Curriculum learning
The DOI matches its Crossref record; 35 of the 40 references it cites were found in a registry or an index.
Abstract
Humans and animals learn much better when the examples are not randomly presented but organized in a meaningful order which illustrates gradually more concepts, and gradually more complex ones. Here, we formalize such training strategies in the context of machine learning, and call them "curriculum learning". In the context of recent research studying the difficulty of training in the presence of non-convex training criteria (for deep deterministic and stochastic neural networks), we explore curriculum learning in various set-ups. The experiments show that significant improvements in generalization can be achieved. We hypothesize that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and, in the case of non-convex criteria, on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Moreover, the lack of acoustical signals could even account for spatial disorientation 91,92.”
From The impact of multisensory integration and perceptual load in virtual reality settings on performance, workload and presence (2021) · cites this paper as [92]
In the citing paper
“This suggests that curriculum-style learning (Bengio et al., 2009), where simpler or archetypal problems are presented first to build a repository of valid heuristics, may potentially bootstrap performance.”
From Dynamic Cheatsheet: Test-Time Learning with Adaptive Memory (2025) · cites this paper as Bengio et al., 2009
Citations
Cites
Reducing the dimensionality of data with neural networks
Hinton, G. E., & Salakhutdinov, R. R. (2006) · Science · doi:10.1126/science.1127647 · Cited by 20,928
A fast learning algorithm for deep belief nets
Hinton, G. E., Osindero, S., & Teh, Y.-W. (2006) · Neural Computation · doi:10.1162/neco.2006.18.7.1527 · Cited by 16,336
Learning Deep Architectures for AI
Bengio, Y. (2009) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000006 · Cited by 9,239
Self-citationExtracting and composing robust features with denoising autoencoders
Vincent, P., Larochelle, H., Bengio, Y., & Manzagol, P.-A. (2008) · Proceedings of the 25th international conference on Machine learning - ICML '08 · doi:10.1145/1390156.1390294 · Cited by 7,436
Self-citationA unified architecture for natural language processing: deep neural networks with multitask learning
Collobert, R., & Weston, J. (2008) · Proceedings of the 25th international conference on Machine learning - ICML '08 · doi:10.1145/1390156.1390177 · Cited by 5,058 · cited 6×
“In Collobert and Weston (2008), the main objective is to learn an embedding for the words as a side-e.ect of learning to compute this score.”
BackgroundMethodInfluentialSelf-citationLearning Deep Architectures for AI
Bengio, Y. (2009) · now publishers, Inc. eBooks · doi:10.1561/9781601982957 · Cited by 4,941
Self-citationGreedy layer-wise training of deep networks
Bengio, Y., Lamblin, P., Popovici, D., & Larochelle, H. (2007) · Advances in Neural Information Processing Systems 19 · doi:10.7551/mitpress/7503.003.0024 · Cited by 4,841
Self-citationA Neural Probabilistic Language Model
Bengio, Y., Ducharme, R., Vincent, P., Jauvin, C., Jaz, K., Hofmann, T., Poggio, T., & Shawe-Taylor, J. S. (2000) · Cited by 3,479
Self-citationLearning and development in neural networks: the importance of starting small
Elman, J. L. (1993) · Cognition · doi:10.1016/0010-0277(93)90058-4 · Cited by 1,988 · cited 8×
“The idea of training a learning machine with a curriculum can be traced back at least to Elman (1993).”
BackgroundMethodInfluentialRestricted Boltzmann machines for collaborative filtering
Salakhutdinov, R., Mnih, A., & Hinton, G. (2007) · Proceedings of the 24th international conference on Machine learning · doi:10.1145/1273496.1273596 · Cited by 1,885
Research works printed but not linked: 3
- 10.5555/89603 · doi:10.5555/89603
- 10.5555/866740 · doi:10.5555/866740
- 10.5555/525571 · doi:10.5555/525571
Cited by
Items 4,501–4,600 of 5,240
Watching the Air Rise: Learning-Based Single-Frame Schlieren Detection
Achermann, F., Haug, J. A., Zumsteg, T., Lawrance, N., Chung, J. J., Kolobov, A., & Siegwart, R. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10611416
“In the first set of experiments, we used held-out flow data from our ROSCINE dataset, recorded in the indoor setup, together with images from the Places Validation Standard dataset to compare the DYBS predictions to the optical flow from the BOS algorithm.”
