Developing a universal and versatile embodied intelligence system presents two primary challenges: the critical embodied data bottleneck, where real-world data is scarce and expensive, and the algorithmic inefficiency of existing methods, which are resource-prohibitive. To address these limitations, we introduce Deliberate Practice Policy Optimization (DPPO), a metacognitive ``Metaloop'' training framework that dynamically alternates between supervised fine-tuning (competence expansion) and reinforcement learning (skill refinement). This enables automatic weakness identification and targeted resource allocation, specifically designed to maximize learning efficiency from sparse, finite data. Theoretically, DPPO can be formalised as a unified preference-learning framework. Empirically, training a vision-language embodied model with DPPO, referred to as Pelican-VL 1.0, yields a 20.3% performance improvement over the base model and surpasses open-source models at the 100B-parameter scale by 10.6%. We are open-sourcing both the models and code, providing the first systematic framework that alleviates the data and resource bottleneck and enables the community to build versatile embodied agents efficiently.
@article{arxiv.2511.16602,
title = {Bridging VLMs and Embodied Intelligence with Deliberate Practice Policy Optimization},
author = {Yi Zhang and Che Liu and Xiancong Ren and Hanchu Ni and Yingji Zhang and Shuai Zhang and Zeyuan Ding and Jiayu Hu and Haozhe Shan and Junbo Qi and Yan Bai and Dengjie Li and Jiachen Luo and Yidong Wang and Yong Dai and Zenglin Xu and Bin Shen and Qifan Wang and Jian Tang and Xiaozhu Ju},
journal= {arXiv preprint arXiv:2511.16602},
year = {2025}
}