Sourav S. Bhabesh

Research focus

Multimodal Models (2)Reasoning & Chain-of-Thought (2)RLHF & Preference Learning (1)Code Generation & Program Synthesis (1)

Frequent co-authors

D. Bespalov (2)Shiyu Yuan (1)Wesley Rose (1)H. Rangwala (1)

Papers (2)

Jul 15, 2026

Shiyu Yuan +4Jul 15, 2026

SD-MAR: Multi-image Analytical Reasoning via Synthetic Data and Reinforcement Learning

VLMs can achieve up to 36.95% higher accuracy in multi-image analytical reasoning tasks, outperforming GPT-4.1 by leveraging synthetic data and a novel reinforcement learning strategy.

Shiyu Yuan, Sourav S. Bhabesh, D. Bespalov +2

Multimodal Models Reasoning & Chain-of-Thought RLHF & Preference Learning

Jul 9, 2026

Yidong Ouyang +3Jul 9, 2026·also Amazon Science

Reinforcing the Generation Order of Multimodal Masked Diffusion Models

Learning the generation order in multimodal tasks can boost performance by over 4%—a game changer for DLMs.

Yidong Ouyang, Zhe Wang, Sourav S. Bhabesh +1

Code Generation & Program Synthesis Multimodal Models Reasoning & Chain-of-Thought

Search

Sourav S. Bhabesh

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)