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

DevFeed: [SPARK: Self-Correcting Embodied Memory via Physical Re-Observation](<https://devfeed.tech/articles/spark-self-correcting-embodied-memory-via-physical-re-observation-114368.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.00344>)

Author: Dmytro Kurdydyk, Aadi Palnitkar, Sebastian Scherer, Seungchan Kim

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [база знаний](<https://devfeed.tech/topics/tag-c51cc4bf76e5.md>), [Machine Perception](<https://devfeed.tech/topics/machine-perception.md>)

Tags: [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [errors](<https://devfeed.tech/tags/errors.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [knowledge-base](<https://devfeed.tech/tags/knowledge-base.md>), [language](<https://devfeed.tech/tags/language.md>), [map](<https://devfeed.tech/tags/map.md>), [memory](<https://devfeed.tech/tags/memory.md>), [merge](<https://devfeed.tech/tags/merge.md>), [outdoor](<https://devfeed.tech/tags/outdoor.md>), [planner](<https://devfeed.tech/tags/planner.md>), [rag](<https://devfeed.tech/tags/rag.md>), [retrieval-augmented-generation](<https://devfeed.tech/tags/retrieval-augmented-generation.md>), [retrieval-augmented-generation-rag](<https://devfeed.tech/tags/retrieval-augmented-generation-rag.md>), [risk](<https://devfeed.tech/tags/risk.md>), [scale](<https://devfeed.tech/tags/scale.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [spark](<https://devfeed.tech/tags/spark.md>), [spatial](<https://devfeed.tech/tags/spatial.md>)

## AI overview

SPARK addresses errors in embodied retrieval-augmented memory by using a semantic critic to flag questionable knowledge-base regions, a planner to direct physical revisits, and a selective merge to add, remove, or replace information. In tests on four outdoor AirSim maps with two degraded viewpoint conditions and 158 fixed embodied questions, one corrective episode improved question-answering performance on every map while storing 13-41% fewer objects than append-only updating; gains reached 2.7x on one map-condition pair. Correction through Episode 4 raised all maps above their undegraded reference performance. Ablations report larger mean gains from targeted than untargeted revisits under the same budget, and comparable gains from selective merging and append-only updating with a smaller knowledge base. These are reported simulation results.

## Source excerpt

arXiv:2611.00344v1 Announce Type: cross Abstract: 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. 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. 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. 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 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. 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.