ChunkAct：动作分块让机器人智能体 LLM 调用减少 2.9 倍
ChunkAct: Action Chunking for Agentic Robotics
让机器人智能体一次 LLM 调用跑完一串动作，LIBERO-Pro 上成功率还高 5 个点，省算力。
这篇论文提出 ChunkAct，把动作分块（action chunking）引入 LLM 智能体机器人：一次 LLM 调用启动一串链接的原语，自主执行，只在失败时把控制权交还规划器。作者发现逐原语调用 LLM 大多冗余——子目标确定后原语序列已由记忆规定，ChunkAct 在离线阶段把记忆编译成块、测试时免记忆执行，让 LLM 计算量随决策数而非原语数增长。在 LIBERO-Pro 基准上，ChunkAct 把任务成功率比现有智能体机器人 SOTA 提高 5.0 个百分点，同时 LLM 调用少 2.9 倍、输入 token 少 4.4 倍、输出 token 少 2.3 倍。
ChunkAct: Action Chunking for Agentic Robotics
Agentic robotics places a large language model (LLM) agent in the robot control loop, where it composes perception, motion, and vision-language-action (VLA) tools under the guidance of a memory and thereby generalizes to new tasks without retraining. This generality comes at a steep price in efficiency. Every primitive requires a round of LLM reasoning over the full interaction history, so the cost of an episode grows with the number of primitives. We find this reasoning largely redundant: once a sub-goal is chosen, its primitives follow a routine already prescribed by memory. We present ChunkAct, which introduces action chunking to agentic robotics. Online, a single call launches a chunk of chained primitives, which executes autonomously and returns control only upon failure. Offline, ChunkAct compiles the memory into these chunks, so execution is memory-free at test time and the planner is reserved for decisions that require reasoning. LLM computation therefore scales with the number of decisions rather than the number of primitives. On LIBERO-Pro, ChunkAct improves task success by 5.0 points over the state of the art in agentic robotics while requiring 2.9x fewer LLM calls, 4.4x fewer input tokens, and 2.3x fewer output tokens.
