arXiv 2610.31162 · cs.CV · 2026-10-22
HiSTEVE 1.0: Harnessing Human-Interface VLA Agents for Open-World Games
Zecong Tang, Zian Guan, Yuncheng Guo, Zhanqiu Zhang, Yiwen Guo, Bo Dai
Tencent
一句话要点
HiSTEVE 1.0 提出一种层级式视觉语言代理,结合结构化记忆与规划模块,在Minecraft中实现了长时域自主决策和原生界面控制,达到顶尖的原子任务表现并完成从零开始的复杂建造任务。
HiSTEVE 1.0 achieves state-of-the-art atomic task performance and long-horizon gameplay in Minecraft via a hierarchical VLA agent with structured planning and native interface control.
要点
- 现有虚拟代理在开放世界游戏中难以兼顾长期自主决策与原生界面操作的可靠性,存在能力割裂和交互干扰问题。
- 提出层级视觉语言代理(HiSTEVE),配备结构化记忆与视觉语言规划器,通过策略耦合的任务完成检测器判断子任务结束,并使用专用RMSNorm分支和鼠标动作嵌入分离GUI与非GUI交互。
- 在MineStudio Bench上原子任务准确率达85.0%,OpenHA平均成功率37.4%,成功完成从零开始制作铁镐等复杂长时任务,表现媲美人类玩家。
- 实现了视觉语言代理在开放世界中的高效协同控制,但未明确说明其泛化能力或对动态环境的鲁棒性。
English overview
HiSTEVE 1.0 introduces a hierarchical visual-language agent for Minecraft that integrates long-horizon planning with native GUI control. It uses structured memory, a VLM planner, and policy-coupled completion detection to manage subtasks. Separate regime-specific branches and embeddings mitigate interaction interference. The agent achieves top-tier performance on benchmark tasks and successfully completes complex, human-like long-horizon gameplay.
只根据 arXiv 官方摘要,由开源模型整理,经数字/专名自动核对(必须在摘要中出现),以论文原文为准。
阅读 arXiv 原文 PDF Hugging Face 讨论
主题与学科
同主题:机器人与具身
- Has GPT-6-Astra Achieved Embodied Intelligence? ▲421
- In-Context Learning for Robots: Methods and Applications ▲388
- GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives ▲310
- HexaShell: A Thin Harness for Exploring the Limits of VLM Agents as Robot Coding Agents ▲294
- Reflex: Visually Grounded Reactive Humanoid Control ▲154
同主题:智能体
- Raven: The Harness of Harnesses for Composable Agentic Intelligence 2026-09-27 · ▲748
- False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents 2026-09-30 · ▲642
- Omni-IO Skills: Harnessing Your Agent Omni-Native 2026-09-25 · ▲485
- BizAgentBench: Unveiling Hidden Challenges in Deploying Long-Horizon Agents for Real-World Business 2026-10-14 · ▲388
- Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generation 2026-10-14 · ▲364
同机构:Tencent
- WeVisOne-VL: Unified Visual Perception from Objects to Pixels 2026-10-22
- Making MLLM Latents Matter: Factor-Aware Predictive Learning for Cinematic Video Generation 2026-10-22
- Looking Ahead: Modeling Future Evolution in Autoregressive Video Generation 2026-10-21
- Training the Whole Agent, Not Just the Model 2026-10-20
- Tree-Aware Posterior Distillation for Long-Horizon Agentic Reinforcement Learning 2026-10-16
数据说明
元数据来自 arXiv API(已核实);点赞来自 Hugging Face 每日论文;引用来自 Semantic Scholar。最近刷新 2026-10-22。
要点由本站根据摘要自写,不转载摘要原文,请阅读原文。