首页 > AI前沿 > SPARK: Self-Correcting Embodied Memory via Physical Re-Observation

SPARK: Self-Correcting Embodied Memory via Physical Re-Observation

arXiv自然语言 2026-10-31 04:24 16 阅读 查看原文

Retrieval-augmented generation (RAG) systems have seen growing adoption in embodied settings as language-queryable memories of the environment. However, existing approaches often assume a reliable single-episode knowledge base (KB) or accumulate observations across episodes without correcting prior errors. Such errors can persist through incorrect semantic labels, inaccurate object positions, and inconsistent spatial relationships.

SPARK: A Corrective Multi-Episode Framework

To address this, we present SPARK, a corrective multi-episode framework for embodied RAG. After the first episode, a semantic critic flags regions of the KB, a targeted planner generates revisits, and a selective merge adds, removes, or replaces KB information based on the new observations.

Evaluation Setup

We evaluate SPARK on four outdoor AirSim maps under two degraded viewpoint conditions, using perception-built KBs, a visibility-constrained oracle used only for evaluation, and a fixed set of 158 embodied question answering (EQA) questions scored before and after correction.

Results

A single corrective episode improves downstream EQA on every map under both conditions while storing 13–41% fewer objects than append-only updating, with gains up to 2.7x on one map-condition pair. Extending correction through Episode 4 brings every map above the EQA performance of its undegraded reference.

Ablations

Ablations show that the critic's three signals each contribute to performance, that targeted revisits achieve a larger mean gain than untargeted ones under the same budget, and that selective merging achieves gains comparable to append-only updating with a smaller KB.

Conclusion

Overall, SPARK demonstrates the value of correcting embodied memory through targeted re-observation, particularly as embodied systems scale and the risk of accumulated errors increases.