

The Smallest Scale of Hierarchy Survey (SSH) III. Dwarf-dwarf satellite merging phenomena in the low-mass regime

Authors: Elena Sacchi, Michele Bellazzini, Francesca Annibali, Monica Tosi, Giacomo Beccari, John M. Cannon, Laura C. Hunter, Diego Paris, Sambit Roychowdhury, Lila Schisgal, Liese van Zee, Michele Cignoni, Felice Cusano, Roelof S. de Jong, Leslie Hunt, Raffaele Pascale

Abstract: We present new deep, wide-field Large Binocular Telescope (LBT) $g$ and $r$ imaging data from the Smallest Scale of Hierarchy Survey (SSH) revealing previously undetected tidal features and stellar streams in the outskirts of six dwarf irregular galaxies (NGC 5238, UGC 6456, UGC 6541, UGC 7605, UGC 8638, and UGC 8760) with stellar masses in the range $1.2 \times 10^7$ M$_{\odot}$ to… ▽ More We present new deep, wide-field Large Binocular Telescope (LBT) $g$ and $r$ imaging data from the Smallest Scale of Hierarchy Survey (SSH) revealing previously undetected tidal features and stellar streams in the outskirts of six dwarf irregular galaxies (NGC 5238, UGC 6456, UGC 6541, UGC 7605, UGC 8638, and UGC 8760) with stellar masses in the range $1.2 \times 10^7$ M$_{\odot}$ to $1.4 \times 10^8$ M$_{\odot}$. The six dwarfs are located 1-2 Mpc away from large galaxies, implying that the observed distortions are unlikely to be due to tidal effects from a nearby, massive companion. At the dwarfs' distances of $\sim$3-4 Mpc, the identified tidal features are all resolved into individual stars in the LBT images and appear to be made of a population older than 1-2 Gyr, excluding the possibility that they result from irregular and asymmetric star formation episodes that are common in gas-rich dwarf galaxies. The most plausible explanation is that we are witnessing the hierarchical merging assembling of these dwarfs with their satellite populations, a scenario also supported by the peculiar morphology and disturbed velocity field of their HI component. From the SSH sample we estimate a fraction of late type dwarfs showing signs of merging with satellites of $\sim$13\%, in agreement with other recent independent studies and theoretical predictions within the $Λ$CDM cosmological framework. △ Less

Submitted 3 June, 2024; originally announced June 2024.

Comments: 12 pages, 11 figures including one in the appendix, accepted for publication by A&A

Revealing the hidden Dirac gap in a topological antiferromagnet using Floquet-Bloch manipulation

Authors: Nina Bielinski, Rajas Chari, Julian May-Mann, Soyeun Kim, Jack Zwettler, Yujun Deng, Anuva Aishwarya, Subhajit Roychowdhury, Chandra Shekhar, Makoto Hashimoto, Donghui Lu, Jiaqiang Yan, Claudia Felser, Vidya Madhavan, Zhi-Xun Shen, Taylor L. Hughes, Fahad Mahmood

Abstract: Manipulating solids using the time-periodic drive of a laser pulse is a promising route to generate new phases of matter. Whether such `Floquet-Bloch' manipulation can be achieved in topological magnetic systems with disorder has so far been unclear. In this work, we realize Floquet-Bloch manipulation of the Dirac surface-state mass of the topological antiferromagnet (AFM) MnBi$_2$Te$_4$. Using ti… ▽ More Manipulating solids using the time-periodic drive of a laser pulse is a promising route to generate new phases of matter. Whether such `Floquet-Bloch' manipulation can be achieved in topological magnetic systems with disorder has so far been unclear. In this work, we realize Floquet-Bloch manipulation of the Dirac surface-state mass of the topological antiferromagnet (AFM) MnBi$_2$Te$_4$. Using time- and angle-resolved photoemission spectroscopy (tr-ARPES), we show that opposite helicities of mid-infrared circularly polarized light result in substantially different Dirac mass gaps in the AFM phase, despite the equilibrium Dirac cone being massless. We explain our findings in terms of a Dirac fermion with a random mass. Our results underscore Floquet-Bloch manipulation as a powerful tool for controlling topology even in the presence of disorder, and for uncovering properties of materials that may elude conventional probes. △ Less

Submitted 26 May, 2024; originally announced May 2024.

Exploring Ordinality in Text Classification: A Comparative Study of Explicit and Implicit Techniques

Authors: Siva Rajesh Kasa, Aniket Goel, Karan Gupta, Sumegh Roychowdhury, Anish Bhanushali, Nikhil Pattisapu, Prasanna Srinivasa Murthy

Abstract: Ordinal Classification (OC) is a widely encountered challenge in Natural Language Processing (NLP), with applications in various domains such as sentiment analysis, rating prediction, and more. Previous approaches to tackle OC have primarily focused on modifying existing or creating novel loss functions that \textbf{explicitly} account for the ordinal nature of labels. However, with the advent of… ▽ More Ordinal Classification (OC) is a widely encountered challenge in Natural Language Processing (NLP), with applications in various domains such as sentiment analysis, rating prediction, and more. Previous approaches to tackle OC have primarily focused on modifying existing or creating novel loss functions that \textbf{explicitly} account for the ordinal nature of labels. However, with the advent of Pretrained Language Models (PLMs), it became possible to tackle ordinality through the \textbf{implicit} semantics of the labels as well. This paper provides a comprehensive theoretical and empirical examination of both these approaches. Furthermore, we also offer strategic recommendations regarding the most effective approach to adopt based on specific settings. △ Less

Submitted 20 May, 2024; originally announced May 2024.

Comments: Findings of ACL 2024

ERATTA: Extreme RAG for Table To Answers with Large Language Models

Authors: Sohini Roychowdhury, Marko Krema, Anvar Mahammad, Brian Moore, Arijit Mukherjee, Punit Prakashchandra

Abstract: Large language models (LLMs) with retrieval augmented-generation (RAG) have been the optimal choice for scalable generative AI solutions in the recent past. However, the choice of use-cases that incorporate RAG with LLMs have been either generic or extremely domain specific, thereby questioning the scalability and generalizability of RAG-LLM approaches. In this work, we propose a unique LLM-based… ▽ More Large language models (LLMs) with retrieval augmented-generation (RAG) have been the optimal choice for scalable generative AI solutions in the recent past. However, the choice of use-cases that incorporate RAG with LLMs have been either generic or extremely domain specific, thereby questioning the scalability and generalizability of RAG-LLM approaches. In this work, we propose a unique LLM-based system where multiple LLMs can be invoked to enable data authentication, user query routing, data retrieval and custom prompting for question answering capabilities from data tables that are highly varying and large in size. Our system is tuned to extract information from Enterprise-level data products and furnish real time responses under 10 seconds. One prompt manages user-to-data authentication followed by three prompts to route, fetch data and generate a customizable prompt natural language responses. Additionally, we propose a five metric scoring module that detects and reports hallucinations in the LLM responses. Our proposed system and scoring metrics achieve >90% confidence scores across hundreds of user queries in the sustainability, financial health and social media domains. Extensions to the proposed extreme RAG architectures can enable heterogeneous source querying using LLMs. △ Less

Submitted 14 May, 2024; v1 submitted 6 May, 2024; originally announced May 2024.

Comments: 5 pages, 3 tables, Asilomar SSC Conference, 2024

DEVILS/MIGHTEE/GAMA/DINGO: The Impact of SFR Timescales on the SFR-Radio Luminosity Correlation

Authors: Robin H. W. Cook, Luke J. M. Davies, Jonghwan Rhee, Catherine L. Hale, Sabine Bellstedt, Jessica E. Thorne, Ivan Delvecchio, Jordan D. Collier, Richard Dodson, Simon P. Driver, Benne W. Holwerda, Matt J. Jarvis, Kenda Knowles, Claudia Lagos, Natasha Maddox, Martin Meyer, Aaron S. G. Robotham, Sambit Roychowdhury, Kristof Rozgonyi, Nicholas Seymour, Malgorzata Siudek, Matthew Whiting, Imogen Whittam

Abstract: The tight relationship between infrared luminosity (L$_\mathrm{TIR}$) and 1.4 GHz radio continuum luminosity (L$_\mathrm{1.4GHz}$) has proven useful for understanding star formation free from dust obscuration. Infrared emission in star-forming galaxies typically arises from recently formed, dust-enshrouded stars, whereas radio synchrotron emission is expected from subsequent supernovae. By leverag… ▽ More The tight relationship between infrared luminosity (L$_\mathrm{TIR}$) and 1.4 GHz radio continuum luminosity (L$_\mathrm{1.4GHz}$) has proven useful for understanding star formation free from dust obscuration. Infrared emission in star-forming galaxies typically arises from recently formed, dust-enshrouded stars, whereas radio synchrotron emission is expected from subsequent supernovae. By leveraging the wealth of ancillary far-ultraviolet - far-infrared photometry from the Deep Extragalactic VIsible Legacy Survey (DEVILS) and Galaxy and Mass Assembly (GAMA) surveys, combined with 1.4 GHz observations from the MeerKAT International GHz Tiered Extragalactic Exploration (MIGHTEE) survey and Deep Investigation of Neutral Gas Origins (DINGO) projects, we investigate the impact of timescale differences between far-ultraviolet - far-infrared and radio-derived star formation rate (SFR) tracers. We examine how the SED-derived star formation histories (SFH) of galaxies can be used to explain discrepancies in these SFR tracers, which are sensitive to different timescales. Galaxies exhibiting an increasing SFH have systematically higher L$_\mathrm{TIR}$ and SED-derived SFRs than predicted from their 1.4 GHz radio luminosity. This indicates that insufficient time has passed for subsequent supernovae-driven radio emission to accumulate. We show that backtracking the SFR(t) of galaxies along their SED-derived SFHs to a time several hundred megayears prior to their observed epoch will both linearise the SFR-L$_\mathrm{1.4GHz}$ relation and reduce the overall scatter. The minimum scatter in the SFR(t)-L$_\mathrm{1.4GHz}$ is reached at 200 - 300 Myr prior, consistent with theoretical predictions for the timescales required to disperse the cosmic ray electrons responsible for the synchrotron emission. △ Less

Submitted 1 May, 2024; originally announced May 2024.

Comments: 23 pages, 13 figures, 2 tables. Accepted for publication in MNRAS

Observations on Building RAG Systems for Technical Documents

Authors: Sumit Soman, Sujoy Roychowdhury

Abstract: Retrieval augmented generation (RAG) for technical documents creates challenges as embeddings do not often capture domain information. We review prior art for important factors affecting RAG and perform experiments to highlight best practices and potential challenges to build RAG systems for technical documents. Retrieval augmented generation (RAG) for technical documents creates challenges as embeddings do not often capture domain information. We review prior art for important factors affecting RAG and perform experiments to highlight best practices and potential challenges to build RAG systems for technical documents. △ Less

Submitted 31 March, 2024; originally announced April 2024.

