
Courtesy: Apple Intelligence
Somjit Roy
Ph.D. in Statistics
Department of Statistics, Texas A&M University
-----
SciML || Bayesian Modeling & Computation
-----

Courtesy: Apple Intelligence
Ph.D. in Statistics
Department of Statistics, Texas A&M University
-----
SciML || Bayesian Modeling & Computation
-----
Our paper on Almost Perfect Mutually Unbiased Bases that are Sparse got published in Journal of Statistical Theory and Practice.
Our abstract on TAVIE: Tangent Approximation for Variational Inference in different Exponential Families got selected in IISA Conference 2024, for a talk in the contributed session.
Selected as the Workflow Workshop Organizer in Statistics Graduate Student Association (SGSA), for the academic year 2024-2025.
Selected to attend the Summer School on Computational Materials Science (CMS3), organized by TAMU Materials Science and Engineering Department, TAMU HPRC and National Science Foundation (NSF).
Successfully passed the compulsory first-year Qualifying Examinations (QE) in Ph.D. Statistics @Texas A&M Statistics.
Successfully passed the compulsory first-year Qualifying Examinations (QE) in Ph.D. Statistics @Texas A&M Statistics.
Our work on Almost Perfect Mutually Unbiased Bases that are Sparse has been arXived.
Collaboration with Arroyave Lab @Texas A&M Materials Science, working on TPT grant towards Bayesian methodologies in autonomous materials discovery.
Joined the doctoral (Ph.D.) program in Statistics at Department of Statistics, Texas A&M University.
Recipient of the R.C. Bose Memorial Book Prize, granted by Calcutta Statistical Association for securing the highest marks in first year of MSc. Statistics in Department of Statistics, University of Calcutta.
Received the Science Academies Summer Research Fellowship, jointly granted by Indian Academy of Science (IASc), Indian National Science Academy (INSA) and National Academy of Sciences (NASI).
Awarded the OPHI Summer School grant by Oxford Poverty and Human Development Initiative (OPHI), University of Oxford.
Awarded the IAOS 2022 Travel & Conference grant by IAOS, Eurostat, Statistics Poland, World Bank and UNSD.
Our work on A Heuristic Framework to Search for Approximate Mutually Unbiased Bases got published in CSCML 2022.
I'm a second-year Ph.D. student in Statistics at Department of Statistics, Texas A&M University, where I spend my days wrangling complex data, building fancy models, and convincing my laptop that Bayesian inference is, in fact, solvable. I'm fortunate to be advised by Dr. Bani K. Mallick (TAMU Statistics) and co-advised by Dr. Debdeep Pati (UW-Madison Statistics)—who ensures my model, related theory and code makes more sense than my coffee-fueled late-night ideas.
My research primarily focuses on Statistical Modeling and Computation, where I develop both theoretical and methodological frameworks to analyze complex real-world data. My work spans Scientific Machine Learning (SciML), Bayesian Machine Learning algorithms, Variational Inference (VI), Bayesian Optimization (BO), Tree & Graphical Modeling, and Symbolic Regression.
I also actively collaborate with the Computational Materials Science Lab in the Department of Materials Science and Engineering, Texas A&M University. Our research integrates SciML and Bayesian Optimization within Computational Materials Science, with the broader goal of advancing autonomous materials discovery.
Bayesian Machine Learning (ML) algorithms for modeling complex real-world data arising from multidisciplinary fields.
Variational Inference (VI; popularly known as Variational Bayes (VB)) for approximate Bayesian inference.
Theoretical and Methodological underpinnings of Bayesian Tree & Graphical models aimed at predictive modeling.
Variable Selection for identifying important variables through appropriate selection priors, while accounting for uncertainty quantification and model complexity.
Bayesian Optimization (BO), Subspace Reduction methods and Gaussian Process priors towards autonomous materials discovery.
Exploring structures like Mutually Unbiased Bases (MUBs) within the domain of Quantum Cryptography by using combinatorial designs and heuristic search algorithms.
A fully Bayesian model-based framework for the identification of physicochemically meaningful descriptors or materials genes, that are obtained as a composition of primary features (atomic radii, ionization energy and other physical materials' properties) and elementary algebraic operators (+, -, /, x, sin, cos, log, exp). (In Progress).
A joint work with Department of Materials Science and Engineering, TAMU aimed at developing scalable BO techniques for autonomous and optimal materials discovery. (In Progress).
Robust and scalable model-based approach for the Bayesian optimization framework with extensions to Bayesian Additive Regression Trees (BART). (In Progress).
An extension of the tangent-transform variational inference approach to a large class of flexible probability models including linear regression with heavy-tailed errors (i.e., strongly super-Gaussian likelihoods) as well as Bayesian quantile regression. (Submitted).
(2024) Ajeet Kumar, Subhamoy Maitra and Somjit Roy. Almost Perfect Mutually Unbiased Bases that are Sparse. Journal of Statistical Theory and Practice 18, 61 (2024). [Paper].
(2022) Sreejit Chaudhury, Ajeet Kumar, Subhamoy Maitra, Somjit Roy, and Sourav Sen Gupta. A Heuristic Framework to Search for Approximate Mutually Unbiased Bases. In: Dolev, S., Katz, J., Meisels, A. (eds) Cyber Security, Cryptology, and Machine Learning. CSCML 2022. Lecture Notes in Computer Science, vol 13301. Springer, Cham. [Paper].
bayesestdft: Estimating the Degrees of Freedom of the Student's t-Distribution under a Bayesian Framework. Maintainer and Developer: Somjit Roy. Contributor: Se Yoon Lee. Github version.
GoodFitSBM: Monte Carlo goodness-of-fit tests for Stochastic Block Models. Developer: Somjit Roy. Maintainer and Co-developer: Soham Ghosh. Contributor: Dr. Debdeep Pati. Github version.
gamblers.ruin.gameplay:One-Dimensional Random Walks Through Simulation of the Gambler's Ruin Problem. Maintainer and Developer: Somjit Roy.
YatesAlgo.FactorialExp.SR:Yates' Algorithm in 2^n Factorial Experiment. Maintainer and Developer: Somjit Roy. Selected for a talk in the useR regional conference in Basel, Switzerland, July 2023.
 
