AI Paper Atlas

Research map / Video / Video-language understanding

Learning to Deny: Action Denial in Multimodal Large Language Models

Raiyaan Abdullah, Shehreen Azad, Yogesh Rawat

ECCV 2026 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
preserve context but remove defining motion (action-denial negatives) vs action-presence clips20 MLLMs · UCF101-ADaccuracy (%): Action-Presence 97.5 -> Overall-AD 34.1 (Ovis2.5-9B)Hurt
replace generic None option with explicit denial statement (ablation)Ovis2.5-9B · UCF101-AD Overall-ADaccuracy (%): 34.1 -> 74.7Improved
remove the primary distractor from the options (ablation)Ovis2.5-9B · UCF101-AD Overall-ADaccuracy (%): 34.1 -> 93.9Improved
reduce answer options from 11 to 4 (ablation)8 high-performing MLLMs · UCF101-AD Overall-ADaccuracy (%): VideoLLaMA3-7B 51.5 -> 43.7; Qwen2.5-VL-7B 25.1 -> 30.5Mixed
use reasoning (thinking) variantsMLLMs · UCF101-AD Overall-ADaccuracy (%): Ovis2.5-9B 34.1 -> 40.4; Kimi-VL 29.6 -> 20.9Mixed
increase model scale (ablation)Qwen2.5-VL family · UCF101-AD Overall-ADaccuracy (%): 19.8 (3B) -> 25.1 (7B) -> 45.7 (72B)Improved
add CausalAct-0 causal graph promptOvis2.5-9B · UCF101, K400, HMDB51, SSv2, Diving48, FineGym99 denialdenial accuracy (%): HMDB51 51.2 -> 74.3; SSv2 16.0 -> 37.4; UCF101 78.7 -> 86.1Improved
add CausalAct zero-shot promptVideoLLaMA3 · UCF101-ADOverall-ADHurt
finetune small MLLMs on graph-based VQA tasks (CausalAct)2B-3B MLLMs · UCF101-AD Action-Denial testaccuracyImproved
use pruned or random graph instead of full causal graph (ablation)CausalAct finetuned · UCF101-ADaccuracyHurt
use pruned or random graph in zero-shot (ablation)CausalAct-0 · UCF101-ADaccuracyNo clear effect
freeze vision encoder during graph finetuning (ablation)CausalAct finetuning · UCF101-ADaccuracy: about -15% on averageHurt

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