Comments: Published as a Tiny Paper at ICLR 2024

ACM Class: I.2.7

Observation of Chiral Surface State in Superconducting NbGe$_2$

Authors: Mengyu Yao, Martin Gutierrez-Amigo, Subhajit Roychowdhury, Ion Errea, Alexander Fedorov, Vladimir N. Strocov, Maia G. Vergniory, Claudia Felser

Abstract: The interplay between topology and superconductivity in quantum materials harbors rich physics ripe for discovery. In this study, we investigate the topological properties and superconductivity of the nonsymmorphic chiral superconductor NbGe$_2$ using high-resolution angle-resolved pho-toemission spectroscopy (ARPES), transport measurements, and ab initio calculations. The ARPES data revealed exot… ▽ More The interplay between topology and superconductivity in quantum materials harbors rich physics ripe for discovery. In this study, we investigate the topological properties and superconductivity of the nonsymmorphic chiral superconductor NbGe$_2$ using high-resolution angle-resolved pho-toemission spectroscopy (ARPES), transport measurements, and ab initio calculations. The ARPES data revealed exotic chiral surface states on the (100) surface originating from the inherent chiral crystal structure. Supporting calculations indicate that NbGe$_2$ likely hosts elusive Weyl fermions in its bulk electronic structure. Furthermore, we uncovered the signatures of van Hove singularities that can enhance many-body interactions. Additionally, transport measurements demonstrated that NbGe$_2$ exhibits superconductivity below 2K. Overall, our comprehensive results provide the first concrete evidence that NbGe$_2$ is a promising platform for investigating the interplay between non-trivial band topology, possible Weyl fermions, van Hove singularities, and superconductivity in chiral quantum materials. △ Less

Submitted 4 April, 2024; v1 submitted 5 March, 2024; originally announced March 2024.

MIGHTEE-HI: HI galaxy properties in the large scale structure environment at z~0.37 from a stacking experiment

Authors: Francesco Sinigaglia, Giulia Rodighiero, Ed Elson, Alessandro Bianchetti, Mattia Vaccari, Natasha Maddox, Anastasia A. Ponomareva, Bradley S. Frank, Matt J. Jarvis, Barbara Catinella, Luca Cortese, Sambit Roychowdhury, Maarten Baes, Jordan D. Collier, Olivier Ilbert, Ali A. Khostovan, Sushma Kurapati, Hengxing Pan, Isabella Prandoni, Sambatriniaina H. A. Rajohnson, Mara Salvato, Srikrishna Sekhar, Gauri Sharma

Abstract: We present the first measurement of HI mass of star-forming galaxies in different large scale structure environments from a blind survey at $z\sim 0.37$. In particular, we carry out a spectral line stacking analysis considering $2875$ spectra of colour-selected star-forming galaxies undetected in HI at $0.23 < z < 0.49$ in the COSMOS field, extracted from the MIGHTEE-HI Early Science datacubes, ac… ▽ More We present the first measurement of HI mass of star-forming galaxies in different large scale structure environments from a blind survey at $z\sim 0.37$. In particular, we carry out a spectral line stacking analysis considering $2875$ spectra of colour-selected star-forming galaxies undetected in HI at $0.23 < z < 0.49$ in the COSMOS field, extracted from the MIGHTEE-HI Early Science datacubes, acquired with the MeerKAT radio telescope. We stack galaxies belonging to different subsamples depending on three different definitions of large scale structure environment: local galaxy overdensity, position inside the host dark matter halo (central, satellite, or isolated), and cosmic web type (field, filament, or knot). We first stack the full star-forming galaxy sample and find a robust HI detection yielding an average galaxy HI mass of $M_{\rm HI}=(8.12\pm 0.75)\times 10^9\, {\rm M}_\odot$ at $\sim 11.8σ$. Next, we investigate the different subsamples finding a negligible difference in $M_{\rm HI}$ as a function of the galaxy overdensity. We report an HI excess compared to the full sample in satellite galaxies ($M_{\rm HI}=(11.31\pm1.22)\times 10^9$, at $\sim 10.2 σ$) and in filaments ($M_{\rm HI}=(11.62\pm 0.90)\times 10^9$. Conversely, we report non-detections for the central and knot galaxies subsamples, which appear to be HI-deficient. We find the same qualitative results also when stacking in units of HI fraction ($f_{\rm HI}$). We conclude that the HI amount in star-forming galaxies at the studied redshifts correlates with the large scale structure environment. △ Less

Submitted 1 March, 2024; originally announced March 2024.

Comments: Accepted for publication in MNRAS. 15 figures, 3 tables

Universal Translational and Rotational Mobility Expressions of Phoretic and Self-phoretic Particles with Arbitrary Interaction Potentials

Authors: Arkava Ganguly, Souradeep Roychowdhury, Ankur Gupta

Abstract: The mobility of externally-driven phoretic propulsion of particles is evaluated by simultaneously solving the solute conservation equation, interaction potential equation, and the modified Stokes equation. While accurate, this approach is cumbersome, especially when the interaction potential decays slowly compared to the particle size. In contrast to external phoresis, the motion of self-phoretic… ▽ More The mobility of externally-driven phoretic propulsion of particles is evaluated by simultaneously solving the solute conservation equation, interaction potential equation, and the modified Stokes equation. While accurate, this approach is cumbersome, especially when the interaction potential decays slowly compared to the particle size. In contrast to external phoresis, the motion of self-phoretic particles is typically estimated by relating the translation and rotation velocities with the local slip velocity. While this approach is convenient and thus widely used, it is only valid when the interaction decay length is significantly smaller than the particle size. Here, by employing the Lorentz reciprocal theorem, we combine the benefits of two approaches and derive unified mobility expressions with arbitrary interaction potentials such that the expressions predict the translation and rotation velocities for both externally driven and self-propelling particles. We show that these expressions can conveniently recover the well-known mobility relationships of external electrophoresis and diffusiophoresis for arbitrary double-layer thickness. Additionally, we show that for a spherical microswimmer, our derived expressions relax to the slip velocity calculations in the limit of thin interaction lengthscales. We also employ the derived mobility expressions to calculate the velocities of an autophoretic Janus particle. We find that there is significant dampening in the translation velocity even when the interaction length is an order of magnitude larger than the particle size. Finally, we study the motion of a catalytically self-propelled particle, while it also propels due to external concentration gradients, and demonstrate how the two propulsion modes compete with each other. △ Less

Submitted 29 February, 2024; originally announced February 2024.

Untangle charge-order dependent bulk states from surface effects in a topological kagome metal ScV$_6$Sn$_6$

Authors: Zi-Jia Cheng, Sen Shao, Byunghoon Kim, Tyler A. Cochran, Xian P. Yang, Changjiang Yi, Yu-Xiao Jiang, Junyi Zhang, Md Shafayat Hossain, Subhajit Roychowdhury, Turgut Yilmaz, Elio Vescovo, Alexei Fedorov, Shekhar Chandra, Claudia Felser, Guoqing Chang, M. Zahid Hasan

Abstract: Kagome metals with charge density wave (CDW) order exhibit a broad spectrum of intriguing quantum phenomena. The recent discovery of the novel kagome CDW compound ScV$_6$Sn$_6$ has spurred significant interest. However, understanding the interplay between CDW and the bulk electronic structure has been obscured by a profusion of surface states and terminations in this quantum material. Here, we emp… ▽ More Kagome metals with charge density wave (CDW) order exhibit a broad spectrum of intriguing quantum phenomena. The recent discovery of the novel kagome CDW compound ScV$_6$Sn$_6$ has spurred significant interest. However, understanding the interplay between CDW and the bulk electronic structure has been obscured by a profusion of surface states and terminations in this quantum material. Here, we employ photoemission spectroscopy and potassium dosing to elucidate the complete bulk band structure of ScV$_6$Sn$_6$, revealing multiple van Hove singularities near the Fermi level. We surprisingly discover a robust spin-polarized topological Dirac surface resonance state at the M point within the two-fold van Hove singularities. Assisted by the first-principle calculations, the temperature dependence of the $k_z$- resolved ARPES spectrum provides unequivocal evidence for the proposed $\sqrt{3}$$\times$$\sqrt{3}$$\times3$ charge order over other candidates. Our work not only enhances the understanding of the CDW-dependent bulk and surface states in ScV$_6$Sn$_6$ but also establishes an essential foundation for potential manipulation of the CDW order in kagome materials. △ Less

Submitted 3 February, 2024; originally announced February 2024.

Comments: To appear in PRB

How Robust are LLMs to In-Context Majority Label Bias?

Authors: Karan Gupta, Sumegh Roychowdhury, Siva Rajesh Kasa, Santhosh Kumar Kasa, Anish Bhanushali, Nikhil Pattisapu, Prasanna Srinivasa Murthy

Abstract: In the In-Context Learning (ICL) setup, various forms of label biases can manifest. One such manifestation is majority label bias, which arises when the distribution of labeled examples in the in-context samples is skewed towards one or more specific classes making Large Language Models (LLMs) more prone to predict those labels. Such discrepancies can arise from various factors, including logistic… ▽ More In the In-Context Learning (ICL) setup, various forms of label biases can manifest. One such manifestation is majority label bias, which arises when the distribution of labeled examples in the in-context samples is skewed towards one or more specific classes making Large Language Models (LLMs) more prone to predict those labels. Such discrepancies can arise from various factors, including logistical constraints, inherent biases in data collection methods, limited access to diverse data sources, etc. which are unavoidable in a real-world industry setup. In this work, we study the robustness of in-context learning in LLMs to shifts that occur due to majority label bias within the purview of text classification tasks. Prior works have shown that in-context learning with LLMs is susceptible to such biases. In our study, we go one level deeper and show that the robustness boundary varies widely for different models and tasks, with certain LLMs being highly robust (~90%) to majority label bias. Additionally, our findings also highlight the impact of model size and the richness of instructional prompts contributing towards model robustness. We restrict our study to only publicly available open-source models to ensure transparency and reproducibility. △ Less

Submitted 27 December, 2023; originally announced December 2023.

Comments: 6 pages, 3 figures, 2 table. Accepted at Workshop on Responsible Language Modeling, AAAI 2024, (www.aaai.org)

Journey of Hallucination-minimized Generative AI Solutions for Financial Decision Makers

Authors: Sohini Roychowdhury

Abstract: Generative AI has significantly reduced the entry barrier to the domain of AI owing to the ease of use and core capabilities of automation, translation, and intelligent actions in our day to day lives. Currently, Large language models (LLMs) that power such chatbots are being utilized primarily for their automation capabilities for software monitoring, report generation etc. and for specific perso… ▽ More Generative AI has significantly reduced the entry barrier to the domain of AI owing to the ease of use and core capabilities of automation, translation, and intelligent actions in our day to day lives. Currently, Large language models (LLMs) that power such chatbots are being utilized primarily for their automation capabilities for software monitoring, report generation etc. and for specific personalized question answering capabilities, on a limited scope and scale. One major limitation of the currently evolving family of LLMs is 'hallucinations', wherein inaccurate responses are reported as factual. Hallucinations are primarily caused by biased training data, ambiguous prompts and inaccurate LLM parameters, and they majorly occur while combining mathematical facts with language-based context. Thus, monitoring and controlling for hallucinations becomes necessary when designing solutions that are meant for decision makers. In this work we present the three major stages in the journey of designing hallucination-minimized LLM-based solutions that are specialized for the decision makers of the financial domain, namely: prototyping, scaling and LLM evolution using human feedback. These three stages and the novel data to answer generation modules presented in this work are necessary to ensure that the Generative AI chatbots, autonomous reports and alerts are reliable and high-quality to aid key decision-making processes. △ Less

Submitted 17 November, 2023; originally announced November 2023.

