Pooyan Fazli

EgoP $^{2}$ C: Internalizing Plan-Perceive-Compose for Spatial Reasoning in Egocentric Video

Chaoyu Li, Arka Sadhu, Yuliang Li +3

Answering complex spatial reasoning questions from egocentric video requires determining what evidence is needed, acquiring targeted observations, and combining them through multip…

Video Generation Models: A Survey of Post-Training and Alignment

Chaoyu Li, Xiaoyi Gu, Yogesh Kulkarni +10

Video generation has rapidly progressed from short, low-quality clips to high-resolution, long-duration sequences with complex spatiotemporal dynamics. Despite strong generative pr…

Video-FLAIR: Not Whether to Reason, But How

Yogesh Kulkarni, Pooyan Fazli

Multimodal queries can require different types of reasoning. Some can be answered via perceptual reasoning, extracting information directly from the visual signal, while others req…

CASHEW: Stabilizing Multimodal Reasoning via Iterative Trajectory Aggregation

Chaoyu Li, Fei Tao, Pooyan Fazli

Vision-language models achieve strong performance across a wide range of multimodal understanding and reasoning tasks, yet their multi-step reasoning remains unstable. Repeated sam…

EgoVITA: Learning to Plan and Verify for Egocentric Video Reasoning

Yogesh Kulkarni, Pooyan Fazli

Egocentric video understanding requires procedural reasoning under partial observability and continuously shifting viewpoints. Current multimodal large language models (MLLMs) stru…

FrameOracle: Learning What to See and How Much to See in Videos

Chaoyu Li, Tianzhi Li, Fei Tao +6

Vision-language models (VLMs) advance video understanding but operate under tight computational budgets, making performance dependent on selecting a small, high-quality subset of f…

AVATAR: Reinforcement Learning to See, Hear, and Reason Over Video

Yogesh Kulkarni, Pooyan Fazli

Multimodal reasoning over long-horizon video is challenging due to the need for precise spatiotemporal fusion and alignment across modalities. While recent methods such as Group Re…

ReGATE: Learning Faster and Better with Fewer Tokens in MLLMs

Chaoyu Li, Yogesh Kulkarni, Pooyan Fazli

The computational cost of training multimodal large language models (MLLMs) grows rapidly with the number of processed tokens. Existing efficiency methods mainly target inference v…

VideoPASTA: 7K Preference Pairs That Matter for Video-LLM Alignment

Yogesh Kulkarni, Pooyan Fazli

Video-language models (Video-LLMs) excel at understanding video content but struggle with spatial relationships, temporal ordering, and cross-frame continuity. To address these lim…

ChartQA-X: Generating Explanations for Visual Chart Reasoning

Shamanthak Hegde, Pooyan Fazli, Hasti Seifi

The ability to explain complex information from chart images is vital for effective data-driven decision-making. In this work, we address the challenge of generating detailed expla…

VideoA11y: Method and Dataset for Accessible Video Description

Chaoyu Li, Sid Padmanabhuni, Maryam Cheema +2

Video descriptions are crucial for blind and low vision (BLV) users to access visual content. However, current artificial intelligence models for generating descriptions often fall…

VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding

Chaoyu Li, Eun Woo Im, Pooyan Fazli

Multimodal large language models (MLLMs) have recently shown significant advancements in video understanding, excelling in content reasoning and instruction-following tasks. Howeve…

VideoSAVi: Self-Aligned Video Language Models without Human Supervision

Yogesh Kulkarni, Pooyan Fazli

Recent advances in video-large language models (Video-LLMs) have led to significant progress in video understanding. Current preference optimization methods often rely on proprieta…

VidComposition: Can MLLMs Analyze Compositions in Compiled Videos?

Yolo Y. Tang, Junjia Guo, Hang Hua +9

The advancement of Multimodal Large Language Models (MLLMs) has enabled significant progress in multimodal understanding, expanding their capacity to analyze video content. However…