面向部分可观测下时序知识图谱记忆的神经符号元策略

View PDF HTML (experimental)

Abstract:Partially observable reinforcement learning requires deciding what to retain, retrieve, and forget over time. We introduce a neuro-symbolic meta-policy that learns which symbolic memory heuristic to apply at each decision point while keeping execution symbolic. Our setting uses temporal knowledge-graph memory in RoomKG, where hidden state and observations are represented as Resource Description Framework (RDF) graphs and memory is augmented with temporal RDF triple annotations. The model combines knowledge-graph encoding of memory contents with value heads for question answering, exploration, and forgetting, yielding a controller that is both adaptive and inspectable. This gives the work a direct Semantic Web grounding through RDF-based representation, annotation-compatible graph semantics, and graph-based symbolic operations over explicit memory state. On train/test room splits at long-term memory capacity of 512, the qualifier-aware StarE-GNN configuration achieves the best held-out performance among the compared symbolic, neural, and neuro-symbolic systems while preserving step-level traceability of memory-management decisions.

Comments:	Accepted at ISWC 2026 (Research Track), LNCS 17126, pp. 191-209. v5: accepted-manuscript notice on the first page, wider margins
Subjects:	Artificial Intelligence (cs.AI)
Cite as:	arXiv:2607.18368 [cs.AI]
	(or arXiv:2607.18368v5 [cs.AI] for this version)
	https://doi.org/10.48550/arXiv.2607.18368 arXiv-issued DOI via DataCite
Journal reference:	ISWC 2026, LNCS 17126, pp. 191-209
Related DOI:	https://doi.org/10.1007/978-3-032-42015-2_10 DOI(s) linking to related resources

Submission history

From: Taewoon Kim [view email]
[v1] Mon, 20 Jul 2026 17:55:00 UTC (210 KB)
[v2] Fri, 24 Jul 2026 20:23:23 UTC (194 KB)
[v3] Tue, 28 Jul 2026 13:21:18 UTC (200 KB)
[v4] Sat, 3 Oct 2026 07:26:30 UTC (194 KB)
[v5] Thu, 29 Oct 2026 16:35:11 UTC (193 KB)