Comments: 4 pages, 2 Figures

Gender-Based Comparative Study of Type 2 Diabetes Risk Factors in Kolkata, India: A Machine Learning Approach

Authors: Rahul Jain, Anoushka Saha, Gourav Daga, Durba Bhattacharya, Madhura Das Gupta, Sourav Chowdhury, Suparna Roychowdhury

Abstract: Type 2 diabetes mellitus represents a prevalent and widespread global health concern, necessitating a comprehensive assessment of its risk factors. This study aimed towards learning whether there is any differential impact of age, Lifestyle, BMI and Waist to height ratio on the risk of Type 2 diabetes mellitus in males and females in Kolkata, West Bengal, India based on a sample observed from the… ▽ More Type 2 diabetes mellitus represents a prevalent and widespread global health concern, necessitating a comprehensive assessment of its risk factors. This study aimed towards learning whether there is any differential impact of age, Lifestyle, BMI and Waist to height ratio on the risk of Type 2 diabetes mellitus in males and females in Kolkata, West Bengal, India based on a sample observed from the out-patient consultation department of Belle Vue Clinic in Kolkata. Various machine learning models like Logistic Regression, Random Forest, and Support Vector Classifier, were used to predict the risk of diabetes, and performance was compared based on different predictors. Our findings indicate a significant age-related increase in risk of diabetes for both males and females. Although exercising and BMI was found to have significant impact on the risk of Type 2 diabetes in males, in females both turned out to be statistically insignificant. For both males and females, predictive models based on WhtR demonstrated superior performance in risk assessment compared to those based on BMI. This study sheds light on the gender-specific differences in the risk factors for Type 2 diabetes, offering valuable insights that can be used towards more targeted healthcare interventions and public health strategies. △ Less

Submitted 14 October, 2023; originally announced November 2023.

Comments: 10 pages, 7 tables,3 figures, submitted to a conference

Hallucination-minimized Data-to-answer Framework for Financial Decision-makers

Authors: Sohini Roychowdhury, Andres Alvarez, Brian Moore, Marko Krema, Maria Paz Gelpi, Federico Martin Rodriguez, Angel Rodriguez, Jose Ramon Cabrejas, Pablo Martinez Serrano, Punit Agrawal, Arijit Mukherjee

Abstract: Large Language Models (LLMs) have been applied to build several automation and personalized question-answering prototypes so far. However, scaling such prototypes to robust products with minimized hallucinations or fake responses still remains an open challenge, especially in niche data-table heavy domains such as financial decision making. In this work, we present a novel Langchain-based framewor… ▽ More Large Language Models (LLMs) have been applied to build several automation and personalized question-answering prototypes so far. However, scaling such prototypes to robust products with minimized hallucinations or fake responses still remains an open challenge, especially in niche data-table heavy domains such as financial decision making. In this work, we present a novel Langchain-based framework that transforms data tables into hierarchical textual data chunks to enable a wide variety of actionable question answering. First, the user-queries are classified by intention followed by automated retrieval of the most relevant data chunks to generate customized LLM prompts per query. Next, the custom prompts and their responses undergo multi-metric scoring to assess for hallucinations and response confidence. The proposed system is optimized with user-query intention classification, advanced prompting, data scaling capabilities and it achieves over 90% confidence scores for a variety of user-queries responses ranging from {What, Where, Why, How, predict, trend, anomalies, exceptions} that are crucial for financial decision making applications. The proposed data to answers framework can be extended to other analytical domains such as sales and payroll to ensure optimal hallucination control guardrails. △ Less

Submitted 9 November, 2023; originally announced November 2023.

Comments: 11 pages, 5 figures, 4 tables

Tackling Concept Shift in Text Classification using Entailment-style Modeling

Authors: Sumegh Roychowdhury, Karan Gupta, Siva Rajesh Kasa, Prasanna Srinivasa Murthy, Alok Chandra

Abstract: Pre-trained language models (PLMs) have seen tremendous success in text classification (TC) problems in the context of Natural Language Processing (NLP). In many real-world text classification tasks, the class definitions being learned do not remain constant but rather change with time - this is known as Concept Shift. Most techniques for handling concept shift rely on retraining the old classifie… ▽ More Pre-trained language models (PLMs) have seen tremendous success in text classification (TC) problems in the context of Natural Language Processing (NLP). In many real-world text classification tasks, the class definitions being learned do not remain constant but rather change with time - this is known as Concept Shift. Most techniques for handling concept shift rely on retraining the old classifiers with the newly labelled data. However, given the amount of training data required to fine-tune large DL models for the new concepts, the associated labelling costs can be prohibitively expensive and time consuming. In this work, we propose a reformulation, converting vanilla classification into an entailment-style problem that requires significantly less data to re-train the text classifier to adapt to new concepts. We demonstrate the effectiveness of our proposed method on both real world & synthetic datasets achieving absolute F1 gains upto 7% and 40% respectively in few-shot settings. Further, upon deployment, our solution also helped save 75% of labeling costs overall. △ Less

Submitted 6 November, 2023; originally announced November 2023.

Journal ref: NeurIPS 2023 - Workshop on Distribution Shifts

Integrability and non-integrability for holographic dual of Matrix model and non-Abelian T-dual of AdS$_5\times$S$^5$

Authors: Jitendra Pal, Sourav Roychowdhury

Abstract: In this paper we study integrability and non-integrability for type-IIA supergravity background dual to deformed plane wave matrix model. From the bulk perspective, we estimate various chaos indicators that clearly shows chaotic string dynamics in the limit of small value of the parameter $L$ present in the theory. On the other hand, the string dynamics exhibits a non-chaotic motion for the large… ▽ More In this paper we study integrability and non-integrability for type-IIA supergravity background dual to deformed plane wave matrix model. From the bulk perspective, we estimate various chaos indicators that clearly shows chaotic string dynamics in the limit of small value of the parameter $L$ present in the theory. On the other hand, the string dynamics exhibits a non-chaotic motion for the large value of the parameter $L$ and therefore presumably an underlying integrable structure. Our findings reveals that the parameter $L$ in the type-IIA background acts as an interpolation between a non-integrable theory to an integrable theory in dual SCFTs. △ Less

Submitted 20 May, 2024; v1 submitted 18 October, 2023; originally announced October 2023.

Comments: 1+23 pages; 15 Figs; Major revision; v3; Accepted to Nucl. Phys. B

Journal ref: Nucl. Phys. B 1004 (2024) 116570

Cellular automata in the light of COVID-19

Authors: Sourav Chowdhury, Suparna Roychowdhury, Indranath Chaudhuri

Abstract: Currently, the world has been facing the brunt of a pandemic due to a disease called COVID-19 for the last 2 years. To study the spread of such infectious diseases it is important to not only understand their temporal evolution but also the spatial evolution. In this work, the spread of this disease has been studied with a cellular automata (CA) model to find the temporal and the spatial behavior… ▽ More Currently, the world has been facing the brunt of a pandemic due to a disease called COVID-19 for the last 2 years. To study the spread of such infectious diseases it is important to not only understand their temporal evolution but also the spatial evolution. In this work, the spread of this disease has been studied with a cellular automata (CA) model to find the temporal and the spatial behavior of it. Here, we have proposed a neighborhood criteria which will help us to measure the social confinement at the time of the disease spread. The two main parameters of our model are (i) disease transmission probability (q) which helps us to measure the infectivity of a disease and (ii) exponent (n) which helps us to measure the degree of the social confinement. Here, we have studied various spatial growths of the disease by simulating this CA model. Finally we have tried to fit our model with the COVID-19 data of India for various waves and have attempted to match our model predictions with regards to each wave to see how the different parameters vary with respect to infectivity and restrictions in social interaction. △ Less

Submitted 31 July, 2023; originally announced July 2023.

Comments: 12 pages, 43 figures and presented in ICNDA 2022

Effect of air pollution on the growth of diabetic population

Authors: Sourav Chowdhury, Suparna Roychowdhury, Indranath Chaudhuri

Abstract: Diabetes mellitus is a disease which is currently a huge health hazard globally. The cases of diabetes had increased by a significant amount in past decades. Also it has been predicted that it will further increase in future. Diabetes depends on various factors like obesity, physical inactivity. Also diabetes can depend on various environmental issues. In this article, our main focus is to study t… ▽ More Diabetes mellitus is a disease which is currently a huge health hazard globally. The cases of diabetes had increased by a significant amount in past decades. Also it has been predicted that it will further increase in future. Diabetes depends on various factors like obesity, physical inactivity. Also diabetes can depend on various environmental issues. In this article, our main focus is to study the dependence of the diabetic cases on the air pollution. We have used the data for diabetic population and PM2.5 concentration in the air for five countries from 2010 to 2021. Here we have studied the correlation between the diabetic cases data and PM2.5 concentration data. Also, we have done the linear regression analysis to find whether this correlation is statistically significant. △ Less

Submitted 31 July, 2023; originally announced July 2023.

Comments: 4 pages, 5 figures, presented in the International Conference 'International Conference on Climate Change: Global Cooperation' at St. Xavier's College (Autonomous), Kolkata, India

Simulating the spread of COVID-19 with cellular automata: A new approach

Authors: Sourav Chowdhury, Suparna Roychowdhury, Indranath Chaudhuri

Abstract: Between the years 2020 to 2022, the world was hit by the pandemic of COVID-19 giving rise to an extremely grave situation. The global economy was badly hurt due to the consequences of various intervention strategies (like social distancing, lockdown) which were applied by different countries to control this pandemic. There are multiple speculations that humanity will again face such pandemics in t… ▽ More Between the years 2020 to 2022, the world was hit by the pandemic of COVID-19 giving rise to an extremely grave situation. The global economy was badly hurt due to the consequences of various intervention strategies (like social distancing, lockdown) which were applied by different countries to control this pandemic. There are multiple speculations that humanity will again face such pandemics in the future. Thus it is very important to learn and gain knowledge about the spread of such infectious diseases and the various factors which are responsible for it. In this study, we have extended our previous work (Chowdhury et.al., 2022) on the probabilistic cellular automata (CA) model to reproduce the spread of COVID-19 in several countries by modifying its earlier used neighbourhood criteria. This modification gives us the liberty to adopt the effect of different restrictions like lockdown and social distancing in our model. We have done some theoretical analysis for initial infection and simulations to gain insights into our model. We have also studied the data from eight countries for COVID-19 in a window of 876 days and compared it with our model. We have developed a proper framework to fit our model on the data for confirmed cases of COVID-19 and have also re-checked the goodness of the fit with the data of the deceased cases for this pandemic. This model fits well with different peaks of COVID-19 data for all the eight countries and can be possibly generalized for a global prediction. △ Less

Submitted 26 July, 2023; originally announced July 2023.

