![[2602.17913] From Lossy to Verified: A Provenance-Aware Tiered Memory for Agents](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.17913] From Lossy to Verified: A Provenance-Aware Tiered Memory for Agents
Summary
The paper presents TierMem, a novel memory framework for agents that balances the need for accurate evidence with efficiency, reducing latency and token consumption while maintaining high accuracy.
Why It Matters
This research addresses the critical challenge of managing memory in AI agents, particularly in balancing the trade-off between using compressed summaries and retaining detailed logs. By improving evidence retrieval, it enhances the reliability of AI responses, which is vital for applications in various fields, including healthcare and customer service.
Key Takeaways
- TierMem utilizes a two-tier memory system to optimize evidence retrieval.
- The framework reduces input tokens by 54.1% and latency by 60.7% while maintaining high accuracy.
- It addresses the issue of unverifiable omissions in AI responses by linking summaries to raw logs.
Computer Science > Databases arXiv:2602.17913 (cs) [Submitted on 20 Feb 2026] Title:From Lossy to Verified: A Provenance-Aware Tiered Memory for Agents Authors:Qiming Zhu, Shunian Chen, Rui Yu, Zhehao Wu, Benyou Wang View a PDF of the paper titled From Lossy to Verified: A Provenance-Aware Tiered Memory for Agents, by Qiming Zhu and 4 other authors View PDF HTML (experimental) Abstract:Long-horizon agents often compress interaction histories into write-time summaries. This creates a fundamental write-before-query barrier: compression decisions are made before the system knows what a future query will hinge on. As a result, summaries can cause unverifiable omissions -- decisive constraints (e.g., allergies) may be dropped, leaving the agent unable to justify an answer with traceable evidence. Retaining raw logs restores an authoritative source of truth, but grounding on raw logs by default is expensive: many queries are answerable from summaries, yet raw grounding still requires processing far longer contexts, inflating token consumption and latency. We propose TierMem, a provenance-linked framework that casts retrieval as an inference-time evidence allocation problem. TierMem uses a two-tier memory hierarchy to answer with the cheapest sufficient evidence: it queries a fast summary index by default, and a runtime sufficiency router Escalates to an immutable raw-log store only when summary evidence is insufficient. TierMem then writes back verified findings as new summary u...
