Curriculum Learning

Follow this topic Ask about this topic

Curriculum learning studies whether the order and difficulty of training can help models learn more efficiently or generalize better. Common approaches include easy-to-hard schedules, self-paced sampling based on loss or confidence, adaptive teacher-student selection, and staged curricula for tasks, model representations, or reinforcement-learning environments. It is used in settings from language-model reasoning and preference tuning to robotics, vision, and scientific prediction, with evaluations on reasoning benchmarks, image datasets, and simulated or physical control tasks. Results are mixed: curricula can improve learning speed and data efficiency, but benefits depend on how difficulty is measured, how pacing matches the learner, and whether comparisons control for data exposure and training budget.

Latest papers

176 papers29 papers in the last four weeks, up 53% on the four weeks before. 0.2% of all new papers.

Heterogeneous Co-Evolution Learning for Tool-Efficient Search Agents with Zero Labels
Oct 29, 2026Qin-Wen Luo, Sheng Ren, Xiang Chen +6LLM Agent Training LLM Agent Efficiency
Nanjing University of Aeronautics and Astronautics, Nanjing, China · Didichuxing Co. Ltd, Beijing, China
Lev-3B: A Data Recipe for Compact Probabilistic Log Decision Models
Oct 28, 2026Leo Betthauser, Yingqi Liu, Uday Shah +5Language Model Calibration Curriculum Learning
Cisco
Evolve on Graph: Self-Evolving Knowledge Graph Reasoning without Question Annotations
Oct 26, 2026Zongyuan Sun, Shuai Zhao, Ke Shu +8LLM Reasoning with Graphs Knowledge Graph Question Answering
Beijing University of Posts and Telecommunications, Beijing, China · University of Helsinki, Helsinki, Finland · University of Amsterdam, Amsterdam, Netherlands +1
Curriculum-Guided Acquisition of Hierarchical Task Knowledge from LLM-Generated Experience
Oct 26, 2026Yuesheng Xu, Hector Munoz-AvilaCurriculum Learning LLM Planning
Computer Science & Engineering, Lehigh University; Bethlehem, PA 18015-3084 USA
MTA-Zero: Self-Improving Multimodal Deep Search via Online QA-Solver Co-training
Oct 24, 2026Xiangyu Peng, Jiaxin Zhang, An Yan +4Multimodal Search Agents Curriculum Learning
Supervision-Adaptive Curriculum and Guarded Anti-Curriculum Learning for Dense Prediction
Oct 21, 2026Chunming He, Dingming Zhang, Fengyang Xiao +2Curriculum Learning Dense Prediction
Department of Biomedical Engineering, Duke University, Durham, NC 27708 USA · Adobe Inc., San Jose, CA, USA
Curriculum Learning for Generative Flexible Docking
Oct 20, 2026Wenyin Zhou, Hossein AzizpourStructure-Based Drug Design Curriculum Learning
KTH Royal Institute of Technology · Science for Life Laboratory
GRADUS: Discovering and Recovering Agent Lessons from the Environment's Latent Syllabus
Oct 20, 2026Qing Ping, Wei Yang, Panpan Xu +4LLM Agent Training LLM Agent Trajectory Synthesis
AWS AI · University of Southern California · BMW Group Research and Technology
Predictive Bandit Curriculum Reinforcement Learning via Gradient-Based Sample Group Discovery
Oct 19, 2026Liang Xu, Shufan Shen, Zhongni Hou +5Curriculum RL Curriculum Learning
State Key Lab. of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy of Sciences · Meituan
Curriculum Learning Teaches LLMs Long-Horizon Implicit Reasoning, Up To A Limit
Oct 19, 2026Jimson Huang, Maxwell Nickell, Ruowang Zhang +1Long-Horizon LLM Reasoning Implicit Reasoning in Language Models
Purdue University West Lafayette, IN 47907, USA
PRECOCIAL: Pretraining with Code Curriculums via Iterative Search for Accelerated Learning
Oct 18, 2026Edward Gu, Ali Payani, Kevin EllisSynthetic Data Pretraining Synthetic Data Generation for Language Models
Cornell University · Cisco Research
CRISP: Progressive Illumination Curriculum for Recurrent Low-Light Video Enhancement with Semantic Priors
Oct 16, 2026Ruirui Lin, Xin Yan Lim, Yini Li +2Video Restoration Low-Light Image Enhancement
University of Bristol
Perceive, Solve, Reason: Progressive Capability Formation in Time Series-LLMs
Oct 16, 2026Xin Qiu, Junlong Tong, Wei Zhang +1Time Series Reasoning Time Series Language Modeling
EIT-NLP Lab, Eastern Institute of Technology, Ningbo · Zhejiang University · Shanghai Jiao Tong University
Path to Prior-Data Fitted Relational Models via Learned Latent Multi-table Representations