Comments: 22 pages, 57 figures

Integrability and non-integrability for marginal deformations of 4d $\mathcal N = 2$ SCFTs

Authors: Jitendra Pal, Sourav Roychowdhury, Arindam Lala, Dibakar Roychowdhury

Abstract: We study integrability and non-integrability for marginal deformations of 4d $\mathcal N =2$ SCFTs. We estimate various chaos indicators for the bulk theory which clearly shows the onset of a chaotic string dynamics in the limit of large deformations. On the other hand, for small values of the deformation parameter, the resulting dynamics exhibits a non-chaotic motion and therefore presumably an u… ▽ More We study integrability and non-integrability for marginal deformations of 4d $\mathcal N =2$ SCFTs. We estimate various chaos indicators for the bulk theory which clearly shows the onset of a chaotic string dynamics in the limit of large deformations. On the other hand, for small values of the deformation parameter, the resulting dynamics exhibits a non-chaotic motion and therefore presumably an underlying integrable structure. Our analysis reveals that the $γ$-deformation in the type-IIA theory could be interpreted as an interpolation between a class of integrable $\mathcal N =2$ SCFTs and a class of non-integrable $\mathcal N =1$ SCFTs at strong coupling. We also generalise our results in the presence of the flavor branes. △ Less

Submitted 16 October, 2023; v1 submitted 22 July, 2023; originally announced July 2023.

Comments: 1+22 pages; 27 Figs; Reference added; Appendices added; v2; Accepted to JHEP

Journal ref: JHEP 10 (2023) 173

Galaxy And Mass Assembly (GAMA): The group HI mass as a function of halo mass

Authors: Ajay Dev, Simon P. Driver, Martin Meyer, Sambit Roychowdhury, Jonghwan Rhee, Adam R. H. Stevens, Claudia del P. Lagos, Joss Bland-Hawthorn, Barbara Catinella, A. M. Hopkins, Jonathan Loveday, Danail Obreschkow, Steven Phillipps, Aaron S. G. Robotham

Abstract: We determine the atomic hydrogen (HI) to halo mass relation (HIHM) using Arecibo Legacy Fast ALFA survey HI data at the location of optically selected groups from the Galaxy and Mass Assembly (GAMA) survey. We make direct HI detections for 37 GAMA groups. Using HI group spectral stacking of 345 groups, we study the group HI content as function of halo mass across a halo mass range of… ▽ More We determine the atomic hydrogen (HI) to halo mass relation (HIHM) using Arecibo Legacy Fast ALFA survey HI data at the location of optically selected groups from the Galaxy and Mass Assembly (GAMA) survey. We make direct HI detections for 37 GAMA groups. Using HI group spectral stacking of 345 groups, we study the group HI content as function of halo mass across a halo mass range of $10^{11} - 10^{14.7}\text{ M}_\odot$. We also correct our results for Eddington bias. We find that the group HI mass generally rises as a function of halo mass from $1.3\%$ of the halo mass at $10^{11.6} \text{M}_\odot$ to $0.4\%$ at $10^{13.7} \text{M}_\odot$ with some indication of flattening towards the high-mass end. Despite the differences in optical survey limits, group catalogues, and halo mass estimation methods, our results are consistent with previous group HI-stacking studies. Our results are also consistent with mock observations from SHARK and IllustrisTNG. △ Less

Submitted 22 May, 2023; originally announced May 2023.

Comments: Accepted in MNRAS; 18 pages, 12 figures

Quantum oscillations revealing topological band in kagome metal ScV6Sn6

Authors: Changjiang Yi, Xiaolong Feng, Ning Mao, Premakumar Yanda, Subhajit Roychowdhury, Yang Zhang, Claudia Felser, Chandra Shekhar

Abstract: Compounds with kagome lattice structure are known to exhibit Dirac cones, flat bands, and van Hove singularities, which host numerous versatile quantum phenomena. Inspired by these intriguing properties, we investigate the temperature and magnetic field dependent electrical transports along with the theoretical calculations of ScV6Sn6, a nonmagnetic charge density wave (CDW) compound. At low tempe… ▽ More Compounds with kagome lattice structure are known to exhibit Dirac cones, flat bands, and van Hove singularities, which host numerous versatile quantum phenomena. Inspired by these intriguing properties, we investigate the temperature and magnetic field dependent electrical transports along with the theoretical calculations of ScV6Sn6, a nonmagnetic charge density wave (CDW) compound. At low temperatures, the compound exhibits Shubnikov-de Haas quantum oscillations, which help to design the Fermi surface (FS) topology. This analysis reveals the existence of several small FSs in the Brillouin zone, combined with a large FS. Among them, the FS possessing Dirac band is a non-trivial and generates a non-zero Berry phase. In addition, the compound also shows the anomalous Hall-like behaviour up to the CDW with the CDW phase, ScV6Sn6 presents a unique material example of the versatile HfFe6Ge6 family and provides various promising opportunities to explore the series further. △ Less

Submitted 22 January, 2024; v1 submitted 8 May, 2023; originally announced May 2023.

Comments: Published version, 19 Pages, 5 figures with supplementary

Journal ref: Phys. Rev. B 109, 035124 (2024)

Softening of a flat phonon mode in the kagome ScV$_6$Sn$_6$

Authors: A. Korshunov, H. Hu, D. Subires, Y. Jiang, D. Călugăru, X. Feng, A. Rajapitamahuni, C. Yi, S. Roychowdhury, M. G. Vergniory, J. Strempfer, C. Shekhar, E. Vescovo, D. Chernyshov, A. H. Said, A. Bosak, C. Felser, B. Andrei Bernevig, S. Blanco-Canosa

Abstract: The long range electronic modulations recently discovered in the geometrically frustrated kagome lattice have opened new avenues to explore the effect of correlations in materials with topological electron flat bands. The observation of the lattice response to the emergent new phases of matter, a soft phonon mode, has remained elusive and the microscopic origin of charge density waves (CDWs) is st… ▽ More The long range electronic modulations recently discovered in the geometrically frustrated kagome lattice have opened new avenues to explore the effect of correlations in materials with topological electron flat bands. The observation of the lattice response to the emergent new phases of matter, a soft phonon mode, has remained elusive and the microscopic origin of charge density waves (CDWs) is still unknown. Here, we show, for the first time, a complete melting of the ScV$_ 6$Sn$_ 6$ (166) kagome lattice. The low energy phonon with propagation vector $\frac{1}{3} \frac{1}{3} \frac{1}{2}$ collapses at 98 K, without the emergence of long-range charge order, which sets in with a propagation vector $\frac{1}{3} \frac{1}{3} \frac{1}{3}$. The CDW is driven (but locks at a different vector) by the softening of an overdamped phonon flat plane at k$_z$=$π$. We observe broad phonon anomalies in momentum space, pointing to (1) the existence of approximately flat phonon bands which gain some dispersion due to electron renormalization, and (2) the effects of the momentum dependent electron-phonon interaction in the CDW formation. Ab initio and analytical calculations corroborate the experimental findings to indicate that the weak leading order phonon instability is located at the wave vector $\frac{1}{3} \frac{1}{3} \frac{1}{2}$ of a rather flat collapsed mode. We analytically compute the phonon frequency renormalization from high temperatures to the soft mode, and relate it to a peak in the orbital-resolved susceptibility, obtaining an excellent match with both ab initio and experimental results, and explaining the origin of the approximately flat phonon dispersion. Our data report the first example of the collapse of a softening of a flat phonon plane and promote the 166 compounds of the kagome family as primary candidates to explore correlated flat phonon-topological flat electron physics. △ Less

Submitted 18 April, 2023; originally announced April 2023.

Comments: 10 pages, 4 figures

Journal ref: Nat. Commun. 14, 6646 (2023)

NUMSnet: Nested-U Multi-class Segmentation network for 3D Medical Image Stacks

Authors: Sohini Roychowdhury

Abstract: Semantic segmentation for medical 3D image stacks enables accurate volumetric reconstructions, computer-aided diagnostics and follow up treatment planning. In this work, we present a novel variant of the Unet model called the NUMSnet that transmits pixel neighborhood features across scans through nested layers to achieve accurate multi-class semantic segmentations with minimal training data. We an… ▽ More Semantic segmentation for medical 3D image stacks enables accurate volumetric reconstructions, computer-aided diagnostics and follow up treatment planning. In this work, we present a novel variant of the Unet model called the NUMSnet that transmits pixel neighborhood features across scans through nested layers to achieve accurate multi-class semantic segmentations with minimal training data. We analyze the semantic segmentation performance of the NUMSnet model in comparison with several Unet model variants to segment 3-7 regions of interest using only 10% of images for training per Lung-CT and Heart-CT volumetric image stacks. The proposed NUMSnet model achieves up to 20% improvement in segmentation recall with 4-9% improvement in Dice scores for Lung-CT stacks and 2.5-10% improvement in Dice scores for Heart-CT stacks when compared to the Unet++ model. The NUMSnet model needs to be trained by ordered images around the central scan of each volumetric stack. Propagation of image feature information from the 6 nested layers of the Unet++ model are found to have better computation and segmentation performances than propagation of all up-sampling layers in a Unet++ model. The NUMSnet model achieves comparable segmentation performances to existing works, while being trained on as low as 5\% of the training images. Also, transfer learning allows faster convergence of the NUMSnet model for multi-class semantic segmentation from pathology in Lung-CT images to cardiac segmentations in Heart-CT stacks. Thus, the proposed model can standardize multi-class semantic segmentation on a variety of volumetric image stacks with minimal training dataset. This can significantly reduce the cost, time and inter-observer variabilities associated with computer-aided detections and treatment. △ Less

Submitted 5 April, 2023; originally announced April 2023.

Comments: 15 pages, 10 pages, 8 tables

Chaotic dynamics of off-equatorial orbits around pseudo-Newtonian compact objects with dipolar halos

Authors: Saikat Das, Suparna Roychowdhury

Abstract: In this paper, we implement a generalised pseudo-Newtonian potential to study the off-equatorial orbits inclined at a certain angle with the equatorial plane around Schwarzschild and Kerr-like compact object primaries surrounded by a dipolar halo of matter. The chaotic dynamics of the orbits are detailed for both non-relativistic and special-relativistic test particles. The dependence of the degre… ▽ More In this paper, we implement a generalised pseudo-Newtonian potential to study the off-equatorial orbits inclined at a certain angle with the equatorial plane around Schwarzschild and Kerr-like compact object primaries surrounded by a dipolar halo of matter. The chaotic dynamics of the orbits are detailed for both non-relativistic and special-relativistic test particles. The dependence of the degree of chaos on the Kerr parameter $a$ and the inclination angle $i$ is established individually using widely used indicators, such as the Poincaré Maps and the Maximum Lyapunov Exponents. Although the orbits' chaoticity has a positive correlation with $i$, the growth in the chaotic behaviour is not systematic. There is a threshold value of the inclination angle $i_{\text{c}}$, after which the degree of chaos sharply increases. On the other hand, the chaoticity of the inclined orbits anti-correlates with $a$ throughout its entire range. However, the negative correlation is systematic at lower values of the inclination angle. At higher values of $i$, the degree of chaos increases rapidly below a threshold value of the Kerr parameter, $a_{\text{c}}$. Above this threshold value, the correlation becomes weak. Furthermore, we establish a qualitative correlation between the threshold values and the overall chaoticity of the system. The studies performed with different orbital parameters and several initial conditions reveal the intricate nature of the system. △ Less

Submitted 29 December, 2023; v1 submitted 26 March, 2023; originally announced March 2023.

