The Year in AI Papers
The year/Independent research

Paper 2511.11113

VIDEOP2R: Video Understanding from Perception to Reasoning

Yifan Jiang, Yueying Wang, Rui Zhao, Toufiq Parag, Zhimin Chen, Zhenyu Liao, Jayakrishnan Unnikrishnan

Published
Nov 2025
Research lab
Independent
Citations
8
GitHub
Not linked

01 In brief

Summary

The paper introduces VIDEOP2R, a process-aware reinforcement fine-tuning (RFT) framework for video understanding that explicitly separates perception and reasoning.

It addresses limitations of existing video RFT methods, which treat video reasoning as a single process and assign a single final reward, leading to poor credit assignment.

VIDEOP2R consists of two stages: in the SFT stage, a three-step pipeline generates VIDEOP2R-CoT-162K, a high-quality process-aware chain-of-thought dataset with separate perception and reasoning traces; in the RL stage, a novel process-aware group relative policy optimization (PA-GRPO) algorithm provides separate rewards for perception and reasoning, improving credit assignment.

Experiments on seven benchmarks show VIDEOP2R achieves state-of-the-art performance on six, with average accuracy gains of 1.9%–9.1% over the base model.

Ablations confirm the effectiveness of process-aware modeling and PA-GRPO, and show that perception outputs are information-sufficient for downstream reasoning.

The framework also mitigates think-answer mismatch and improves training efficiency by reducing advantage collapse.

02 From the paper

Abstract

Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. Our project page is available at https://videop2r.github.io/videop2r/.