Sahil Garg, PhD
Machine Learning Researcher at Morgan Stanley

Smiley face

Research Interests

I enjoy conducting problem driven research. In the past 13 years, I had the privilege of collaborating with many researchers from diverse fields, and accordingly, some of the research interests are listed as below.

Deep Neural Nets

Generative AI, out-of-distribution detection, continual learning, domain adaptation (distributional invariance), information-theoretic clustering or kNN search, transfer entropy or mutual information estimation, robust & interpretative classification, scaling laws for DNNs, highly noisy or irregular timeseries analysis, curriculum learning, multi-scale anomaly (fraud) detection in database systems.

Natural Language Processing

Relation extraction for cancer biology, dialog modeling for psychiatry, abstract meaning representations, discourse analysis for measuring thought disorder & complexity, metaphor detection, interpretative document summary with LLMs, curriculum pretraining of LLMs, OOD detection in LLMs.

Machine Learning

Nonstationary kernels, informative sensing, locality sensitive hashing for representation learning, convolution kernels for text, information theoretic representation learning, nearly unsupervised, attention mechanism in kernels via nonstationarity, interpretative, data-efficiency, neuro-inspired, sparse modeling, network science.


Employment and Education

At present, I am enjoying ML research at Morgan Stanley, continuing my passion for problem-driven research. Previously, I explored the field of computational psychiatry as a postdoctoral fellow in the Icahn School of Medicine at Mount Sinai under the mentorship of Prof. Cheryl Corcoran and Dr. Guillermo Cecchi. Prior to that, I was advised by Prof. Aram Galstyan for my PhD thesis at USC. In the past life, I had the privilege of learning under the guidance of Prof. Nora Ayanian at USC, Prof. Amarjeet Singh at IIIT Delhi, and Prof. Fabio Ramos.

Selected Papers

Deep Generative Sampling in the Dual Divergence Space: A Data-efficient & Interpretative Approach for Generative AI. Garg et al. 2024. Preprint PDF.

Empowering Time Series Analysis with Large Language Models: A Survey. Jiang et al. 2024. Preprint PDF

Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting. Rasul et al. NeurIPS 2023 Workshop: R0-FoMo. PDF, Code.

In- or Out-of-Distribution Detection via Dual Divergence Estimation. Garg et al. UAI-23. PDF, Code.

Information Theoretic Clustering via Divergence Maximization among Clusters. Garg et al. UAI-23. PDF, Code.

Estimating Transfer Entropy under Long Ranged Dependencies. Garg et al. UAI-22. PDF, Code.

Increased Metaphor Production in Open-Ended Speech Samples of Patients With Prodromal and Developed Schizophrenia Detected with NLP. Srivastava et al. Biological Psychiatry, 2022.

Negative symptoms and speech pauses in youths at clinical high risk for psychosis. Stanislawski et al. NPJ Schizophrenia, 2021. PDF

NERO: a biomedical named-entity (recognition) ontology with a large, annotated corpus reveals meaningful associations through text embedding. Wang et al. NPJ Systems Biology and Applications, 2021. PDF

Linking language features to clinical symptoms and multimodal imaging in individuals at clinical high risk for psychosis. Haas et al. European Psychiatry, 2020. PDF

Modeling Dialogues with Hashcode Representations: A Nonparametric Approach. Garg et al. AAAI-20. PDF.

Nearly-Unsupervised Hashcode Representations for Relation Extraction. Garg et al. EMNLP-19. PDF, Code.

Kernelized Hashcode Representations for Relation Extraction. Garg et al. AAAI-19. PDF, Code.

Stochastic Learning of Nonstationary Kernels for Natural Language Modeling. Garg et al. 2017. PDF.

Neurogenesis-Inspired Dictionary Learning: Online Model Adaption in a Changing World. Sahil Garg*, Irina Rish, Guillermo Cecchi, Aurelie Lozano. IJCAI-17. PDF, Code.

Extracting Biopathway Interactions using Semantic Parsing of Biomedical Text. Garg et al. AAAI-16. PDF, Code.

Persistent Monitoring of Stochastic Spatio-temporal Phenomena with a Small Team of Robots. Garg et al. RSS-14. PDF.

Learning Nonstationary Space-Time Models for Environmental Monitoring. Garg et al. AAAI-12. PDF, Code.

See Google Scholar for an extended list of the papers.

This site is text-positive and defiantly retro (hand-crafted HTML 1.0).

sahil.garg.cs at gmail.com