ProGraph matches FullContext on multi-hop memory

Summary

Current LLM agent memory systems fail at multi-hop reasoning across sessions. ProGraph trades graph indexing overhead for substring traversal of LLM-generated profiles, reducing API calls while handling 5-hop queries—directly improving agent reliability in long-running applications.

Why it matters

Current LLM agent memory systems fail at multi-hop reasoning across sessions. ProGraph trades graph indexing overhead for substring traversal of LLM-generated profiles, reducing API calls while handling 5-hop queries—directly improving agent reliability in long-running applications.

Sources

1. 1.profile expansion -- substring-matched traversal of entity names that naturally appear in LLM-written profile narratives, a minimal alternative to explicit knowledge-graph construction
2. 2.compression residuals -- exact dates, quantities, and named items co-extracted with each profile update at zero extra API cost
3. 3.ProGraph averages 80.1% on MemHop (matching the FullContext reference) and 78.4% on LoCoMo (exceeding FullContext by 11.3pp)
4. 4.Long-term memory is essential for LLM agents that interact across sessions, yet current memory benchmarks primarily evaluate single-hop recall, leaving multi-hop association largely unmeasured