Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

state-conditioned memory

State-conditioned memory is an artificial intelligence memory framework in which stored information and past experiences are dynamically accessed, selected, or compiled based on an agent's real-time execution state. Unlike static memory systems that inject fixed historical context at the start of a task, state-conditioned memory continuously aligns relevant knowledge with the agent's evolving environmental conditions and current operational needs. This dynamic alignment allows autonomous systems, such as embodied agents, to receive targeted guidance tailored to each specific decision-making step, thereby reducing context misalignment and improving efficiency across changing environments.

1 item

MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents

MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents

Xintao Ding, Xinrui Wang, Yifan Yang, Hao Wu, Shiqi Jiang, Qianxi Zhang, Liang Mi, Hanxin Zhu, Kunxi Li, Yunxin Liu, Zhibo Chen, Ting Cao

OrganizationsHuazhong University of Science and TechnologyMicrosoftNanjing UniversityTsinghua UniversityUniversity of Science and Technology of China

Why you should read this

Introduces a state-conditioned memory compilation framework for embodied agents that dynamically translates past experience into real-time text and latent guidance, boosting task success by up to 129% while cutting per-step latency by 60%.

Existing memory systems for embodied agents typically inject retrieved memory as static context at episode start, a paradigm we term Ahead-of-time Monolithic Memory Injection (AMMI). However, this static design quickly becomes misaligned with the agent's evolving state and may degrade lightweight executors below the no-memory baseline. To address this, we propose MemCompiler, which reframes memory utilization as State-Conditioned Memory Compilation. A learned Memory Compiler reads a structured Brief State capturing the agent's current execution state and dynamically selects and compiles only relevant memory into executable guidance. This guidance is delivered through a text channel and a latent Soft-Mem channel that preserves perceptual information not expressible in text. Across Alf World, EmbodiedBench, and ScienceWorld, MemCompiler consistently improves over no-memory across open-source backbones (up to +129%), matches or approaches frontier closed-source systems, and reduces per-step latency by 60%, demonstrating that state-aware memory compilation improves both effectiveness and efficiency.

Added

2026-09-29