APPO: Attention-guided Perception Policy Optimization for Video Reasoning

2026-02-27

AI Summary

This work addresses the limitation that current video reasoning performance is primarily constrained by fine-grained perceptual capabilities rather than high-level reasoning, with perception enhancement typically relying on costly fine-grained annotations. To overcome this, the authors propose Attention-guided Perceptual Policy Optimization (APPO), an algorithm that, for the first time, reveals the dominant role of perception in video reasoning. APPO introduces a token-level dense reward mechanism to optimize intra-group perceptual tokens within critical video frames without requiring additional annotations. The method is compatible with models of varying scales (3B/7B) and consistently outperforms GRPO and DAPO across multiple video benchmarks, achieving performance gains of 0.5% to 4%, thereby demonstrating its effectiveness and generalizability.

Technology Category

• Computer Vision: Video Understanding & Activity Analysis
• Search and Optimization: Learning to Search
• Knowledge Representation and Reasoning: Argumentation

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Complex video reasoning, actually, relies excessively on fine-grained perception rather than on expert (e.g., Ph.D, Science)-level reasoning. Through extensive empirical observation, we have recognized the critical impact of perception. In particular, when perception ability is almost fixed, enhancing reasoning from Qwen3-8B to OpenAI-o3 yields only 0.7% performance improvement. Conversely, even minimal change in perception model scale (from 7B to 32B) boosts performance by 1.4%, indicating enhancing perception, rather than reasoning, is more critical to improve performance. Therefore, exploring how to enhance perception ability through reasoning without the need for expensive fine-grained annotation information is worthwhile. To achieve this goal, we specially propose APPO, the Attention-guided Perception Policy Optimization algorithm that leverages token-level dense rewards to improve model's fine-grained perception. The core idea behind APPO is to optimize those tokens from different responses that primarily focus on the same crucial video frame (called intra-group perception tokens). Experimental results on diverse video benchmarks and models with different scales (3/7B) demonstrate APPO consistently outperforms GRPO and DAPO (0.5%~4%). We hope our work provides a promising approach to effectively enhance model's perception abilities through reasoning in a low-cost manner, serving diverse scenarios and demands.

Problem

• video reasoning
• perception enhancement
• fine-grained perception
• annotation-free learning
• model perception

Innovation

• perception enhancement
• attention-guided optimization
• video reasoning
• token-level reward
• annotation-free learning