Comments: 23 pages, 11 figures. A corrected version of the manuscript. Accepted for publication in Chaos, Solitons & Fractals

Spin $ 2 $ spectrum for marginal deformations of 4d $ \mathcal{N}=2 $ SCFTs

Authors: Sourav Roychowdhury, Dibakar Roychowdhury

Abstract: We compute spin $ 2 $ spectrum associated with massive graviton fluctuations in $γ$-deformed Gaiotto-Maldacena background those are holographically dual to marginal deformations of $\mathcal{N}=2$ SCFTs in four dimensions. Under the special circumstances, we analytically estimate the spectra both for the $ γ$- deformed Abelian T dual (ATD) as well as the non-Abelian T dual (NATD) cases where we re… ▽ More We compute spin $ 2 $ spectrum associated with massive graviton fluctuations in $γ$-deformed Gaiotto-Maldacena background those are holographically dual to marginal deformations of $\mathcal{N}=2$ SCFTs in four dimensions. Under the special circumstances, we analytically estimate the spectra both for the $ γ$- deformed Abelian T dual (ATD) as well as the non-Abelian T dual (NATD) cases where we retain ourselves upto leading order in the deformation parameter. Our analysis reveals a continuous spectra which is associated with the breaking of the $ U(1) $ isometry (along the directions of the internal manifold) in the presence of the $ γ$- deformation. We also comment on the effects of adding flavour branes into the picture and the nature of the associated spin $ 2 $ operators in the dual $ \mathcal{N}=1 $ SCFTs. △ Less

Submitted 2 March, 2023; v1 submitted 30 January, 2023; originally announced January 2023.

Comments: 1+19 pages ; v2 ; Accepted to JHEP

Journal ref: JHEP 03 (2023) 083

The resolved scaling relations in DustPedia: Zooming in on the local Universe

Authors: Viviana Casasola, Simone Bianchi, Laura Magrini, Aleksandr V. Mosenkov, Francesco Salvestrini, Maarten Baes, Francesco Calura, Letizia P. Cassara', Christopher J. R. Clark, Edvige Corbelli, Jacopo Fritz, Frederic Galliano, Elisabetta Liuzzo, Suzanne Madden, Angelos Nersesian, Francesca Pozzi, Sambit Roychowdhury, Ivano Baronchelli, Matteo Bonato, Carlotta Gruppioni, Lara Pantoni

Abstract: We perform a homogeneous analysis of an unprecedented set of spatially resolved scaling relations (SRs) between ISM components and other properties in the range of scales 0.3-3.4 kpc. We also study some ratios: dust-to-stellar, dust-to-gas, and dust-to-metal. We use a sample of 18 large, spiral, face-on DustPedia galaxies. All the SRs are moderate/strong correlations except the dust-HI SR that doe… ▽ More We perform a homogeneous analysis of an unprecedented set of spatially resolved scaling relations (SRs) between ISM components and other properties in the range of scales 0.3-3.4 kpc. We also study some ratios: dust-to-stellar, dust-to-gas, and dust-to-metal. We use a sample of 18 large, spiral, face-on DustPedia galaxies. All the SRs are moderate/strong correlations except the dust-HI SR that does not exist or is weak for most galaxies. The SRs do not have a universal form but each galaxy is characterized by distinct correlations, affected by local processes and galaxy peculiarities. The SRs hold starting from 0.3 kpc, and if a breaking down scale exists it is < 0.3 kpc. By evaluating all galaxies at 3.4 kpc, differences due to peculiarities of individual galaxies are cancelled out and the corresponding SRs are consistent with those of whole galaxies. By comparing subgalactic and global scales, the most striking result emerges from the SRs involving ISM components: the dust-total gas SR is a good correlation at all scales, while the dust-H2 and dust-HI SRs are good correlations at subkpc/kpc and total scales, respectively. For the other explored SRs, there is a good agreement between small and global scales and this may support the picture where the main physical processes regulating the properties and evolution of galaxies occur locally. Our results are consistent with the hypothesis of self-regulation of the SF process. The analysis of subgalactic ratios shows that they are consistent with those derived for whole galaxies, from low to high z, supporting the idea that also these ratios could be set by local processes. Our results highlight the heterogeneity of galaxy properties and the importance of resolved studies on local galaxies in the context of galaxy evolution. They also provide observational constraints to theoretical models and updated references for high-z studies. △ Less

Submitted 28 October, 2022; originally announced October 2022.

Comments: 42 pages, 11 figures and 5 tables in the main text, 2 figures and 1 table in Appendix. Accepted for publication in A&A

Journal ref: A&A 668, A130 (2022)

Deep Investigation of Neutral Gas Origins (DINGO): HI stacking experiments with early science data

Authors: Jonghwan Rhee, Martin Meyer, Attila Popping, Sabine Bellstedt, Simon P. Driver, Aaron S. G. Robotham, Matthew Whiting, Ivan K. Baldry, Sarah Brough, Michael J. I. Brown, John D. Bunton, Richard Dodson, Benne W. Holwerda, Andrew M. Hopkins, Bärbel S. Koribalski, Karen Lee-Waddell, Ángel R. López-Sánchez, Jon Loveday, Elizabeth Mahony, Sambit Roychowdhury, Kristóf Rozgonyi, Lister Staveley-Smith

Abstract: We present early science results from Deep Investigation of Neutral Gas Origins (DINGO), an HI survey using the Australian Square Kilometre Array Pathfinder (ASKAP). Using ASKAP sub-arrays available during its commissioning phase, DINGO early science data were taken over $\sim$ 60 deg$^{2}$ of the Galaxy And Mass Assembly (GAMA) 23 h region with 35.5 hr integration time. We make direct detections… ▽ More We present early science results from Deep Investigation of Neutral Gas Origins (DINGO), an HI survey using the Australian Square Kilometre Array Pathfinder (ASKAP). Using ASKAP sub-arrays available during its commissioning phase, DINGO early science data were taken over $\sim$ 60 deg$^{2}$ of the Galaxy And Mass Assembly (GAMA) 23 h region with 35.5 hr integration time. We make direct detections of six known and one new sources at $z < 0.01$. Using HI spectral stacking, we investigate the HI gas content of galaxies at $0.04 < z< 0.09$ for different galaxy colours. The results show that galaxy morphology based on optical colour is strongly linked to HI gas properties. To examine environmental impacts on the HI gas content of galaxies, three sub-samples are made based on the GAMA group catalogue. The average HI mass of group central galaxies is larger than those of satellite and isolated galaxies, but with a lower HI gas fraction. We derive a variety of HI scaling relations for physical properties of our sample, including stellar mass, stellar mass surface density, $NUV-r$ colour, specific star formation rate, and halo mass. We find that the derived HI scaling relations are comparable to other published results, with consistent trends also observed to $\sim$0.5 dex lower limits in stellar mass and stellar surface density. The cosmic HI densities derived from our data are consistent with other published values at similar redshifts. DINGO early science highlights the power of HI spectral stacking techniques with ASKAP. △ Less

Submitted 20 October, 2022; v1 submitted 18 October, 2022; originally announced October 2022.

Comments: 27 pages, 25 figures, 10 tables, accepted for publication in MNRAS

Probing galaxy evolution through HI 21-cm emission and absorption: current status and prospects with the Square Kilometre Array

Authors: Rajeshwari Dutta, Sushma Kurapati, J. N. H. S. Aditya, Omkar Bait, Mousumi Das, Prasun Dutta, K. Indulekha, Meera Nandakumar, Narendra Nath Patra, Nirupam Roy, Sambit Roychowdhury

Abstract: One of the major science goals of the Square Kilometre Array (SKA) is to understand the role played by atomic hydrogen (HI) gas in the evolution of galaxies throughout cosmic time. The hyperfine transition line of the hydrogen atom at 21-cm is one of the best tools to detect and study the properties of HI gas associated with galaxies. In this article, we review our current understanding of HI gas… ▽ More One of the major science goals of the Square Kilometre Array (SKA) is to understand the role played by atomic hydrogen (HI) gas in the evolution of galaxies throughout cosmic time. The hyperfine transition line of the hydrogen atom at 21-cm is one of the best tools to detect and study the properties of HI gas associated with galaxies. In this article, we review our current understanding of HI gas and its relationship with galaxies through observations of the 21-cm line both in emission and absorption. In addition, we provide an overview of the HI science that will be possible with SKA and its pre-cursors and pathfinders, i.e. HI 21-cm emission and absorption studies of galaxies from nearby to high redshifts that will trace various processes governing galaxy evolution. △ Less

Submitted 21 September, 2022; v1 submitted 19 September, 2022; originally announced September 2022.

Comments: 31 pages, 7 figures, accepted on 27 May 2022 for publication in the Journal of Astrophysics and Astronomy (to appear in the special issue on "Indian participation in the SKA"), figure 4 has been updated

Droplet Migration in the Presence of a Reacting Surfactant at Low Péclet Numbers

Authors: Souradeep Roychowdhury, Rajarshi Chattopadhyay, Rahul Mangal, Dipin S. Pillai

Abstract: A surfactant-laden droplet of one fluid dispersed in another immiscible fluid serves as an artificial model system capable of mimicking microbial swimmers. Either an interfacial chemical reaction or the process of solubilization generates gradients in interfacial tension resulting in a Marangoni flow. The resulting fluid flow propels the droplet toward a region of lower interfacial tension. The ad… ▽ More A surfactant-laden droplet of one fluid dispersed in another immiscible fluid serves as an artificial model system capable of mimicking microbial swimmers. Either an interfacial chemical reaction or the process of solubilization generates gradients in interfacial tension resulting in a Marangoni flow. The resulting fluid flow propels the droplet toward a region of lower interfacial tension. The advective transport of surfactants sustains the active propulsion of these droplets. In these systems, the local interfacial tension is affected by the interfacial reaction kinetics as well as convection and diffusion induced concentration gradients. The migration of such a surfactant-laden viscous droplet undergoing an interfacial reaction, suspended in a background Poiseuille flow is investigated. The focus is specifically on the role of the surface reaction that generates a non-uniform interfacial coverage of the surfactant, which in turn dictates the migration velocity of the droplet in the background flow. Assuming negligible interface deformation and fluid inertia, the Lorentz reciprocal theorem is used to analytically determine the migration velocity of the droplet using regular perturbation expansion in terms of the surface Péclet number. We show that the presence of interfacial reaction affects the magnitude of both stream-wise and cross-stream migration velocity of the droplet in a background Poiseuille flow. We conclude that the stream-wise migration velocity is not of sufficient strength to exhibit positive rheotaxis as observed in recent experimental observations. Additional effects such as the hydrodynamic interactions with the adjacent wall may be essential to capture the same. △ Less

Submitted 18 August, 2022; originally announced August 2022.

