CRAFT: Adapting VLA Models to Contact-rich Manipulation via Force-aware Curriculum Fine-tuning

• URL: http://arxiv.org/abs/2602.12532v1
• Date: Fri, 13 Feb 2026 02:28:21 GMT
• Title: CRAFT: Adapting VLA Models to Contact-rich Manipulation via Force-aware Curriculum Fine-tuning
• Abstract summary: Vision-Language-Action models can execute general instructions, but struggle with contact-rich manipulation tasks.<n>We introduce CRAFT, a force-aware curriculum fine-tuning framework to regulate vision and language embeddings during early training.<n>We show that CRAFT consistently improves task success, generalizes to unseen objects and novel task variations, and adapts effectively across diverse VLA architectures.
• Score: 46.57805525532354
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Vision-Language-Action (VLA) models have shown a strong capability in enabling robots to execute general instructions, yet they struggle with contact-rich manipulation tasks, where success requires precise alignment, stable contact maintenance, and effective handling of deformable objects. A fundamental challenge arises from the imbalance between high-entropy vision and language inputs and low-entropy but critical force signals, which often leads to over-reliance on perception and unstable control. To address this, we introduce CRAFT, a force-aware curriculum fine-tuning framework that integrates a variational information bottleneck module to regulate vision and language embeddings during early training. This curriculum strategy encourages the model to prioritize force signals initially, before progressively restoring access to the full multimodal information. To enable force-aware learning, we further design a homologous leader-follower teleoperation system that collects synchronized vision, language, and force data across diverse contact-rich tasks. Real-world experiments demonstrate that CRAFT consistently improves task success, generalizes to unseen objects and novel task variations, and adapts effectively across diverse VLA architectures, enabling robust and generalizable contact-rich manipulation.

Related papers

• DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination [39.716584408571265]
  Dexterous manipulation involves contact-rich and fine-grained interactions with the physical world.<n>We present DeCAL, a vision-language-action model that unifies understanding, imagination and action generation for contact-rich manipulation.
  arXiv  Detail & Related papers  (2026-09-08T17:47:43Z)
• CR-VLA-Force: Learning Control-aware Compliance VLA Model for Robust Contact-rich Robotic Manipulation [22.8919569384701]
  We propose a Control-aware Compliance VLA framework for reactive control.<n>The CC-VLA model employs a multimodal mixture-of-experts (MoE) to encode force signal sequences and vision-language fused feature.<n>We show that CC-VLA significantly improves success rates in challenging force-perception tasks and enhances force-control precision.
  arXiv  Detail & Related papers  (2026-09-05T02:54:51Z)
• Simple-to-Complex Structured Demonstrations for Vision-Language-Action Learning [9.340058165047717]
  Vision-Language-Action (VLA) models have demonstrated strong capabilities in robotic manipulation.<n>We identify demonstration organization as a fundamental yet overlooked aspect of imitation learning.<n>We propose a simple-to-complex structured demonstration collection strategy for VLA learning using a dual-arm robotic platform.
  arXiv  Detail & Related papers  (2026-07-06T01:38:43Z)
• TACO: TActile World Model as a Self-COrrector forScalable VLA Post-Training [79.94818813802682]
  We introduce TACO, a tactile-aware world-model-driven framework for scalable VLA post-training in contact-rich manipulation.<n>To incorporate tactile corrective supervision into VLA post-training, TACO combines knowledge-insulated tactile adaptation with advantage-conditioned training.
  arXiv  Detail & Related papers  (2026-07-03T00:23:30Z)
• Decoupling the Declarative from the Procedural in Vision-Language-Action Models [65.22639791239932]
  Policy trained to clone a behavior from object-specific demonstrations must generalize beyond that object.<n>We propose w$2$VLA, a new Vision-Language-Action model with restructured information flow.<n>Unlike popular, state-of-the-art VLAs, we show that our modular approach successfully decouples knowledge representations.
  arXiv  Detail & Related papers  (2026-06-19T14:43:58Z)
• Tactile Modality Fusion for Vision-Language-Action Models [22.788833830429766]
  We propose TacFiLM, a lightweight modality-fusion approach that integrates visual-tactile signals into vision-language-action (VLA) models.<n>Results show consistent improvements in success rate, direct insertion performance, completion time, and force stability across both in-distribution and out-of-distribution tasks.
  arXiv  Detail & Related papers  (2026-03-15T20:57:51Z)
• SELF-VLA: A Skill Enhanced Agentic Vision-Language-Action Framework for Contact-Rich Disassembly [5.973392871392723]
  Vision-language-action (VLA) models have shown impressive performance on standard robotic manipulation tasks.<n>We propose SELF-VLA, an agentic VLA framework that integrates explicit disassembly skills.
  arXiv  Detail & Related papers  (2026-03-10T22:30:28Z)
• On-the-Fly VLA Adaptation via Test-Time Reinforcement Learning [61.38273866396522]
  Vision-Language-Action models have emerged as a powerful paradigm for general-purpose robot learning.<n>Current methods remain unsuitable for challenging simulated- or physical-world deployments.<n>We introduce a Test-Time Reinforcement Learning for VLAs framework that enables on-the-fly policy adaptation during inference.
  arXiv  Detail & Related papers  (2026-01-11T01:51:30Z)
• ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation [62.58034332427291]
  ForceVLA is a novel end-to-end manipulation framework.<n>It treats external force sensing as a first-class modality within VLA systems.
  arXiv  Detail & Related papers  (2025-05-28T09:24:25Z)
• Robotic Control via Embodied Chain-of-Thought Reasoning [86.6680905262442]
  Key limitation of learned robot control policies is their inability to generalize outside their training data.<n>Recent works on vision-language-action models (VLAs) have shown that the use of large, internet pre-trained vision-language models can substantially improve their robustness and generalization ability.<n>We introduce Embodied Chain-of-Thought Reasoning (ECoT) for VLAs, in which we train VLAs to perform multiple steps of reasoning about plans, sub-tasks, motions, and visually grounded features before predicting the robot action.
  arXiv  Detail & Related papers  (2024-07-11T17:31:01Z)

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.