Contrastive Learning with Global Representation for Face Anti-spoofing
Wang, K., Dong, J., Feng, G., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-5594-3_36
Song, H., Dabre, R., Chu, C., Fujita, A., & Kurohashi, S. (2024) · Journal of Information Processing · doi:10.2197/ipsjjip.32.628 · Cited by 1
Unveiling the Decision-Making Process in Reinforcement Learning with Genetic Programming
Eberhardinger, M., Rupp, F., Maucher, J., & Maghsudi, S. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-7181-3_28 · Cited by 2
Ren, X., Wang, M., Qin, T., et al. (2024) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2024.08.19.608587
An Effective Sample Learning: Biasing Samples with High Class UncertaintyPreprint
Yang, M., He, Q., Huang, J., & Liu, Z. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4945276
Zhong, S. (2024) · IEEE International Geoscience and Remote Sensing Symposium (IGARSS) · doi:10.1109/igarss53475.2024.10640514
Learning to Deliberate: Multi-pass Decoding for Document-Grounded Conversations
Qiu, J., & Yang, Y. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650055
“The training loss is the sum of these three components: Furthermore, we employ curriculum learing [13] to train the model to improve performance and convergence rate [14].”
MethodAffinity-Weighted RandAugment for Problem-Oriented Augmentation
Park, Y., Takase, T., Kameyama, K., & Onishi, M. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651376
“Starting with learning simple data and transitioning to more complex data, to which RA is applied, may benefit from the curriculum learning [21].”
BackgroundDe-Redundancy Distillation And Feature Shift Correction For Cross-Domain Few-Shot Learning
Shuai, Q., Feng, L., Yue, L., Li, Z., & Xu, L. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651484
“Curriculum learning is a training strategy where easier tasks are given higher priority initially, followed by gradually increasing the weights assigned to more challenging tasks [23].”
BackgroundAbsorb What You Need: Accelerating Exploration via Valuable Knowledge Extraction
Wu, Y., Wu, Z., Yan, R., Tao, P., & Xing, J. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651098 · Cited by 1
“Curriculum learning [3]has shown promising results in contexts where rewards are sparse.”
BackgroundSuro, F., Michel, F., & Stratulat, T. (2024) · Artificial Intelligence · doi:10.1016/j.artint.2024.104228 · Cited by 1
Joint Self-Paced Learning and Concept Factorization for Multi-view Clustering
Hu, X., Xiong, D., & Chai, L. (2024) · Chinese Control Conference (CCC) · doi:10.23919/ccc63176.2024.10662328
Case-Based Deduction for Entailment Tree Generation
Shi, J., Ding, X., & Liu, T. (2024) · Mathematics · doi:10.3390/math12182893 · Cited by 1
Peng, N., Guo, S., Yuan, S., Kitsuregawa, M., & Gu, Y. (2024) · Web Intelligence · doi:10.3233/web-240279
“This paradigm was first raised and applie by Yoshua Bengio [3].”
Adaptive Data Optimization for Cross-domain Action Recognition in Low-Light Environment
Liu, H., Yang, H., & Wang, D. (2024) · IEEE Conference on Cybernetics and Intelligent Systems · doi:10.1109/cis-ram61939.2024.10673309
“And this progressive learning is possible to achieve results that are somewhat similar to course learning [20].”
BackgroundRepAn: Enhanced Annealing through Re-parameterization
Fei, X., Zheng, X., Wang, Y., Chao, F., Wu, C., & Cao, L. (2024) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52733.2024.00554 · Cited by 3
“In addition, other training techniques such as boot-strap [44], hard example mining [55], and curriculum learning [3, 24] can also be used for progressive training.”
BackgroundRobust Self-Paced Incremental Learning for Multitemporal Remote Sensing Image Classification
Li, H., Niu, P., Gong, M., Xing, L., Wu, Y., & Qin, A. K. (2024) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2024.3466309
“Bengio et al. [42] proposed the curriculum learning (CL) method in 2004.”
MethodSelf-Auxiliary Distillation for Sample Efficient Learning in Google-Scale Recommenders
Zhang, Y., Wang, R., Li, X., et al. (2024) · ACM Conference on Recommender Systems (RecSys) · doi:10.1145/3640457.3688041
“Different adaptors [14], curriculum learning strategies [2] and other functions (e.g. max) can be used as 𝑓 .”
