Kulin Shah

I am a Ph.D. student in Computer Science department at the University of Texas Austin advised by Prof. Adam Klivans. My research interest broadly lies in generative models. I like to understand (empirically and theoretically) building blocks of modern generative models with the hope of improving them.

I was a student researcher at Google with Rina Panigrahy working on understanding and improving reasoning capabilities of language models during June 2023-March 2024. Before joining Ph.D., I was a research fellow at Microsoft Research India, where I worked with Amit Deshpande and Navin Goyal. I graduated with my B.Tech. (Hons.) in Computer Science from IIIT-Hyderabad, where I worked with Naresh Manwani.


Papers

(α-β denotes alphabetical order and * denotes equal contribution)



Train for the Worst, Plan for the Best: Understanding Token Ordering in Masked Diffusions PDF
Jaeyeon Kim*, Kulin Shah*, Vasilis Kontonis, Sham M. Kakade, Sitan Chen
Preprint

Does Generation Require Memorization? Creative Diffusion using Ambient Diffusion PDF Code
Kulin Shah, Alkis Kalavasis, Giannis Daras, Adam Klivans
Preprint

Causal Language Modeling can Elicit Search and Reasoning Capabilities on Puzzles PDF Code
Kulin Shah, Nishanth Dikkala, Xin Wang, Rina Panigrahy NeurIPS 2024 | Conference on Neural Information Processing Systems

Unrolled denoising networks provably learn optimal Bayesian inference PDF
Aayush Karan*, Kulin Shah*, Sitan Chen, Yonina C. Eldar NeurIPS 2024 | Conference on Neural Information Processing Systems

Learning general Gaussian mixtures with efficient score matching PDF
(α-β) Sitan Chen, Vasilis Kontonis, Kulin Shah
Preprint

Simple Mechanisms for Representing, Indexing and Manipulating Concepts PDF
(α-β)Yuanzhi Li, Raghu Meka, Rina Panigrahy, Kulin Shah
Preprint

Learning Mixtures of Gaussians Using the DDPM Objective PDF Poster
Kulin Shah, Sitan Chen, Adam Klivans
NeurIPS 2023 | Conference on Neural Information Processing Systems

Ambient Diffusion: Learning Clean Distributions From Corrupted Data PDF Code
Giannis Daras, Kulin Shah, Yuval Dagan, Aravind Gollakota, Alexandros G. Dimakis, Adam Klivans NeurIPS 2023 | Conference on Neural Information Processing Systems

Learning and Generalization in Overparameterized Normalizing Flows PDF Poster Code
Kulin Shah, Amit Deshpande, Navin Goyal AISTATS 2022 | International Conference on Artificial Intelligence and Statistics

RISAN: Robust Instance Specific Deep Abstention Network PDF
Bhavya Kalra, Kulin Shah, Naresh Manwani
UAI, 2021 | Conference on Uncertainty in Artificial Intelligence
Selected for Oral presentation

Rawlsian Fair Adaptation of Deep Learning Classifiers PDF
Kulin Shah, Pooja Gupta, Amit Deshpande, Chiranjib Bhattacharyya
AIES, 2021 | AAAI/ACM Conference on AI, Ethics, and Society

Online Active Learning of Reject Option Classifiers PDF
Kulin Shah, Naresh Manwani
AAAI, 2020 | AAAI Conference on Artificial Intelligence
Selected for Oral presentation

Sparse Reject Option Classifier using Successive Linear Programming PDF
Kulin Shah, Naresh Manwani
AAAI, 2019 | AAAI Conference on Artificial Intelligence
Selected for Oral presentation

Ingredients for happiness: Modeling constructs via semi-supervised content driven inductive transfer learning PDF
Kulin Shah, Naresh Manwani
AFFCON workshop @ AAAI, 2019 | AAAI-19 Workshop on affective content analysis

PLUME: Polyhedral Learning Using Mixture of Experts PDF
Kulin Shah, PS Sastry, Naresh Manwani
Preprint




Wepage inspired by minimal research theme.

Flag Counter