World Model Self-Distillation: Training World Models to Solve General Tasks

• URL: http://arxiv.org/abs/2606.12072v1
• Date: Wed, 10 Jun 2026 13:40:19 GMT
• Title: World Model Self-Distillation: Training World Models to Solve General Tasks
• Abstract summary: We propose a scalable framework that elicits task-solving ability in pretrained video models by combining self-distillation and reinforcement learning.<n>We transfer execution knowledge from caption-guided generation to instruction-conditioned task solving without curated task-video supervision.<n> Experiments show that the Executor surpasses the Demonstrator under our VLM-based evaluation protocol and transfers competitively to robotic tasks.
• Score: 23.602187351200183
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Pretrained video generators are promising visual world models that exhibit emergent task-solving abilities; however, their reliance on detailed textual descriptions limits their direct use for planning and decision-making. Existing approaches either outsource this reasoning to language or vision-language models, or rely on supervised fine-tuning with paired task-execution videos, which are costly to collect and difficult to scale. We propose a scalable framework that elicits task-solving ability in such models by combining self-distillation with reinforcement learning. Given an unlabeled scene image, a vision-language model generates a candidate task and a detailed step-by-step solution. The solution conditions a pretrained video diffusion model, the Demonstrator; we distill its behavior into an Executor conditioned only on the image and a short task prompt. This transfers execution knowledge from caption-guided generation to instruction-conditioned task solving without curated task-video supervision. We further improve the Executor with reinforcement learning from VLM feedback, exploiting the asymmetry between judging whether a sampled video satisfies a task and generating the solution. Experiments on our proposed WorldTasks-Benchmark and the DreamGen robotics benchmark show that the Executor surpasses the Demonstrator under our VLM-based evaluation protocol and transfers competitively to robotic tasks.

Related papers

• SparTa: Sparse Graphical Task Models from a Handful of Demonstrations [18.00619071013106]
  We focus on inferring what the robot should achieve in the task, rather than how to do so.<n>We propose a demonstration segmentation and pooling approach that extracts a series of manipulation graphs and estimates distributions over object states.<n>We deploy the fitted models both in simulation and on a real robot, demonstrating that the resulting task representations support reliable execution across environments.
  arXiv  Detail & Related papers  (2026-02-18T21:54:35Z)
• Exploring Conditions for Diffusion models in Robotic Control [70.27711404291573]
  We explore leveraging pre-trained text-to-image diffusion models to obtain task-adaptive visual representations for robotic control.<n>We find that naively applying textual conditions yields minimal or even negative gains in control tasks.<n>We propose ORCA, which introduces learnable task prompts that adapt to the control environment and visual prompts that capture fine-grained, frame-specific details.
  arXiv  Detail & Related papers  (2025-10-17T10:24:14Z)
• UniVid: Unifying Vision Tasks with Pre-trained Video Generation Models [12.21686773633269]
  Large language models, trained on extensive corpora, successfully unify diverse linguistic tasks within a single generative framework.<n>Such modeling requires task-specific pre-training across modalities and sources, which is costly and limits scalability to unseen tasks.<n>We propose UniVid, a framework that fine-tunes a video diffusion transformer to handle various vision tasks without task-specific modifications.
  arXiv  Detail & Related papers  (2025-09-26T01:43:40Z)
• Latent Action Pretraining Through World Modeling [1.988007188564225]
  We propose LAWM, a model-agnostic framework to pretrain imitation learning models in a self-supervised way.<n>Our framework is designed to be effective for transferring across tasks, environments, and embodiments.
  arXiv  Detail & Related papers  (2025-09-22T21:19:10Z)
• Subtask-Aware Visual Reward Learning from Segmented Demonstrations [97.80917991633248]
  This paper introduces REDS: REward learning from Demonstration with Demonstrations, a novel reward learning framework.<n>We train a dense reward function conditioned on video segments and their corresponding subtasks to ensure alignment with ground-truth reward signals.<n>Our experiments show that REDS significantly outperforms baseline methods on complex robotic manipulation tasks in Meta-World.
  arXiv  Detail & Related papers  (2025-02-28T01:25:37Z)
• TWIST & SCOUT: Grounding Multimodal LLM-Experts by Forget-Free Tuning [54.033346088090674]
  We introduce TWIST & SCOUT, a framework that equips pre-trained MLLMs with visual grounding ability.<n>To fine-tune the model effectively, we generate a high-quality synthetic dataset we call SCOUT.<n>This dataset provides rich supervision signals, describing a step-by-step multimodal reasoning process.
  arXiv  Detail & Related papers  (2024-10-14T13:35:47Z)
• Harnessing Large Language Models for Training-free Video Anomaly Detection [34.76811491190446]
  Video anomaly detection (VAD) aims to temporally locate abnormal events in a video. Training-based methods are prone to be domain-specific, thus being costly for practical deployment. We propose LAnguage-based VAD (LAVAD), a method tackling VAD in a novel, training-free paradigm.
  arXiv  Detail & Related papers  (2024-04-01T09:34:55Z)
• Look, Remember and Reason: Grounded reasoning in videos with language models [5.3445140425713245]
  Multi-temporal language models (LM) have recently shown promising performance in high-level reasoning tasks on videos. We propose training an LM end-to-end on low-level surrogate tasks, including object detection, re-identification, tracking, to endow the model with the required low-level visual capabilities. We demonstrate the effectiveness of our framework on diverse visual reasoning tasks from the ACRE, CATER, Something-Else and STAR datasets.
  arXiv  Detail & Related papers  (2023-06-30T16:31:14Z)
• Dense Video Object Captioning from Disjoint Supervision [77.47084982558101]
  We propose a new task and model for dense video object captioning. This task unifies spatial and temporal localization in video. We show how our model improves upon a number of strong baselines for this new task.
  arXiv  Detail & Related papers  (2023-06-20T17:57:23Z)
• Learning and Verification of Task Structure in Instructional Videos [85.511888642497]
  We introduce a new pre-trained video model, VideoTaskformer, focused on representing the semantics and structure of instructional videos. Compared to prior work which learns step representations locally, our approach involves learning them globally. We introduce two new benchmarks for detecting mistakes in instructional videos, to verify if there is an anomalous step and if steps are executed in the right order.
  arXiv  Detail & Related papers  (2023-03-23T17:59:54Z)
• Goal-Conditioned End-to-End Visuomotor Control for Versatile Skill Primitives [89.34229413345541]
  We propose a conditioning scheme which avoids pitfalls by learning the controller and its conditioning in an end-to-end manner. Our model predicts complex action sequences based directly on a dynamic image representation of the robot motion. We report significant improvements in task success over representative MPC and IL baselines.
  arXiv  Detail & Related papers  (2020-03-19T15:04:37Z)

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.