Zhao, X., Qu, L., Feng, T., Cai, J., & Tang, B. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3681056 · Cited by 3
“2) simultaneously optimizing such entangled features 𝑥 = { 𝑥 ,𝑥 } for both text and visual modalities (i.e., imbalanced multimodal features) poses a special challenge for the classifier [24, 25].”
BackgroundFrolov, S., Moser, B. B., Palacio, S., & Dengel, A. R. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3680742
“Curriculum Learning: The idea of monotonically increasing the difficulty of tasks is related to curriculum learning (CL) [1] which introduces more complex concepts gradually, instead of randomly presenting training data.”
Improving clinical expertise in large language models using electronic medical recordsPreprint
Zhu, L., Liu, J., Wang, J., Zhang, W., Jiang, S., Yang, H., Wang, C., Ye, Q., Ruan, T., Rui, X., et al. (2024) · Research Square · doi:10.21203/rs.3.rs-5285540/v1 · Cited by 1
SACL: Sequential Augmentation with Curriculum Learning in Dataset Level
Ma, B., & Kong, F. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-9437-9_10
Chinese Personalized Commonsense Understanding and Reasoning Based on Curriculum-Learning
Yang, Y., Li, W., Fan, X., Deng, W., Liu, J., Diao, Y., & Tuerxun, P. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-9434-8_17 · Cited by 1
Instruction-Driven Game Engine: A Poker Case StudyPreprint
Wu, H., Liu, X., Wang, Y., & Zhao, H. (2024) · Qeios · doi:10.32388/ardyd9
Failed Goal Aware Hindsight Experience Replay
Kim, T., Jeong, H., & Har, D. (2024) · Lecture notes in networks and systems · doi:10.1007/978-3-031-70687-5_4
Manduca, G. M., Furnari, A., & Farinella, G. M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-78395-1_5
Clinical Learning Order-Guided Deep Neural Network for Brain Tumor Segmentation
Zhang, P., Jinke, W. A. N. G., Cheng, Y., & Tamura, S. (2024) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2024edp7222
“Currently, these learning patterns are generally embedded into deep neural networks through curriculum learning [14].”
Deep learning with Gaussian continuation
Ilersich, A. F., & Nair, P. B. (2024) · Foundations of Data Science · doi:10.3934/fods.2024053
MasterPlan: A Reinforcement Learning Based Scheduler for Archive Storage
Chen, X.-Y., Xu, E., Mo, D., Lu, R., Wu, H., Ding, D., & Xue, G. (2024) · ACM Transactions on Architecture and Code Optimization · doi:10.1145/3708542 · Cited by 1
“Besides, we employ a curriculum learning [12] strategy to prevent the agent from receiving massive negative rewards in the early stages.”
Kim, J., & Song, E. (2024) · Journal of electromagnetic engineering and science · doi:10.26866/jees.2025.3.r.261
Tazawa, R., Torii, T., & Hidaka, S. (2024) · International Conference on Intelligent Informatics and Biomedical Sciences (ICIIBMS) · doi:10.1109/iciibms62405.2024.10792875
“One effective resolution for these persistent challenges against DRL is curriculum learning [5].”
Assessing and Scoring Difficulty of Hard-to-Solve Data in Summarization Tasks
Jung, J., Seo, H., Namgoong, H., & Jung, S. (2024) · IEEE Access · doi:10.1109/access.2024.3519548
Speaker Diarization for Unlimited Number of Speakers Using Dynamic Linear
Wang, S., & Zhang, C. (2024) · International Symposium on Chinese Spoken Language Processing (ISCSLP) · doi:10.1109/iscslp63861.2024.10800134
Robot Generating Data for Learning Generalizable Visual Robotic Manipulation
Li, Y., Yuan, Y., Cui, J., Huan, H., Fu, W., Gao, J., Xu, Z., & Wu, Y. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10801742 · Cited by 4
MFC-EQ: Mean-Field Control with Envelope Q-learning for Moving Decentralized Agents in Formation
Lin, Q., & Ma, H. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802293 · Cited by 1
“Similar to [10], we employ the scheme of curriculum learning [29] to gradually introduce more challenging tasks into the training data.”
Villarreal, M., Poudel, B., Wickman, R., Shen, Y., & Li, W. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802857 · Cited by 7
“AutoJoin is also easier to implement than Shen et al. [3] as perturbations are randomly sampled within a moving range that is determined by linear curriculum learning [10].”