Comments: 2 figures

Unusual gas structure in an otherwise normal spiral galaxy hosting GRB 171205A / SN 2017iuk

Authors: M. Arabsalmani, S. Roychowdhury, F. Renaud, A. Burkert, E. Emsellem, E. Le Floc'h, E. Pian

Abstract: We study the structure of atomic hydrogen (HI) in the host galaxy of GRB 171205A / SN 2017iuk at z=0.037 through HI 21cm emission line observations with the Karl G. Jansky Very Large Array. These observations reveal unusual morphology and kinematics of the HI in this otherwise apparently normal galaxy. High column density, cold HI is absent from an extended North-South region passing by the optica… ▽ More We study the structure of atomic hydrogen (HI) in the host galaxy of GRB 171205A / SN 2017iuk at z=0.037 through HI 21cm emission line observations with the Karl G. Jansky Very Large Array. These observations reveal unusual morphology and kinematics of the HI in this otherwise apparently normal galaxy. High column density, cold HI is absent from an extended North-South region passing by the optical centre of the galaxy, but instead is extended towards the South, on both sides of the galaxy. Moreover, the HI kinematics do not show a continuous change along the major axis of the galaxy as expected in a classical rotating disk. We explore several scenarios to explain the HI structure and kinematics in the galaxy: feedback from a central starburst and/or an active galactic nucleus, ram pressure stripping, accretion, and tidal interaction from a companion galaxy. All of these options are ruled out. The most viable remaining explanation is the penetrating passage of a satellite through the disk only a few Myr ago, redistributing the HI in the GRB host without yet affecting its stellar distribution. It can also lead to the rapid formation of peculiar stars due to a violent induced shock. The location of GRB 171205A in the vicinity of the distorted area suggests that its progenitor star(s) originated in extreme conditions that share the same origin as the peculiarities in HI. This could explain the atypical location of GRB 171205A in its host galaxy. △ Less

Submitted 14 June, 2022; originally announced June 2022.

Comments: This is the Accepted Manuscript version of an article accepted for publication in The Astronomical Journal. IOP Publishing Ltd is not responsible for any errors or omissions in this version of the manuscript or any version derived from it. The manuscript has 8 pages and 2 figures

CRUSH: Contextually Regularized and User anchored Self-supervised Hate speech Detection

Authors: Souvic Chakraborty, Parag Dutta, Sumegh Roychowdhury, Animesh Mukherjee

Abstract: The last decade has witnessed a surge in the interaction of people through social networking platforms. While there are several positive aspects of these social platforms, the proliferation has led them to become the breeding ground for cyber-bullying and hate speech. Recent advances in NLP have often been used to mitigate the spread of such hateful content. Since the task of hate speech detection… ▽ More The last decade has witnessed a surge in the interaction of people through social networking platforms. While there are several positive aspects of these social platforms, the proliferation has led them to become the breeding ground for cyber-bullying and hate speech. Recent advances in NLP have often been used to mitigate the spread of such hateful content. Since the task of hate speech detection is usually applicable in the context of social networks, we introduce CRUSH, a framework for hate speech detection using user-anchored self-supervision and contextual regularization. Our proposed approach secures ~ 1-12% improvement in test set metrics over best performing previous approaches on two types of tasks and multiple popular english social media datasets. △ Less

Submitted 4 May, 2022; v1 submitted 13 April, 2022; originally announced April 2022.

Comments: Accepted in NAACL HLT 2022 (Long Paper)

ACM Class: I.2.7; J.4

ALMA/ACA CO Survey of the IC 1459 and NGC 4636 Groups: Environmental Effects on the Molecular Gas of Group Galaxies

Authors: Bumhyun Lee, Jing Wang, Aeree Chung, Luis C. Ho, Ran Wang, Tomonari Michiyama, Juan Molina, Yongjung Kim, Li Shao, Virginia Kilborn, Shun Wang, Xuchen Lin, Dawoon E. Kim, B. Catinella, L. Cortese, N. Deg, H. Dénes, A. Elagali, Bi-Qing For, D. Kleiner, B. S. Koribalski, K. Lee-Waddell, J. Rhee, K. Spekkens, T. Westmeier , et al. (8 additional authors not shown)

Abstract: We present new results of a 12CO(J=1-0) imaging survey using the Atacama Compact Array (ACA) for 31 HI detected galaxies in the IC 1459 and NGC 4636 groups. This is the first CO imaging survey for loose galaxy groups. We obtained well-resolved CO data (~0.7-1.5 kpc) for a total of 16 galaxies in two environments. By comparing our ACA CO data with the HI and UV data, we probe the impacts of the gro… ▽ More We present new results of a 12CO(J=1-0) imaging survey using the Atacama Compact Array (ACA) for 31 HI detected galaxies in the IC 1459 and NGC 4636 groups. This is the first CO imaging survey for loose galaxy groups. We obtained well-resolved CO data (~0.7-1.5 kpc) for a total of 16 galaxies in two environments. By comparing our ACA CO data with the HI and UV data, we probe the impacts of the group environment on the cold gas components (CO and HI gas) and star formation activity. We find that CO and/or HI morphologies are disturbed in our group members, some of which show highly asymmetric CO distributions (e.g., IC 5264, NGC 7421, and NGC 7418). In comparison with isolated galaxies in the xCOLD GASS sample, our group galaxies tend to have low star formation rates and low H2 gas fractions. Our findings suggest that the group environment can change the distribution of cold gas components, including the molecular gas, and star formation properties of galaxies. This is supporting evidence that preprocessing in the group-like environment can play an important role in galaxy evolution. △ Less

Submitted 31 August, 2023; v1 submitted 12 April, 2022; originally announced April 2022.

Comments: 42 pages, 29 figures, 6 tables, published in ApJS

Penrose limits in massive type-IIA AdS$_3$ background

Authors: Sourav Roychowdhury, Prasanta K. Tripathy

Abstract: In this paper we consider the non-Abelian T-dual geometry of the type $IIB$ supergravity theory on $AdS_3\times S^3\times T^4$ background along a convenient $SU(2)$ subgroup of the $SO(4)$ R-symmetry. We examine various null geodesics of the resulting massive type $IIA$ supergravity theory and investigate the Penrose limits along these geodesics. We find that one of the resulting backgrounds admit… ▽ More In this paper we consider the non-Abelian T-dual geometry of the type $IIB$ supergravity theory on $AdS_3\times S^3\times T^4$ background along a convenient $SU(2)$ subgroup of the $SO(4)$ R-symmetry. We examine various null geodesics of the resulting massive type $IIA$ supergravity theory and investigate the Penrose limits along these geodesics. We find that one of the resulting backgrounds admits pp-wave geometry in the neighbourhood of a suitable null geodesic. We carry out the supersymmetry analysis of the resulting pp-wave geometry and observe that it preserves sixteen supercharges. Further we comment on the possible gauge theory dual of the resulting pp-wave background. △ Less

Submitted 15 May, 2022; v1 submitted 27 March, 2022; originally announced March 2022.

Comments: 1+18 pages; v2; Minor modifications; Accepted to Phys. Rev. D

Journal ref: Phys. Rev. D. 105, 106024 (2022)

Semi-supervised and Deep learning Frameworks for Video Classification and Key-frame Identification

Authors: Sohini Roychowdhury

Abstract: Automating video-based data and machine learning pipelines poses several challenges including metadata generation for efficient storage and retrieval and isolation of key-frames for scene understanding tasks. In this work, we present two semi-supervised approaches that automate this process of manual frame sifting in video streams by automatically classifying scenes for content and filtering frame… ▽ More Automating video-based data and machine learning pipelines poses several challenges including metadata generation for efficient storage and retrieval and isolation of key-frames for scene understanding tasks. In this work, we present two semi-supervised approaches that automate this process of manual frame sifting in video streams by automatically classifying scenes for content and filtering frames for fine-tuning scene understanding tasks. The first rule-based method starts from a pre-trained object detector and it assigns scene type, uncertainty and lighting categories to each frame based on probability distributions of foreground objects. Next, frames with the highest uncertainty and structural dissimilarity are isolated as key-frames. The second method relies on the simCLR model for frame encoding followed by label-spreading from 20% of frame samples to label the remaining frames for scene and lighting categories. Also, clustering the video frames in the encoded feature space further isolates key-frames at cluster boundaries. The proposed methods achieve 64-93% accuracy for automated scene categorization for outdoor image videos from public domain datasets of JAAD and KITTI. Also, less than 10% of all input frames can be filtered as key-frames that can then be sent for annotation and fine tuning of machine vision algorithms. Thus, the proposed framework can be scaled to additional video data streams for automated training of perception-driven systems with minimal training images. △ Less

Submitted 25 March, 2022; originally announced March 2022.

Comments: 9 pages, 7 images, 3 tables

Heating of the intracluster medium by buoyant bubbles and sound waves

Authors: Asif Iqbal, Subhabrata Majumdar, Biman B. Nath, Suparna Roychowdhury

Abstract: Active galactic nuclei (AGN) powered by the central Super-Massive Black Holes (SMBHs) play a major role in modifying the thermal properties of the intracluster medium (ICM). In this work, we implement two AGN heating models: (i) by buoyant cavities rising through stratified ICM (effervescent model) and, (ii) by viscous and conductive dissipation of sound waves (acoustic model). Our aim is to deter… ▽ More Active galactic nuclei (AGN) powered by the central Super-Massive Black Holes (SMBHs) play a major role in modifying the thermal properties of the intracluster medium (ICM). In this work, we implement two AGN heating models: (i) by buoyant cavities rising through stratified ICM (effervescent model) and, (ii) by viscous and conductive dissipation of sound waves (acoustic model). Our aim is to determine whether these heating models are consistent with ICM observables and if one is preferred over the other. We assume an initial entropy profile of ICM that is expected from the purely gravitational infall of the gas in the potential of the dark matter halo. We then incorporate heating, radiative cooling, and thermal conduction to study the evolution of ICM over the age of the clusters. Our results are: (i) Both the heating processes can produce comparable thermal profiles of the ICM with some tuning of relevant parameters. (ii) Thermal conduction is crucially important, even at the level of 10\% of the Spitzer values, in transferring the injected energy beyond the central regions, and without which the temperature/entropy profiles are unrealistically high. (iii) The required injected AGN power scales with cluster mass as $M_{\rm vir}^{1.5}$ for both models. (iv) The required AGN luminosity is comparable with the observed radio jet power, reinforcing the idea that AGNs are the dominant heating source in clusters. (v) Finally, we estimate that the fraction of the total AGN luminosity available as the AGN mechanical luminosity at $0.02r_{500}$ is less than 0.05\%. △ Less

Submitted 3 November, 2022; v1 submitted 24 March, 2022; originally announced March 2022.

