Title: EvolvingGrasp: Evolutionary Grasp Generation via Efficient Preference Alignment Show Chinese title

Title: 进化抓取：通过高效偏好对齐的进化抓取生成

Abstract: Dexterous robotic hands often struggle to generalize effectively in complex environments due to the limitations of models trained on low-diversity data. However, the real world presents an inherently unbounded range of scenarios, making it impractical to account for every possible variation. A natural solution is to enable robots learning from experience in complex environments, an approach akin to evolution, where systems improve through continuous feedback, learning from both failures and successes, and iterating toward optimal performance. Motivated by this, we propose EvolvingGrasp, an evolutionary grasp generation method that continuously enhances grasping performance through efficient preference alignment. Specifically, we introduce Handpose wise Preference Optimization (HPO), which allows the model to continuously align with preferences from both positive and negative feedback while progressively refining its grasping strategies. To further enhance efficiency and reliability during online adjustments, we incorporate a Physics-aware Consistency Model within HPO, which accelerates inference, reduces the number of timesteps needed for preference finetuning, and ensures physical plausibility throughout the process. Extensive experiments across four benchmark datasets demonstrate state of the art performance of our method in grasp success rate and sampling efficiency. Our results validate that EvolvingGrasp enables evolutionary grasp generation, ensuring robust, physically feasible, and preference-aligned grasping in both simulation and real scenarios. Show Chinese abstract

Abstract: 灵活的机械手在复杂环境中往往难以有效泛化，这是由于在低多样性数据上训练的模型存在局限性。然而，现实世界呈现出本质上无限多样的场景，因此考虑所有可能的变化是不现实的。一个自然的解决方案是让机器人在复杂环境中通过经验进行学习，这种方法类似于进化，其中系统通过持续的反馈不断改进，从失败和成功中学习，并逐步向最佳性能迭代。受此启发，我们提出了EvolvingGrasp，这是一种通过高效偏好对齐持续提升抓取性能的进化抓取生成方法。具体来说，我们引入了手部姿态偏好优化（HPO），使模型能够在不断精进其抓取策略的同时，持续与正负反馈中的偏好对齐。为了进一步提高在线调整过程中的效率和可靠性，我们在HPO中引入了一个物理感知的一致性模型，该模型加快了推理速度，减少了偏好微调所需的步数，并在整个过程中确保物理合理性。在四个基准数据集上的大量实验表明，我们的方法在抓取成功率和采样效率方面达到了最先进水平。我们的结果验证了EvolvingGrasp能够实现进化抓取生成，在仿真和真实场景中都能确保稳健、物理可行且偏好对齐的抓取。