2 papers
cs.AI2026
ClawVM: Harness-Managed Virtual Memory for Stateful Tool-Using LLM Agents
Mofasshara Rafique, Laurent Bindschaedler
Stateful tool-using LLM agents treat the context window as working memory, yet today's agent harnesses manage residency and durability as best-effort, causing recurring failures: l…
cs.DB2026
Semantic Caching for OLAP via LLM-Based Query Canonicalization (Extended Version)
Laurent Bindschaedler
Analytical workloads exhibit substantial semantic repetition, yet most production caches key entries by SQL surface form (text or AST), fragmenting reuse across BI tools, notebooks…