AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

An Imitation Learning Approach for Cache Replacement

Evan Liu, Milad Hashemi, Kevin Swersky, Parthasarathy Ranganathan, Junwhan Ahn

ICML 2020 · Reinforcement learning › Visual and representation RL · approach: RL with tree search for discovery · 5 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use end-to-end imitation learning from Belady's instead of LRUParrot · SPEC2006 and Web Searchcache hit rate: 16% higher cache hit rates than LRU averaged over all programs; 13.5% higher raw cache hit rate than LRU on Web SearchImproved
Use end-to-end imitation learning from Belady's instead of GliderParrot · SPEC2006 (13 programs)normalized cache hit rate: 20% higher normalized cache hit rate averaged over all programsMixed
Use Parrot instead of LRUParrot · Web Searchnormalized cache hit rate: 61% higher normalized cache hit rateImproved
Use the byte embedder instead of separate address and PC embeddings (ablation)Parrot · SPEC2006 (13 programs)normalized cache hit rate: full-sized model performs better; byte model achieves an average of 8% higher normalized cache hit rate than GliderHurt
Add reuse-distance prediction as an auxiliary task (ablation)Parrot · SPEC2006 (4 programs) and Web Searchnormalized cache hit rate: 16.8% higher normalized cache hit ratesImproved
Use reuse-distance prediction to directly choose the line with the highest predicted distance rather than as an auxiliary loss (ablation)Parrot · SPEC2006 (4 programs) and Web Searchcache hit rateHurt
Train on-policy with DAgger rather than on oracle-policy states (ablation)Parrot · SPEC2006 (4 programs) and Web Searchnormalized cache hit rate: 9.8% higher normalized cache hit rate overallMixed
Use ranking loss instead of log-likelihood loss (ablation)Parrot · SPEC2006 (4 programs) and Web Searchnormalized cache hit rate: 3.5% normalized cache hit rate improvementImproved
Increase the attended history length up to 80 accesses (ablation)Parrot · SPEC2006normalized cache hit rate: 80Improved
Increase history length beyond 80 accesses (ablation)Parrot · SPEC2006normalized cache hit rate: 80No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy