AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

EHL*: Memory-Budgeted Indexing for Ultrafast Optimal Euclidean Pathfinding

Jinchun Du, Bojie Shen, Muhammad Aamir Cheema

AAAI 2026 · Architectures and efficiency › RNNs and normalization · approach: Distributed LLM inference systems · 1 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
reduce EHL* memory budget from 80% to 20% of EHL-1EHL* · DAO (unknown queries)query time (us): 2.02 -> 2.49No clear effect
reduce EHL* memory budget from 20% to 5% of EHL-1EHL* · DAO (unknown queries)query time (us): 2.49 -> 4.40Hurt
use known query workload when merging (Cluster-2) at 5% budget (ablation)EHL* · DAOquery time (us): 4.40 -> 1.15Improved
merge cells by hub-label similarity under budget instead of larger EHL cell size (EHL-4)EHL*-20 · DAO (unknown queries)query time (us): 3.19 -> 2.49Improved
increase number of query clusters from 2 to 8 at 5% budgetEHL*-5 · DAOquery time (us): 1.15 -> 2.00Hurt
build index with known workload when only 20% of queries match prediction (20% budget) (ablation)EHL* (Known) · DA Cluster-8, 20% on-distribution queriesquery time (us): 2.08 -> 3.09Hurt
build index with known workload when all queries match prediction (20% budget) (ablation)EHL* (Known) · DA Cluster-8, 100% on-distribution queriesquery time (us): 2.04 -> 1.64Improved

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