MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents
About
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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Embodied AI Task Planning | EB-ALFRED | Average Score40 | 72 | |
| Embodied AI | EmbodiedBench EB-Habitat | Base Score93.25 | 53 | |
| Science Experiment Execution | ScienceWorld (test) | Success Rate48.44 | 24 | |
| Household Task Execution | ALFWorld (test) | Success Rate91.45 | 24 | |
| Embodied AI task execution | EmbodiedBench EB-ALFRED | Average Success Rate48.64 | 24 |