Diagnosing and Correcting Bias in MLLM for Long Video Understanding
arXiv:2610.36210v1 Announce Type: new Abstract: Long-video question answering is challenging since the answer often hinges on a few decisive moments scattered throughout the video, while memory constraints force Multimodal Large Language Models (MLLMs) to sample frames at extremely low rates, resul…
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-10-28 04:00 · arXiv cs.CV
Diagnosing and Correcting Bias in MLLM for Long Video Understanding