Comments: 13 pages, 5 figures, Accepted in MNRAS

Journal ref: MNRAS, 518, 2735-2745 (2023)

Heterotic Kerr-Schild Double Field Theory and its double Yang-Mills formulation

Authors: Eric Lescano, Sourav Roychowdhury

Abstract: We present a formulation of heterotic Double Field Theory (DFT), where the fundamental fields are in $O(D,D)$ representations. The theory is obtained splitting an $O(D,D+K)$ duality invariant DFT. This procedure produces a Green-Schwarz mechanism for the generalized metric, and a fundamental gauge field which transforms as a gauge connection only to leading order. After parametrization, the former… ▽ More We present a formulation of heterotic Double Field Theory (DFT), where the fundamental fields are in $O(D,D)$ representations. The theory is obtained splitting an $O(D,D+K)$ duality invariant DFT. This procedure produces a Green-Schwarz mechanism for the generalized metric, and a fundamental gauge field which transforms as a gauge connection only to leading order. After parametrization, the former induces a non-covariant transformation on the metric tensor, which can be removed considering field redefinitions, and an ordinary Green-Schwarz mechanism on the b-field. Within this framework we explore perturbative properties of heterotic DFT. We use a relaxed version of the generalized Kerr-Schild ansatz (GKSA), where the generalized background metric is perturbed up to quadratic order considering a single null vector and the gauge field is linearly perturbed before parametrization. Finally we compare the dynamics of the gauge field and the generalized metric in order to inspect the behavior of the classical double copy correspondence at the DFT level. △ Less

Submitted 25 April, 2022; v1 submitted 23 January, 2022; originally announced January 2022.

Comments: 33 pages, v2: Section 6 (Discussion) with clarifications. We are very grateful to our anonymous JHEP referee. Matches published version

Journal ref: JHEP04(2022)090

The variation of the gas content of galaxy groups and pairs compared to isolated galaxies

Authors: Sambit Roychowdhury, Martin J. Meyer, Jonghwan Rhee, Martin A. Zwaan, Garima Chauhan, Luke J. M. Davies, Sabine Bellstedt, Simon P. Driver, Claudia del P. Lagos, Aaron S. G. Robotham, Joss Bland-Hawthorn, Richard Dodson, Benne W. Holwerda, Andrew M. Hopkins, Maritza A. Lara-Lopez, Angel R. Lopez-Sanchez, Danail Obreschkow, Kristof Rozgonyi, Matthew T. Whiting, Angus H. Wright

Abstract: We measure how the atomic gas (HI) fraction ($f_{HI}={\rm \frac{M_{HI}}{M_{*}}}$) of groups and pairs taken as single units vary with average stellar mass ($\langle {\rm M_*} \rangle$) and average star-formation rate ($\langle {\rm SFR} \rangle$), compared to isolated galaxies. The HI 21 cm emission observation are from (i) archival ALFALFA survey data covering three fields from the GAMA survey (p… ▽ More We measure how the atomic gas (HI) fraction ($f_{HI}={\rm \frac{M_{HI}}{M_{*}}}$) of groups and pairs taken as single units vary with average stellar mass ($\langle {\rm M_*} \rangle$) and average star-formation rate ($\langle {\rm SFR} \rangle$), compared to isolated galaxies. The HI 21 cm emission observation are from (i) archival ALFALFA survey data covering three fields from the GAMA survey (provides environmental and galaxy properties), and (ii) DINGO pilot survey data of one of those fields. The mean $f_{HI}$ for different units (groups/pairs/isolated galaxies) are measured in regions of the log($\langle {\rm M_*} \rangle$) -- log($\langle {\rm SFR} \rangle$) plane, relative to the z $\sim 0$ star-forming main sequence (SFMS) of individual galaxies, by stacking $f_{HI}$ spectra of individual units. For ALFALFA, $f_{HI}$ spectra of units are measured by extracting HI spectra over the full groups/pair areas and dividing by the total stellar mass of member galaxies. For DINGO, $f_{HI}$ spectra of units are measured by co-adding HI spectra of individual member galaxies, followed by division by their total stellar mass. For all units the mean $f_{HI}$ decreases as we move to higher $\langle {\rm M_*} \rangle$ along the SFMS, and as we move from above the SFMS to below it at any $\langle {\rm M_*} \rangle$. From the DINGO-based study, mean $f_{HI}$ in groups appears to be lower compared to isolated galaxies for all $\langle {\rm M_*} \rangle$ along the SFMS. From the ALFALFA-based study we find substantially higher mean $f_{HI}$ in groups compared to isolated galaxies (values for pairs being intermediate) for ${\langle{\rm M_*}\rangle}\lesssim10^{9.5}~{\rm M_{\odot}}$, indicating the presence of substantial amounts of HI not associated with cataloged member galaxies in low mass groups. △ Less

Submitted 10 January, 2022; originally announced January 2022.

Comments: Accepted for publication in ApJ. Main text: 26 pages, 16 figures, 6 tables

Robust Estimation of Average Treatment Effects from Panel Data

Authors: Sayoni Roychowdhury, Indrila Ganguly, Abhik Ghosh

Abstract: In order to evaluate the impact of a policy intervention on a group of units over time, it is important to correctly estimate the average treatment effect (ATE) measure. Due to lack of robustness of the existing procedures of estimating ATE from panel data, in this paper, we introduce a robust estimator of the ATE and the subsequent inference procedures using the popular approach of minimum densit… ▽ More In order to evaluate the impact of a policy intervention on a group of units over time, it is important to correctly estimate the average treatment effect (ATE) measure. Due to lack of robustness of the existing procedures of estimating ATE from panel data, in this paper, we introduce a robust estimator of the ATE and the subsequent inference procedures using the popular approach of minimum density power divergence inference. Asymptotic properties of the proposed ATE estimator are derived and used to construct robust test statistics for testing parametric hypotheses related to the ATE. Besides asymptotic analyses of efficiency and powers, extensive simulation studies are conducted to study the finite-sample performances of our proposed estimation and testing procedures under both pure and contaminated data. The robustness of the ATE estimator is further investigated theoretically through the influence functions analyses. Finally our proposal is applied to study the long-term economic effects of the 2004 Indian Ocean earthquake and tsunami on the (per-capita) gross domestic products (GDP) of five mostly affected countries, namely Indonesia, Sri Lanka, Thailand, India and Maldives. △ Less

Submitted 18 December, 2022; v1 submitted 25 December, 2021; originally announced December 2021.

Comments: To appear in 'Statistical Papers'

A robust prediction from a minimal model of COVID-19 -- Can we avoid the third wave?

Authors: Sourav Chowdhury, Suparna Roychowdhury, Indranath Chaudhuri

Abstract: COVID-19 pandemic is one of the major disasters that humanity has ever faced. In this paper, we try to model the effect of vaccination in controlling the pandemic, particularly in context to the third wave which is predicted to hit globally. Here we have modified the SEIRD model by introducing a vaccination term. One of our main assumptions is that the infection rate (\b{eta}(t)) is oscillatory. T… ▽ More COVID-19 pandemic is one of the major disasters that humanity has ever faced. In this paper, we try to model the effect of vaccination in controlling the pandemic, particularly in context to the third wave which is predicted to hit globally. Here we have modified the SEIRD model by introducing a vaccination term. One of our main assumptions is that the infection rate (\b{eta}(t)) is oscillatory. This oscillatory nature has been discussed earlier in literature with reference to the seasonality of epidemics. However, in our case we invoke this nature of the infection rate (\b{eta}(t)) to model the cyclical behavior of the COVID-19 pandemic within a short period. This study focuses on a minimalistic approach where we have logically deduced that the infection rate (\b{eta}(t)) and the vaccination rate (λ) are the most important parameters while the other parameters can be assumed to be constants throughout the simulation. Finally, we have studied the rich interplay between the infection rate (\b{eta}(t)) and the vaccination rate (λ) on the infectious cases of COVID-19 and made some robust conclusions regarding the global behavior of this pandemic in near future. △ Less

Submitted 16 December, 2021; originally announced December 2021.

Comments: 14 pages, 12 figures, accepted for publication in International Journal of Modern Physics C

Representation Learning for Conversational Data using Discourse Mutual Information Maximization

Authors: Bishal Santra, Sumegh Roychowdhury, Aishik Mandal, Vasu Gurram, Atharva Naik, Manish Gupta, Pawan Goyal

Abstract: Although many pretrained models exist for text or images, there have been relatively fewer attempts to train representations specifically for dialog understanding. Prior works usually relied on finetuned representations based on generic text representation models like BERT or GPT-2. But such language modeling pretraining objectives do not take the structural information of conversational text into… ▽ More Although many pretrained models exist for text or images, there have been relatively fewer attempts to train representations specifically for dialog understanding. Prior works usually relied on finetuned representations based on generic text representation models like BERT or GPT-2. But such language modeling pretraining objectives do not take the structural information of conversational text into consideration. Although generative dialog models can learn structural features too, we argue that the structure-unaware word-by-word generation is not suitable for effective conversation modeling. We empirically demonstrate that such representations do not perform consistently across various dialog understanding tasks. Hence, we propose a structure-aware Mutual Information based loss-function DMI (Discourse Mutual Information) for training dialog-representation models, that additionally captures the inherent uncertainty in response prediction. Extensive evaluation on nine diverse dialog modeling tasks shows that our proposed DMI-based models outperform strong baselines by significant margins. △ Less

Submitted 3 May, 2022; v1 submitted 4 December, 2021; originally announced December 2021.

Comments: Preprint, 15 pages, To appear in NAACL 2022 (Main)

In situ process quality monitoring and defect detection for direct metal laser melting

Authors: Sarah Felix, Saikat Ray Majumder, H. Kirk Mathews, Michael Lexa, Gabriel Lipsa, Xiaohu Ping, Subhrajit Roychowdhury, Thomas Spears

Abstract: Quality control and quality assurance are challenges in Direct Metal Laser Melting (DMLM). Intermittent machine diagnostics and downstream part inspections catch problems after undue cost has been incurred processing defective parts. In this paper we demonstrate two methodologies for in-process fault detection and part quality prediction that can be readily deployed on existing commercial DMLM sys… ▽ More Quality control and quality assurance are challenges in Direct Metal Laser Melting (DMLM). Intermittent machine diagnostics and downstream part inspections catch problems after undue cost has been incurred processing defective parts. In this paper we demonstrate two methodologies for in-process fault detection and part quality prediction that can be readily deployed on existing commercial DMLM systems with minimal hardware modification. Novel features were derived from the time series of common photodiode sensors along with standard machine control signals. A Bayesian approach attributes measurements to one of multiple process states and a least squares regression model predicts severity of certain material defects. △ Less

Submitted 3 December, 2021; originally announced December 2021.

Comments: 16 pages, 4 figures

Journal ref: Sci Rep 12, 8503 (2022)

AutoDC: Automated data-centric processing

Authors: Zac Yung-Chun Liu, Shoumik Roychowdhury, Scott Tarlow, Akash Nair, Shweta Badhe, Tejas Shah

Abstract: AutoML (automated machine learning) has been extensively developed in the past few years for the model-centric approach. As for the data-centric approach, the processes to improve the dataset, such as fixing incorrect labels, adding examples that represent edge cases, and applying data augmentation, are still very artisanal and expensive. Here we develop an automated data-centric tool (AutoDC), si… ▽ More AutoML (automated machine learning) has been extensively developed in the past few years for the model-centric approach. As for the data-centric approach, the processes to improve the dataset, such as fixing incorrect labels, adding examples that represent edge cases, and applying data augmentation, are still very artisanal and expensive. Here we develop an automated data-centric tool (AutoDC), similar to the purpose of AutoML, aims to speed up the dataset improvement processes. In our preliminary tests on 3 open source image classification datasets, AutoDC is estimated to reduce roughly 80% of the manual time for data improvement tasks, at the same time, improve the model accuracy by 10-15% with the fixed ML code. △ Less

Submitted 22 November, 2021; originally announced November 2021.

