AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Memory Caching: RNNs with Growing Memory

Ali Behrouz, Zeman Li, Yuan Deng, Peilin Zhong, Meisam Razaviyayn, Vahab Mirrokni

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Sparse efficient transformers for long context · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
add Gated Residual Memory cachingTitans 760M · Commonsense reasoning averageaverage accuracy: 51.56 -> 52.55Improved
add Gated Residual Memory cachingDLA 1.3B · Commonsense reasoning averageaverage accuracy: 53.72 -> 55.96Improved
add Gated Residual Memory cachingTitans · In-context retrieval averageaverage accuracy: 31.75 -> 40.50Improved
add GRM cachingDLA · S-NIAH-2 16Kaccuracy: 28.2 -> 54.8Improved
use fixed-size segmentation instead of Log-Linear++ logarithmic segmentation (ablation)DLA · S-NIAH-2 16Kaccuracy: 18.0 -> 54.8Improved
add Log-Linear++ segmentationTitans 760M · Commonsense reasoning averageaverage accuracy: 51.56 -> 51.74No clear effect
make gamma depend on segment context, not just input (ablation)Titans (GRM) · Retrieval ablationaccuracy: 33.0 -> 40.5Improved
add input-dependent gating (ablation)Titans (GRM) · Retrieval ablationaccuracy: 32.4 -> 40.5Improved
use deep memory instead of linear memory (ablation)Titans (GRM) · Retrieval ablationaccuracy: 34.5 -> 40.5Improved
use separate u and q projections instead of shared (ablation)Titans (GRM) · Retrieval ablationaccuracy: 36.6 -> 40.5Improved
vary segment size from 8 to 128 (ablation)Memory Caching language modeling · Language modeling perplexityperplexity: 16.38, 14.91, 15.26, 17.09 for 8, 32, 64, 128Mixed
use checkpoints of one memory instead of independent compressors (ablation)Titans (GRM) · Language modeling and reasoningaccuracy: 57.2 -> 58.3Improved

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