AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

An Information-Theoretic Criterion for Efficient Data Synthesis

Hanyu Li, Zhengqi Sun, Xiaotie Deng

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 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
train the judge iteratively on its own outputs with no external rubric anchorImplicit-rubric judge model · Instruction-following judgingacceptance rateHurt
use a binary all-rubrics-satisfied reward (coarse signal) despite a noisy judgeRubric-based RL for instruction following · Out-of-distribution generalizationgeneralization: judge consistency 24%Improved
rebalance positive/negative sources so length and style are decorrelated from correctnessMath/code/logic judge model · Balanced held-out correctness setjudging accuracyImproved
assign all positives to one model family and all negatives to another (spurious length/style signal)Math/code/logic judge model · Balanced held-out correctness setcorrectness accuracyHurt
train judge with binary correctness on math only (high meta-level signal)JudgeRLVR judge · Code, logic, general reasoning transfercross-domain judgingImproved
add mastery-gated length penalty to RL (math only training)CoT compression RL · Code, instruction following, general QAreasoning lengthImproved
separate a fixed evaluator agent from the generator agentLong-running agentic application harness · Agentic app developmentoverall qualityImproved

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