 
 
 
 
 
 
 
 
 
Following are the courses offered by TAMU Statistics, where I was appointed as a Teaching Assistant (TA).
(Spring 2024) STAT 639: Data Mining and Analysis (Graduate Level). Course Instructor: Dr. Raymond Ka Wai Wong.
(Fall 2023) STAT 636: Applied Multivariate Analysis and Statistical Learning (Graduate Level). Course Instructor: Dr. Raymond Ka Wai Wong.
(Summer 2024 - Present) Holding a Research Assistant (RA) position under the supervision of Dr. Bani K. Mallick in the Department of Statistics, Texas A&M University.
(Sep 2022 - May 2023) Research Fellow in the Department of Mathematics, University of Maryland, College Park under Dr. Partha Lahiri.
(Jun 2022 - Aug 2022) IASc-INSA-NASI Summer Research Fellow in the Department of Mathematics and Statistics, Indian Institute of Science Education and Research (IISER) Kolkata under the supervision of Dr. Anirvan Chakraborty.
(Oct 2021 - May 2022) Research Intern in the Applied Statistics Unit (ASU), Indian Statistical Institute (ISI) Kolkata under the supervision of Dr. Subhamoy Maitra.
(May 2021 - Aug 2021) Data Science Intern in Tata Electronics Pvt. Ltd. under the supervision of Nagasubramanian Kothandaraman and Dr. Subhamoy Maitra.
(Dec 2024) IISA 2024: Contributed Talk on TAVIE: Tangent Approximation for Variational Inference in different Exponential Families.
(Feb 2023) An online meet, United Nations Panel Meeting on Small Area Estimation (SAE): eLearning Launch.
(May 2024) CMS3-FAST: Attended the summer school on Computational Materials Science in Texas A&M University.
(Feb 2023) An online meet, United Nations Panel Meeting on Small Area Estimation (SAE): eLearning Launch.
(Feb 2023) An online talk by Richard Valliant, as a part of the JPSM MPSDS seminar.
(Jan 2023) Attended the golden jubilee celebration conference on Indian Association of Productivity, Quality and Reliability (IAPQR).
(Aug 2022) Attended the summer school on Multidimensional Poverty Measurement and Analysis, organized by Oxford Poverty and Human Development Initiative (OPHI), University of Oxford in Universitas Indonesia, Jakarta..
(Apr 2022) IAOS 2022: Attended the conference organized by IAOS in Krakow, Poland.
(Apr 2019) ExploreML Google Workshop: Attended the workshop, organized by Google Developers in St. Xavier's College, Kolkata.
(2024) IISA 2024: TAMU Statistics Department Student Travel Award and NSF Travel Grant for presenting at IISA 2024.
(2024) CMS3-FAST Summer School: Globally selected to attend the summer school, jointly organized by Department of Materials Science and Engineering, TAMU, TAMU HPRC and National Science Foundation (NSF) at Texas A&M University. [Link]
(2022) R.C. Bose Memorial Book Prize: Granted by Calcutta Statistical Association, for securing the highest marks in MSc. first year in Department of Statistics, University of Calcutta. [Link]
(2022) Science Academies Summer Research Fellowship: Selected to be awarded the IASc-INSA-NASI Summer Research Fellowship in the summer of 2022, granted jointly by Indian Academy of Sciences (IASc), Indian National Science Academy (INSA) and The National Academy of Sciences (NASI). [Link]
(2022) OPHI, University of Oxford - Summer School Grant: Globally selected among 15 participants and funded by OPHI, University of Oxford to attend the summer school in Universitas Indonesia, Jakarta, Indonesia.[Link]
(2022) IAOS 2022 Conference and Travel Grant: Globally selected among 50 participants and funded by United Nations Statistics Division (UNSD), Eurostat, The World Bank and Statistics Poland to attend IAOS 2022 held in Krakow, Poland.[Link]
Coming Soon!
BLOC 463B, Department of Statistics, Texas A&M University, 155 Ireland St., 3143 TAMU, College Station, Texas: 7784, USA.