AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

ATLAS: Alibaba Dataset and Benchmark for Learning-Augmented Scheduling

Zhiyun Jiang, Tianming Zhao, Chunqiu xia, Albert Zomaya

ICLR 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Algorithms with explicit predictors

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 task-signature features and empirical Bayes shrinkage via LGBM meta-stackingSingle-Stage baseline · ATLAS PAI predictionCov@50 (%): 36.2 -> 70.6Improved
use classification-first routing (Two-Stage) for recurring jobsLGBM-Meta · recurring jobsCov@50 (%): 77.5 -> 78.0Improved
use signature-based predictions instead of HRAS historyHRAS · ATLAS PAI predictionRMSLE: 1.373 -> 0.871Improved
use Weighted-Rec temporal ensemble instead of meta-stackingLGBM-Meta · ATLAS PAI predictionCov@25 (%): 51.2 -> 39.9Hurt
fully commit to predicted ranking (SPJF) rather than hedge with predicted round-robin (PRR)Meta predictor · Total completion timeratio to SRPT: 1.292 -> 1.106Improved
replace HRAS predictions with Meta predictions in SPJFHRAS-SPJF · Total completion timeratio: 1.686 -> 1.106Improved
use predicted-size SPRPT instead of non-preemptive LAS/FBLAS/FB · Max-stretchrho_S,max: 205.28 -> 18.13Improved
use SPRPT instead of EDF-P with Meta predictionsMeta-EDF · Max-stretchrho_S,max: 204.66 -> 18.13Improved
use EDF-P with Meta predictions instead of LAS/FB for median stretchLAS/FB · Median stretchrho_med: 187.31 -> 4.504Improved
use HRAS predictions with EDF-P instead of Meta predictionsMeta-EDF · Max-stretchrho_S,max: 204.66 -> 1700.5Hurt
use Iso predictor for LPPT instead of MetaMeta-LPPT · Makespanratio: 1.689 -> 1.574Improved

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