Oct 16, 2026Mohammad Sadeq Abolhasani, Viswanath GanapathyRelational Foundation Models Synthetic Data Pretraining
SAP Labs
FCTS: Failure-Guided Curriculum Tree Search Over Generated Terrain for Legged Parkour
Oct 15, 2026Narayanan Palghat Parameswaran, Dong-Ki Kim, Jay Patrikar +4Legged Robotics Curriculum Learning
Carnegie Mellon University · FieldAI
Language Exposure Matters: Distribution and Order in Multilingual BabyLM Pre-training
Oct 14, 2026JiWoo Hwang, Lukas Edman, Alexander FraserLanguage Model Pretraining Multilingual Language Models
School of Computation, Information and Technology, TU Munich · Munich Center for Machine Learning · Munich Data Science Institute
Where to Start, How to Move: Analyzing Curricula under GRPO for Long-Horizon LLM Agents
Oct 14, 2026Minchan Kim, Sungho Bae, Youngjin Kim +1Long-Horizon LLM Agents Curriculum RL
Graduate School of Data Science, Seoul National University · Department of Computer Science and Engineering, Konkuk University
RSIgen: Reciprocal Self-Improvement for Agentic Visual Generation
Oct 13, 2026Yidi Liu, Haozhe Wang, Jie Xiao +3Self-Improving Agents Curriculum Learning
University of Science and Technology of China · The Hong Kong University of Science and Technology · Wan Team, Alibaba Group
MedEvo: Medical Reasoning Evolution through Knowledge Navigation
Oct 13, 2026Xiaoliang Wang, Zhuo Chen, Chuanyang Zheng +11Medical Reasoning Multimodal Reasoning
Zhejiang University · Ant Group · Shanghai Artificial Intelligence Laboratory +1
Hierarchical Curriculum Learning for Surgical Video Generation: Focusing on Bilateral Axillo-Breast Approach Robotic Thyroidectomy Datasets
Oct 13, 2026Youngbin Kong, Seungho Lee, Jae Kyung Shin +3Video Generation T2V Generation
Interdisciplinary Program in Medical Informatics, Seoul National University College of Medicine, Seoul, Republic of Korea · Department of Transdisciplinary Medicine, Seoul National University Hospital, Seoul, Republic of Korea · Department of Surgery, Seoul National University Hospital, Seoul, Republic of Korea +4
Calibrated Disambiguation for Partial Multi-label Learning
Oct 13, 2026Zhuoming Li, Yuheng Jia, Mi Yu +1Multi-Label Classification Noisy-Label Learning
School of Computer Science and Engineering, Southeast University · China Telecom Cloud Computing Corporation, Beijing 100088, China
Anchor-and-Harvest Curriculum: A Fully Unmasked Non-LLM Curriculum for Craftax
Oct 11, 2026Jiayu Zhu, Luyuan Li, Yuyuan Shi +1Curriculum RL Curriculum Learning
Brown University · York University · Sun Yat-sen University +1
Self-Evolving Visual Agent via Margin Reward
Oct 10, 2026Wenxi Yang, Yuzhong Zhao, Baochang Zhang +2Tool-Using Vision-Language Agents RL for Tool Use
UCAS · Microsoft Research Asia · Beihang University
A Step Harder, Not a Step Too Far: The Benefits and Limits of Curriculum Difficulty Progression in GRPO
Oct 9, 2026Ronghan Li, Fei Chen, Mingyu LuCurriculum RL Curriculum Learning
College of Artificial Intelligence, Dalian Maritime University, Dalian, 116026, Liaoning, China. · Information Science and Technology College, Dalian Maritime University, Dalian, 116026, Liaoning, China.
NV-Tesseract-AD: Curriculum-Enhanced Conditional Diffusion for Multivariate Time-Series Anomaly Detection
Oct 8, 2026Aditi Gautam, Manikandan Ravikiran, Saira Qureshi +4Time-Series Anomaly Detection Reconstruction-Based Anomaly Detection
NVIDIA
When to Unpair: Regulating Pairing Dependence in Medical Visual In-Context Learning
Oct 7, 2026Cheng Wan, Chenjun Li, Qingyu ZhaoIn-Context Learning Curriculum Learning
Cornell University · Weill Cornell Medicine
Algorithmic Scratchpads and Curriculum Staging for Arithmetic Reasoning in Tiny Transformers
Oct 6, 2026Sourabh KasliwalNumerical Reasoning in Language Models Neural Algorithmic Reasoning
DART-ES: Difficulty-Aware Reweighting and Targeted Replay for Fine-Tuning LLMs with Evolution Strategies
Oct 4, 2026Zhishen Sun, Hongzhan Wang, Sizhe Dang +2LLM Fine-Tuning Evolution Strategies
Xi’an Jiaotong University · SGIT AI Lab
Anticipating the Consequences of Curriculum Decisions with Large Language Models
Oct 3, 2026Octavio Pappalardo, Nathan Herr, Tim RocktäschelCurriculum Learning Curriculum RL
University College London
ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization
Oct 1, 2026Sungho Park, Wonjoong Kim, Jue Zhang +8Agent Harness Optimization Curriculum Learning
POSTECH · KAIST · Microsoft

