Skip to content
AI News HubLIVE
Original source2 min read

CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents

Summary

CoreMem introduces a resource-efficient edge-cloud memory architecture unified by information geometry. It replaces cosine similarity with a locally adaptive Fisher-Rao metric to mitigate hubness in high-dimensional retrieval, and uses Fisher-guided discrete token distillation for principled context compression. On LOCOMO and LongMemEval-S, it improves open-domain reasoning by 4.51 pp and temporal reasoning by 4.17 pp while operating within an 8 GB VRAM budget.

SourcearXiv Computational LinguisticsAuthor: Jiaqi Chen, Yongqin Zeng, Shaoshen Chen, Yijian Zhang, Hai-Tao Zheng, Chunxia Ma, XiuTeng Zhou
CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[2606.18406] CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents

[Submitted on 16 Jun 2026]

Title:CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents

View a PDF of the paper titled CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents, by Jiaqi Chen and 6 other authors

View PDF HTML (experimental)

Abstract:Personalized dialogue agents require continuous long-term memory to maintain coherent interactions across multiple sessions. However, deploying these capabilities on consumer-grade hardware (e.g., 8 GB VRAM edge devices) introduces severe memory and compute bottlenecks. Existing systems typically rely on isotropic cosine similarity for retrieval and heuristic rules for context compression. These approaches lack a unified theoretical foundation, frequently suffering from the hubness problem in high-dimensional retrieval and syntactic fragmentation during compression. To overcome these limitations, we propose CoreMem, a resource-efficient edge-cloud memory architecture fundamentally unified by information geometry. First, Riemannian retrieval replaces cosine matching with a locally adaptive Fisher-Rao metric, effectively penalizing hub memories via Mahalanobis distance with O(Ndr) Woodbury acceleration for real-time search. Second, Fisher-guided discrete token distillation (FDTD) introduces a hierarchical sentence-to-token compression mechanism. It derives sensitivity scores from Fisher information traces, providing a principled compression-KL tradeoff augmented with explicit structural syntax protection. Evaluated on the LOCOMO and LongMemEval-S benchmarks, CoreMem achieves strong accuracy improvements, yielding substantial gains in Open-domain (+4.51 pp) and Temporal (+4.17 pp) reasoning. Extensive profiling confirms that CoreMem operates seamlessly within a strict 8 GB VRAM budget, successfully bridging the gap between resource-constrained edge devices and the demand for theoretically grounded, lifelong memory agents.

Comments: 15 pages, 5 figures

Subjects:

Computation and Language (cs.CL)

ACM classes: I.2.7; I.2.6

Cite as: arXiv:2606.18406 [cs.CL]

(or arXiv:2606.18406v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2606.18406

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Jiaqi Chen [view email] [v1] Tue, 16 Jun 2026 18:56:23 UTC (7,328 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled CoreMem: Riemannian Retrieval and Fisher-Guided Distillation for Long-Term Memory in Dialogue Agents, by Jiaqi Chen and 6 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CL

new | recent | 2026-06

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Riemannian retrieval uses a locally adaptive Fisher-Rao metric and Mahalanobis distance to penalize hub memories, with Woodbury acceleration for real-time search.
  • Fisher-guided discrete token distillation (FDTD) provides hierarchical sentence-to-token compression with sensitivity scores from Fisher information traces and explicit syntax protection.
  • CoreMem achieves +4.51 pp on open-domain reasoning and +4.17 pp on temporal reasoning on LOCOMO and LongMemEval-S benchmarks.
  • The full system runs within a strict 8 GB VRAM budget, targeting consumer-grade edge devices.

Highlights and analysis are generated automatically and may contain errors. Check the original source.