COMET: Contrastive Motion-Enhanced Temporal Reasoning for Video Multimodal Large Language Models
Open Access
- 9 November 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 16190-16198
Abstract
No abstract availableFunding Information
- National Key Research and Development Program of China (2024YFE0203100)
- Guangdong Provincial Key Laboratory of Ultra High Definition Immersive Media Technology (2024B1212010006)
- Shenzhen Science and Technology Program (JCYJ20230807120800001)
This publication has 9 references indexed in Scilit:
- Flow4Agent: Long-form Video Understanding via Motion Prior from Optical Flow2025
- DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learningNature, 2025
- Video-R1: Reinforcing Video Reasoning in MLLMs2025
- SAMFlow: Eliminating Any Fragmentation in Optical Flow with Segment Anything ModelProceedings of the AAAI Conference on Artificial Intelligence, 2024
- Perception Test: A Diagnostic Benchmark for Multimodal Video Models2023
- NExT-QA: Next Phase of Question-Answering to Explaining Temporal Actions2021
- TSM: Temporal Shift Module for Efficient Video Understanding2019
- The “Something Something” Video Database for Learning and Evaluating Visual Common Sense2017
- Convolutional Two-Stream Network Fusion for Video Action Recognition2016