memory-decay

The open-source benchmark for LLM memory decay. Measure how Naive, RAG, Chunked RAG, Cascading, and SummaryMemory degrade over 100 conversation turns. Ebbinghaus forgetting curves, 5-provider LLM eval, multi-seed CI. No API key needed.

Selective memory for AI agents importance-scored, self-decaying memory that forgets what doesn't matter so retrieval stays fast and relevant over long-running sessions.

interactive, sub-16ms vector explainer built for the DataForge 2026 Pathway Track.It proves why classical State Space Models suffer exponential associative forgetting ($\rho(A)^L$) and visualizes how Pathway's Dragon Hatchling (BDH) synaptic plasticity and BDH CQ continuous latent reasoning eliminate interference without KV-cache explosion.

Memory as advisory for LLM agents — forgets noise, keeps what matters. Lazy exponential decay, four-op API, OpenUSD cold tier. Python ≥3.11. Patent-pending.

A simulation that models how brain networks and memory traces slowly break down over time, similar to what happens in diseases like Alzheimer's.

Biologically-inspired Dual-Q reinforcement learning with asymmetric memory decay. MSc AI dissertation, University of Sussex 2025. Forgetting rewards faster than punishments improves adaptation in dynamic environments.