Reinforcement Learning with Inverse Rewards for World Model Post-training

• URL: http://arxiv.org/abs/2509.23958v1
• Date: Sun, 28 Sep 2025 16:27:47 GMT
• Title: Reinforcement Learning with Inverse Rewards for World Model Post-training
• Authors: Yang Ye, Tianyu He, Shuo Yang, Jiang Bian,
• Abstract summary: We propose Reinforcement Learning with Inverse Rewards to improve action-following in video world models.<n>RLIR derives verifiable reward signals by recovering input actions from generated videos using an Inverse Dynamics Model.
• Score: 29.19830208692156
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: World models simulate dynamic environments, enabling agents to interact with diverse input modalities. Although recent advances have improved the visual quality and temporal consistency of video world models, their ability of accurately modeling human-specified actions remains under-explored. Reinforcement learning presents a promising approach for directly improving the suboptimal action-following capability of pre-trained models, assuming that an appropriate reward function can be defined. However, transferring reinforcement learning post-training methods to world model is impractical due to the prohibitive cost of large-scale preference annotations and the infeasibility of constructing rule-based video verifiers. To address this gap, we propose Reinforcement Learning with Inverse Rewards (RLIR), a post-training framework that derives verifiable reward signals by recovering input actions from generated videos using an Inverse Dynamics Model. By mapping high-dimensional video modality to a low-dimensional action space, RLIR provides an objective and verifiable reward for optimization via Group Relative Policy Optimization. Experiments across autoregressive and diffusion paradigms demonstrate 5-10% gains in action-following, up to 10% improvements in visual quality, and higher human preference scores, establishing RLIR as the first post-training method specifically designed to enhance action-following in video world models.

Related papers

• ReWorld: Multi-Dimensional Reward Modeling for Embodied World Models [27.729654985554372]
  ReWorld is a framework aimed to employ reinforcement learning to align the video-based embodied world models with physical realism, task completion capability, embodiment plausibility and visual quality.<n>We show that ReWorld significantly improves the physical fidelity, logical coherence, embodiment and visual quality of generated rollouts, outperforming previous methods.
  arXiv  Detail & Related papers  (2026-01-18T14:27:10Z)
• 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)
• Vid2World: Crafting Video Diffusion Models to Interactive World Models [38.270098691244314]
  Vid2World is a general approach for leveraging and transferring pre-trained video diffusion models into interactive world models.<n>It performs casualization of a pre-trained video diffusion model by crafting its architecture and training objective to enable autoregressive generation.<n>It introduces a causal action guidance mechanism to enhance action controllability in the resulting interactive world model.
  arXiv  Detail & Related papers  (2025-05-20T13:41:45Z)
• Vision-R1: Evolving Human-Free Alignment in Large Vision-Language Models via Vision-Guided Reinforcement Learning [26.14137626882127]
  Large Vision-Language Models (LVLMs) typically follow a two-stage training paradigm-pretraining and supervised fine-tuning.<n> preference optimization, derived from the language domain, has emerged as an effective post-training reinforcement strategy.<n>We propose Vision-R1, a novel vision-guided R1-like reinforcement learning algorithm for LVLMs that rewards models with definitive vision feedback.
  arXiv  Detail & Related papers  (2025-03-23T10:21:14Z)
• Pre-Trained Video Generative Models as World Simulators [59.546627730477454]
  We propose Dynamic World Simulation (DWS) to transform pre-trained video generative models into controllable world simulators.<n>To achieve precise alignment between conditioned actions and generated visual changes, we introduce a lightweight, universal action-conditioned module.<n> Experiments demonstrate that DWS can be versatilely applied to both diffusion and autoregressive transformer models.
  arXiv  Detail & Related papers  (2025-02-10T14:49:09Z)
• Improving Video Generation with Human Feedback [105.81833319891537]
  We develop a systematic pipeline that harnesses human feedback to mitigate video generation problems.<n>We introduce VideoReward, a multi-dimensional video reward model, and examine how annotations and various design choices impact its rewarding efficacy.
  arXiv  Detail & Related papers  (2025-01-23T18:55:41Z)
• iVideoGPT: Interactive VideoGPTs are Scalable World Models [70.02290687442624]
  World models empower model-based agents to interactively explore, reason, and plan within imagined environments for real-world decision-making. This work introduces Interactive VideoGPT, a scalable autoregressive transformer framework that integrates multimodal signals--visual observations, actions, and rewards--into a sequence of tokens. iVideoGPT features a novel compressive tokenization technique that efficiently discretizes high-dimensional visual observations.
  arXiv  Detail & Related papers  (2024-05-24T05:29:12Z)
• ReCoRe: Regularized Contrastive Representation Learning of World Model [21.29132219042405]
  We present a world model that learns invariant features using contrastive unsupervised learning and an intervention-invariant regularizer. Our method outperforms current state-of-the-art model-based and model-free RL methods and significantly improves on out-of-distribution point navigation tasks evaluated on the iGibson benchmark.
  arXiv  Detail & Related papers  (2023-12-14T15:53:07Z)
• Continual Visual Reinforcement Learning with A Life-Long World Model [55.05017177980985]
  We present a new continual learning approach for visual dynamics modeling.<n>We first introduce the life-long world model, which learns task-specific latent dynamics.<n>Then, we address the value estimation challenge for previous tasks with the exploratory-conservative behavior learning approach.
  arXiv  Detail & Related papers  (2023-03-12T05:08:03Z)
• Reinforcement Learning with Action-Free Pre-Training from Videos [95.25074614579646]
  We introduce a framework that learns representations useful for understanding the dynamics via generative pre-training on videos. Our framework significantly improves both final performances and sample-efficiency of vision-based reinforcement learning.
  arXiv  Detail & Related papers  (2022-03-25T19:44:09Z)
• Generative Adversarial Reward Learning for Generalized Behavior Tendency Inference [71.11416263370823]
  We propose a generative inverse reinforcement learning for user behavioral preference modelling. Our model can automatically learn the rewards from user's actions based on discriminative actor-critic network and Wasserstein GAN.
  arXiv  Detail & Related papers  (2021-05-03T13:14:25Z)

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.