Oct 29, 2026J/K move · Enter open · S save

Heterogeneous Co-Evolution Learning for Tool-Efficient Search Agents with Zero Labels

Qin-Wen Luo, Sheng Ren, Xiang Chen, Rui Liu+5

Nanjing University of Aeronautics and Astronautics, Nanjing, China · Didichuxing Co. Ltd, Beijing, China

LLM Agent Training LLM Agent Efficiency Heterogeneous MARL Curriculum Learning RL for IR

Open paper AskPDF arXiv

Deep search agents extend language models by interleaving reasoning with external retrieval. However, effective search-agent training requires addressing two challenges beyond final-answer optimization. First, the capability frontier of agents is largely determined by training questions, whereas fixed or labeled task collections cannot adapt as agents improve. Second, outcome-only reinforcement learning treats all successful trajectories equally, even when some correct answers rely on redundant searches, creating an inherent tension between answer accuracy and tool efficiency. To address these issues, we introduce ZEST-Search, a zero-label and tool-efficient framework for search-agent training via heterogeneous co-evolution. Starting from unlabeled reference questions, ZEST-Search constructs validated search tasks online through an answer-first and evidence-path-guided generation process. A heterogeneous proposer adaptively constructs training tasks around the solver's evolving capability boundary, while the solver learns search-augmented reasoning through reinforcement learning. To jointly optimize task difficulty and search efficiency, ZEST-Search employs an intermediate-difficulty reward for the proposer and a task-local tool-aware reward for the solver, encouraging informative tasks and correct answers with fewer redundant searches. Experiments on seven general-domain and multi-hop question answering benchmarks demonstrate that ZEST-Search consistently improves search-agent performance while achieving higher tool efficiency without labeled search-training data or expert search trajectories. Further analyses show that heterogeneous generation, adaptive difficulty control, and tool-aware optimization enable more effective search-agent training.