OpenHA：Minecraft 中的一系列开源分层智能体模型
OpenHA: A Series of Open-Source Hierarchical Agentic Models in Minecraft
AI总结：
本文针对Minecraft中VLA分层智能体动作空间选择困境，提出统一高层规划与低层控制的CoA框架，训练出的全能智能体超越基线达到SOTA，并开源了包含超800个任务的OpenHA套件。
摘要：
动作空间的选择是开发强大且可端到端训练的智能体时一个关键但尚未解决的挑战。本文首先在开放式的 Minecraft 中，针对 Vision-Language-Action (VLA) 或分层智能体模型，对突出的抽象动作空间和分词器进行了大规模、系统性的比较。我们的分析表明，没有单一的动作空间是普遍最优的；相反，最有效的抽象高度依赖于任务，这为构建通用智能体带来了困境。为了解决这个问题，我们引入了 Chain of Action (CoA)，这是一个新颖的框架，它在单个单体的 VLA 模型中统一了高层规划和低层控制。CoA 将抽象动作不视为对单独策略的命令，而是作为一个中间推理步骤——类似于思维链——来指导最终可执行动作的生成。此外，我们证明了使用 CoA 范式在多种动作空间的混合数据上训练的 All-in-One 智能体能够学习到更鲁棒且更具泛化性的策略。这个统一的智能体实现了新的最先进水平，相较于强大的专业化基线，提高了整体任务成功率。为了促进可复现的研究，我们发布了 OpenHA (Open Hierarchical Agents) 套件，其中包含我们包含超过 800 个不同任务的全面基准、精选数据集、源代码以及所有预训练模型检查点，网址为 https://github.com/CraftJarvis/OpenHA
英文摘要：
The choice of action spaces is a critical yet unresolved challenge in developing capable, end-to-end trainable agents. This paper first presents a large-scale, systematic comparison of prominent abstracted action spaces and tokenizers for Vision-Language-Action (VLA) or hierarchical agent models in the open-ended Minecraft. Our analysis reveals that no single action space is universally optimal; instead, the most effective abstraction is highly task-dependent, creating a dilemma for building generalist agents. To resolve this, we introduce Chain of Action (CoA), a novel framework that unifies high-level planning and low-level control within a single, monolithic VLA model. CoA treats an abstracted action not as a command for a separate policy, but as an intermediate reasoning step--akin to a chain of thought--that guides the generation of the final, executable action. Furthermore, we demonstrate that an All-in-One agent trained on a diverse mixture of action spaces using the CoA paradigm learns a more robust and generalizable policy. This unified agent achieves a new state-of-the-art, improving the overall task success rate over strong, specialized baselines. To foster reproducible research, we release the OpenHA (Open Hierarchical Agents) suite, which includes our comprehensive benchmark of over 800 distinct tasks, curated datasets, source code, and all pretrained model checkpoints at https://github.com/CraftJarvis/OpenHA
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
