AI Paper Atlas

Research map / Video / Video-language understanding

Diagnosing and Correcting Bias in MLLM for Long Video Understanding

Xusheng Liang, Jianqiao Sun, Hao Zhang, Yulei Niu, Hengshuang Zhao, Jiawei Ma

NeurIPS 2026 · Video › Video-language understanding · approach: Long video understanding · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Studies bias in MLLMs for long-video question answering under sparse frame sampling and proposes a correction method (details unavailable beyond abstract).

Problem

Long-video QA where answers depend on a few decisive moments, but MLLMs must sample frames sparsely due to memory limits, risking omission of key evidence.

Approach

Argues that simply enlarging context size does not fundamentally fix performance, and diagnoses a bias in MLLMs for long video understanding, then corrects it; the specific correction is not stated in the abstract.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add CLEAR counterfactual residual steeringQwen3-VL-8B · MLVUaccuracy: 59.0 -> 66.8Improved
add CLEAR counterfactual residual steeringInternVL3-8B · LVBenchaccuracy: 43.4 -> 49.7Improved
add CLEAR counterfactual residual steeringQwen3-VL-8B · S-MME overallShortErr: 44.0 -> 36.9Improved
add CLEAR counterfactual residual steeringLLaVA-Video-7B · S-MME overallaccuracy: 9.5 -> 12.7Improved
use both question-only and salient-view probes instead of only the question-only probe (ablation)CLEAR on Qwen3-VL-8B · S-MME overallaccuracy: 23.3 -> 26.5Improved
use both probes instead of only the salient-view probe (ablation)CLEAR on Qwen3-VL-8B · S-MME overallaccuracy: 23.7 -> 26.5Improved
add question-only probe alone (ablation)Qwen3-VL-8B · S-MME linguistic-bias casesaccuracy: 22.2 -> 26.1Improved
add salient-view probe alone (ablation)Qwen3-VL-8B · S-MME visual-bias casesaccuracy: 18.8 -> 22.9Improved
raise residual strength alpha from 0.15 to 0.35 (ablation)CLEAR on Qwen3-VL-8B · S-MME overallaccuracy: 26.5 -> 24.8Hurt
lower residual strength alpha from 0.15 to 0.05 (ablation)CLEAR on Qwen3-VL-8B · S-MME overallaccuracy: 26.5 -> 23.4Hurt
scale frames from 64 to 128 without CLEAR (ablation)Qwen3-VL-8B · Video-MME overallaccuracy: 65.1 -> 67.0Improved
add CLEAR at 32 frames versus scaling to 128 frames (ablation)Qwen3-VL-8B · Video-MME overallaccuracy: 67.0 -> 68.4Improved

Surprising findings

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy