ReasonAct: Progressive Training for Fine-Grained Video Reasoning in Small Models

• URL: http://arxiv.org/abs/2508.01533v1
• Date: Sun, 03 Aug 2025 01:25:13 GMT
• Title: ReasonAct: Progressive Training for Fine-Grained Video Reasoning in Small Models
• Authors: Jiaxin Liu, Zhaolu Kang,
• Abstract summary: We introduce ReasonAct, a method that enhances video reasoning in smaller models.<n>We build a foundation with text-only reasoning, then fine-tuning on video, and finally refining with temporal-aware reinforcement learning.<n>Our experiments demonstrate improvements of 17.9, 15.8, and 12.3 points over baselines.
• Score: 2.8939451583731963
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While recent multimodal models have shown progress in vision-language tasks, small-scale variants still struggle with the fine-grained temporal reasoning required for video understanding. We introduce ReasonAct, a method that enhances video reasoning in smaller models through a three-stage training process: first building a foundation with text-only reasoning, then fine-tuning on video, and finally refining with temporal-aware reinforcement learning. We build upon Temporal Group Relative Policy Optimization (T-GRPO) by incorporating temporal consistency modeling into policy optimization. We also propose a biomechanically-motivated sub-action decomposition mechanism that provides graduated rewards for constituent action phases. Through experiments on HMDB51, UCF-101, and Kinetics-400, our 3B-parameter model achieves 67.2%, 94.1%, and 78.9% accuracy respectively, demonstrating improvements of 17.9, 15.8, and 12.3 points over baselines. Ablation studies validate that our progressive training methodology enables smaller models to achieve competitive video reasoning performance while maintaining computational efficiency.

Related papers

• SimpleGVR: A Simple Baseline for Latent-Cascaded Video Super-Resolution [55.14432034345353]
  We study key design principles for latter cascaded video super-resolution models, which are underexplored currently.<n>First, we propose two strategies to generate training pairs that better mimic the output characteristics of the base model, ensuring alignment between the VSR model and its upstream generator.<n>Second, we provide critical insights into VSR model behavior through systematic analysis of (1) timestep sampling strategies, (2) noise augmentation effects on low-resolution (LR) inputs.
  arXiv  Detail & Related papers  (2025-06-24T17:57:26Z)
• ReAgent-V: A Reward-Driven Multi-Agent Framework for Video Understanding [71.654781631463]
  ReAgent-V is a novel agentic video understanding framework.<n>It integrates efficient frame selection with real-time reward generation during inference.<n>Extensive experiments on 12 datasets demonstrate significant gains in generalization and reasoning.
  arXiv  Detail & Related papers  (2025-06-02T04:23:21Z)
• Improved Methods for Model Pruning and Knowledge Distillation [3.8993503758122663]
  MAMA Pruning is a performance optimization technique for large language models like R1 or o3-mini.<n>It effectively reduces model size and computational complexity while maintaining performance comparable to the original unpruned model even at extreme pruned levels.<n>Preliminary experimental results show that our method outperforms and be comparable to state-of-the-art methods across various pruning levels and different downstream computational linguistics tasks.
  arXiv  Detail & Related papers  (2025-05-20T07:53:40Z)
• VAPO: Efficient and Reliable Reinforcement Learning for Advanced Reasoning Tasks [49.0793012627959]
  We present VAPO, a novel framework tailored for reasoning models within the value-based paradigm.<n>VAPO attains a state-of-the-art score of $mathbf60.4$.<n>In direct comparison under identical experimental settings, VAPO outperforms the previously reported results of DeepSeek-R1-Zero-Qwen-32B and DAPO by more than 10 points.
  arXiv  Detail & Related papers  (2025-04-07T14:21:11Z)
• A Temporal Modeling Framework for Video Pre-Training on Video Instance Segmentation [32.65111815620319]
  We present a novel video pre-training approach to enhance VIS models.<n>Our innovation focuses on reducing disparities between the pre-training and fine-tuning stages.<n>Our approach achieves a notable 4.0% increase in average precision on the challenging OVIS dataset.
  arXiv  Detail & Related papers  (2025-03-22T07:01:25Z)
• An Empirical Study on Eliciting and Improving R1-like Reasoning Models [90.52239241349504]
  scaling RL training has become a central technique for implementing such reasoning models.<n>We demonstrate that our RL training approach consistently improves the Qwen2.5-32B base models.<n>We also explore the use of tool manipulation, finding that it significantly boosts the reasoning performance of large reasoning models.
  arXiv  Detail & Related papers  (2025-03-06T15:34:27Z)
• LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging [10.33844295243509]
  We propose a unified framework for model merging based on low-rank estimation of task vectors without the need for access to the base model, named textscLoRE-Merging.<n>Our approach is motivated by the observation that task vectors from fine-tuned models frequently exhibit a limited number of dominant singular values, making low-rank estimations less prone to interference.
  arXiv  Detail & Related papers  (2025-02-15T10:18:46Z)
• STEP: Enhancing Video-LLMs' Compositional Reasoning by Spatio-Temporal Graph-guided Self-Training [87.58996020705258]
  Video Large Language Models (Video-LLMs) have recently shown strong derivation in basic video understanding tasks.<n>Video-LLMs struggle with compositional reasoning that requires multi-step explicit-temporal inference across object relations, interactions and events.<n>We propose STEP, a novel graph-guided self-training method that enables VideoLLMs to generate reasoning-rich finetuning data from any raw videos to improve itself.
  arXiv  Detail & Related papers  (2024-11-29T11:54:55Z)
• The Surprising Effectiveness of Test-Time Training for Few-Shot Learning [59.309477460893916]
  Language models (LMs) have shown impressive performance on tasks within their training distribution, but often struggle with structurally novel tasks.<n>We investigate the effectiveness of test-time training (TTT) as a mechanism for improving LMs' reasoning and few-shot learning capabilities.<n>Our findings highlight the limitations of in-context learning for novel tasks and demonstrate the potential of test-time training to enhance language model adaptability.
  arXiv  Detail & Related papers  (2024-11-11T18:59:45Z)
• MoQuad: Motion-focused Quadruple Construction for Video Contrastive Learning [10.41936704731324]
  This paper presents a simple yet effective sample construction strategy to boost the learning of motion features in video contrastive learning. The proposed method, dubbed Motion-focused Quadruple Construction (MoQuad), augments the instance discrimination by meticulously disturbing the appearance and motion of both the positive and negative samples. By simply applying MoQuad to SimCLR, extensive experiments show that we achieve superior performance on downstream tasks compared to the state of the arts.
  arXiv  Detail & Related papers  (2022-12-21T09:26:40Z)
• When to Update Your Model: Constrained Model-based Reinforcement Learning [50.74369835934703]
  We propose a novel and general theoretical scheme for a non-decreasing performance guarantee of model-based RL (MBRL) Our follow-up derived bounds reveal the relationship between model shifts and performance improvement. A further example demonstrates that learning models from a dynamically-varying number of explorations benefit the eventual returns.
  arXiv  Detail & Related papers  (2022-10-15T17:57:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.