Comments: NeurIPS 2021- Data-Centric AI (DCAI) workshop

QU-net++: Image Quality Detection Framework for Segmentation of Medical 3D Image Stacks

Authors: Sohini Roychowdhury

Abstract: Automated segmentation of pathological regions of interest aids medical image diagnostics and follow-up care. However, accurate pathological segmentations require high quality of annotated data that can be both cost and time intensive to generate. In this work, we propose an automated two-step method that detects a minimal image subset required to train segmentation models by evaluating the qualit… ▽ More Automated segmentation of pathological regions of interest aids medical image diagnostics and follow-up care. However, accurate pathological segmentations require high quality of annotated data that can be both cost and time intensive to generate. In this work, we propose an automated two-step method that detects a minimal image subset required to train segmentation models by evaluating the quality of medical images from 3D image stacks using a U-net++ model. These images that represent a lack of quality training can then be annotated and used to fully train a U-net-based segmentation model. The proposed QU-net++ model detects this lack of quality training based on the disagreement in segmentations produced from the final two output layers. The proposed model isolates around 10% of the slices per 3D image stack and can scale across imaging modalities to segment cysts in OCT images and ground glass opacity (GGO) in lung CT images with Dice scores in the range 0.56-0.72. Thus, the proposed method can be applied for cost effective multi-modal pathology segmentation tasks. △ Less

Submitted 12 April, 2022; v1 submitted 27 October, 2021; originally announced October 2021.

Comments: 4 pages, 7 figures, 1 Table

Journal ref: IEEE EMBC, 2022

Video-Data Pipelines for Machine Learning Applications

Authors: Sohini Roychowdhury, James Y. Sato

Abstract: Data pipelines are an essential component for end-to-end solutions that take machine learning algorithms to production. Engineering data pipelines for video-sequences poses several challenges including isolation of key-frames from video sequences that are high quality and represent significant variations in the scene. Manual isolation of such quality key-frames can take hours of sifting through ho… ▽ More Data pipelines are an essential component for end-to-end solutions that take machine learning algorithms to production. Engineering data pipelines for video-sequences poses several challenges including isolation of key-frames from video sequences that are high quality and represent significant variations in the scene. Manual isolation of such quality key-frames can take hours of sifting through hours worth of video data. In this work, we present a data pipeline framework that can automate this process of manual frame sifting in video sequences by controlling the fraction of frames that can be removed based on image quality and content type. Additionally, the frames that are retained can be automatically tagged per sequence, thereby simplifying the process of automated data retrieval for future ML model deployments. We analyze the performance of the proposed video-data pipeline for versioned deployment and monitoring for object detection algorithms that are trained on outdoor autonomous driving video sequences. The proposed video-data pipeline can retain anywhere between 0.1-20% of the all input frames that are representative of high image quality and high variations in content. This frame selection, automated scene tagging followed by model verification can be completed in under 30 seconds for 22 video-sequences under analysis in this work. Thus, the proposed framework can be scaled to additional video-sequence data sets for automating ML versioned deployments. △ Less

Submitted 15 October, 2021; originally announced October 2021.

Comments: 10 pages, 6 Figures, 5 Tables, conference

Penrose limits in non-Abelian T-dual of Klebanov-Tseytlin Background

Authors: Sourav Roychowdhury, Prasanta K. Tripathy

Abstract: In this paper we consider the Klebanov-Tseytlin background and its non-Abelian T-dual geometry along a suitably chosen $SU(2)$ subgroup of isometries. We analyse the Penrose limits along various null geodesics of both the geometries. We observe that, the Klebanov-Tseytlin geometry does not admit any pp-wave solutions. However, the T-dual background gives rise to pp-wave solution upon taking the Pe… ▽ More In this paper we consider the Klebanov-Tseytlin background and its non-Abelian T-dual geometry along a suitably chosen $SU(2)$ subgroup of isometries. We analyse the Penrose limits along various null geodesics of both the geometries. We observe that, the Klebanov-Tseytlin geometry does not admit any pp-wave solutions. However, the T-dual background gives rise to pp-wave solution upon taking the Penrose limit along some appropriate null geodesic. We comment on the possible gauge theory dual for our pp-wave background. △ Less

Submitted 11 November, 2021; v1 submitted 19 September, 2021; originally announced September 2021.

Comments: 1+32 pages; v2; Supersymmetry discussion added; references added; Accepted to Phys. Rev. D

Journal ref: Phys. Rev. D. 104, 126016 (2021)

Video2Skill: Adapting Events in Demonstration Videos to Skills in an Environment using Cyclic MDP Homomorphisms

Authors: Sumedh A Sontakke, Sumegh Roychowdhury, Mausoom Sarkar, Nikaash Puri, Balaji Krishnamurthy, Laurent Itti

Abstract: Humans excel at learning long-horizon tasks from demonstrations augmented with textual commentary, as evidenced by the burgeoning popularity of tutorial videos online. Intuitively, this capability can be separated into 2 distinct subtasks - first, dividing a long-horizon demonstration sequence into semantically meaningful events; second, adapting such events into meaningful behaviors in one's own… ▽ More Humans excel at learning long-horizon tasks from demonstrations augmented with textual commentary, as evidenced by the burgeoning popularity of tutorial videos online. Intuitively, this capability can be separated into 2 distinct subtasks - first, dividing a long-horizon demonstration sequence into semantically meaningful events; second, adapting such events into meaningful behaviors in one's own environment. Here, we present Video2Skill (V2S), which attempts to extend this capability to artificial agents by allowing a robot arm to learn from human cooking videos. We first use sequence-to-sequence Auto-Encoder style architectures to learn a temporal latent space for events in long-horizon demonstrations. We then transfer these representations to the robotic target domain, using a small amount of offline and unrelated interaction data (sequences of state-action pairs of the robot arm controlled by an expert) to adapt these events into actionable representations, i.e., skills. Through experiments, we demonstrate that our approach results in self-supervised analogy learning, where the agent learns to draw analogies between motions in human demonstration data and behaviors in the robotic environment. We also demonstrate the efficacy of our approach on model learning - demonstrating how Video2Skill utilizes prior knowledge from human demonstration to outperform traditional model learning of long-horizon dynamics. Finally, we demonstrate the utility of our approach for non-tabula rasa decision-making, i.e, utilizing video demonstration for zero-shot skill generation. △ Less

Submitted 9 September, 2021; v1 submitted 8 September, 2021; originally announced September 2021.

Towards a Safety Case for Hardware Fault Tolerance in Convolutional Neural Networks Using Activation Range Supervision

Authors: Florian Geissler, Syed Qutub, Sayanta Roychowdhury, Ali Asgari, Yang Peng, Akash Dhamasia, Ralf Graefe, Karthik Pattabiraman, Michael Paulitsch

Abstract: Convolutional neural networks (CNNs) have become an established part of numerous safety-critical computer vision applications, including human robot interactions and automated driving. Real-world implementations will need to guarantee their robustness against hardware soft errors corrupting the underlying platform memory. Based on the previously observed efficacy of activation clipping techniques,… ▽ More Convolutional neural networks (CNNs) have become an established part of numerous safety-critical computer vision applications, including human robot interactions and automated driving. Real-world implementations will need to guarantee their robustness against hardware soft errors corrupting the underlying platform memory. Based on the previously observed efficacy of activation clipping techniques, we build a prototypical safety case for classifier CNNs by demonstrating that range supervision represents a highly reliable fault detector and mitigator with respect to relevant bit flips, adopting an eight-exponent floating point data representation. We further explore novel, non-uniform range restriction methods that effectively suppress the probability of silent data corruptions and uncorrectable errors. As a safety-relevant end-to-end use case, we showcase the benefit of our approach in a vehicle classification scenario, using ResNet-50 and the traffic camera data set MIOVision. The quantitative evidence provided in this work can be leveraged to inspire further and possibly more complex CNN safety arguments. △ Less

Submitted 16 August, 2021; originally announced August 2021.

Comments: 8 pages, 7 figures

Report number: ISSN 1613-0073

Journal ref: Proceedings of the Workshop on Artificial Intelligence Safety 2021

WALLABY pre-pilot survey: Two dark clouds in the vicinity of NGC 1395

Authors: O. Ivy Wong, A. R. H. Stevens, B. -Q. For, T. Westmeier, M. Dixon, S. -H. Oh, G. I. G. Józsa, T. N. Reynolds, K. Lee-Waddell, J. Román, L. Verdes-Montenegro, H. M. Courtois, D. Pomarède, C. Murugeshan, M. T. Whiting, K. Bekki, F. Bigiel, A. Bosma, B. Catinella, H. Dénes, A. Elagali, B. W. Holwerda, P. Kamphuis, V. A. Kilborn, D. Kleiner , et al. (12 additional authors not shown)

Abstract: We present the Australian Square Kilometre Array Pathfinder (ASKAP) WALLABY pre-pilot observations of two `dark' HI sources (with HI masses of a few times 10^8 Msol and no known stellar counterpart) that reside within 363 kpc of NGC 1395, the most massive early-type galaxy in the Eridanus group of galaxies. We investigate whether these `dark' HI sources have resulted from past tidal interactions o… ▽ More We present the Australian Square Kilometre Array Pathfinder (ASKAP) WALLABY pre-pilot observations of two `dark' HI sources (with HI masses of a few times 10^8 Msol and no known stellar counterpart) that reside within 363 kpc of NGC 1395, the most massive early-type galaxy in the Eridanus group of galaxies. We investigate whether these `dark' HI sources have resulted from past tidal interactions or whether they are an extreme class of low surface brightness galaxies. Our results suggest that both scenarios are possible, and not mutually exclusive. The two `dark' HI sources are compact, reside in relative isolation and are more than 159 kpc away from their nearest HI-rich galaxy neighbour. Regardless of origin, the HI sizes and masses of both `dark' HI sources are consistent with the HI size-mass relationship that is found in nearby low-mass galaxies, supporting the possibility that these HI sources are an extreme class of low surface brightness galaxies. We identified three analogues of candidate primordial `dark' HI galaxies within the TNG100 cosmological, hydrodynamic simulation. All three model analogues are dark matter-dominated, have assembled most of their mass 12-13 Gyr ago, and have not experienced much evolution until cluster infall 1-2 Gyr ago. Our WALLABY pre-pilot science results suggest that the upcoming large area HI surveys will have a significant impact on our understanding of low surface brightness galaxies and the physical processes that shape them. △ Less

Submitted 9 August, 2021; originally announced August 2021.

Comments: 16 pages, 11 figures, accepted for publication in MNRAS