AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Harder Is Better: Boosting Mathematical Reasoning via Difficulty-Aware GRPO and Multi-Aspect Question Reformulation

Yanqi Dai, Yuxiang Ji, Xiao Zhang, Yong Wang, Xiangxiang Chu, Zhiwu Lu

ICLR 2026 · Reinforcement learning › RL from feedback · approach: GRPO stability and variants · 22 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 difficulty-balanced advantage estimation (DGAE) (ablation)DGPO · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 37.71 -> 38.65Improved
add difficulty-aware question weighting (DQW) (ablation)DGPO · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 38.65 -> 39.79Improved
temperature T=2.0 instead of 1.0 in DQW (ablation)DGPO · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 39.03 -> 39.79Improved
temperature T=2.0 instead of 10.0 in DQW (ablation)DGPO · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 39.27 -> 39.79Improved
replace GRPO with DGPOQwen2.5-Math-7B · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 37.61 -> 39.79Improved
add MQR data augmentation to DGPO (ablation)Qwen2.5-Math-7B · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 39.90 -> 42.17Improved
add MQR data to GRPO-trained model setup (MQR alone)Qwen2.5-Math-7B · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 37.61 -> 41.04Improved
add DGPO on top of DAPO (ablation)DAPO · Math benchmarks avg (DGPO combined with other methods)average accuracy: 37.94 -> 39.91Improved
add DGPO on top of GPG (ablation)GPG · Math benchmarks avg (DGPO combined with other methods)average accuracy: 37.93 -> 38.92Improved
add DGPO on top of GSPO (ablation)GSPO · Math benchmarks avg (DGPO combined with other methods)average accuracy: 37.71 -> 39.32Improved
use DGPO instead of GRPO on multimodal GeoQAQwen2.5-VL · GeoQA (multimodal)accuracy: 57.43 -> 59.95Improved
MQR reformulation instead of MetaMath rephrasing (ablation)DGPO · Math benchmarks avg (6 sets, trained on MATH)average accuracy: 40.73 -> 42.17Improved

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