Shenouda, M., Whitney, H. M., Giger, M. L., & Armato, S. G. (2024) · Journal of medical imaging · doi:10.1117/1.jmi.11.6.064503
“The model employed a DenseNet-121 architecture, 16 chosen for its previous success in diagnosing pneumonia and other pathologies on CXRs. 17,18 In addition, it adopted a curriculum (transfer) learning approach, 19 increasing the focus of the classification task towards COVID-19 in the final phase.”
Ho, J., Wang, C., King, C., Feng, C.-W., & Sunkuan, S.-Y. (2024) · IEEE International Conference on Robotic Computing (IRC) · doi:10.1109/irc63610.2024.00024 · Cited by 1
“The proposed task clustering method is based on the concepts of task and cluster learning inspired by [5], which involves solving easier or similar problems first and then enhancing the ability of models later to handle more complicated or supplemental ones of curriculum [6].”
PINR: A physics-integrated neural representation for dynamic fluid scenes
Zhao, Z., Zhou, S., Lu, X., et al. (2025) · Neurocomputing · doi:10.1016/j.neucom.2024.129250
Target-Oriented Dynamic Denosing Curriculum Learning for Multimodel Stance Detection
Suo, Z., & Pan, S. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-2071-5_11
Using Curriculum to Train Multisensory Foraging DRL Agents
Filho, R. F. F., Santos, A. M. M., Rodrigues, H. R., Nogueira, Y. L. B., Vidal, C. A., Cavalcante-Neto, J. B., & Serafim, P. B. S. (2024) · Simpósio Brasileiro de Jogos e Entretenimento Digital (SBGames) · doi:10.5753/sbgames.2024.241119
“This brings more robustness and efficiency to the learning process [Bengio et al. 2009].”
A New Class Correlation-Based Dynamic Sample Weighting Method for Medical Image ClassificationPreprint
Yi, G., Ling, M., Liu, X., et al. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5070203
GIT-CXR: End-to-End Transformer for Chest X-Ray Report GenerationPreprint
Sîrbu, I., Sîrbu, I.-R., Bogojeska, J., & Rebedea, T. E. (2025) · Zürcher Hochschule für Angewandte Wissenschaften digital collection (Zurich University of Applied Sciences) · doi:10.48550/arxiv.2501.02598 · Cited by 4
“First introduced by Bengio et al. (2009), the vanilla form of curriculum learning means to gradually increase the difficulty of the data samples that the model sees during training.”
Balanced coarse-to-fine federated learning for noisy heterogeneous clients
Han, L., Zhai, Y., Jia, Y., Cai, Q., Li, H., & Huang, X. (2025) · Complex & Intelligent Systems · doi:10.1007/s40747-024-01694-8
“Self-paced learning, a variant of curriculum learning (CL) [26], is a training strategy inspired by human and animal learning processes.”
Yang, Y., Li, Y., Zhou, L., et al. (2024) · IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt61566.2024.10832171
A novel framework for join order selection based on reinforcement and representation learning
Liu, C., Kamali, A., Kantere, V., et al. (2024) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata62323.2024.10825394
A depth-aware geometric fusion based view synthesis method for sparse RGB-D input
Zhao, Z., Lu, X., Zhou, S., et al. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.126578
Learning From Mistakes: A Multilevel Optimization Framework
Zhang, L., Garg, B., Sridhara, P., Hosseini, R., & Xie, P. (2025) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2025.3534151
“For example, self-paced learning (SPL) [17] prioritizes training easier samples first, MentorNet [18] uses a meta-learning long short-term memory [38] to calculate weights for data with potentially corrupted labels, and curriculum learning [39] prioritizes easier training sets.”
A Skin Disease Classification Model Combining CAFormer and Curriculum Learning
He, X., & Chen, Y. (2024) · International Conference on Algorithms, Data Mining, and Information Technology (ADMIT) · doi:10.1145/3701100.3701118
“This approach helps the model better understand and master the knowledge, significantly improving both the training speed and the final performance of the model [3].”
MedProm: Bridging Dialogue Gaps in Healthcare with Knowledge-Enhanced Generative Models
Varshney, D., Behera, N. K., Katari, P., & Ekbal, A. (2025) · ACM Transactions on Computing for Healthcare · doi:10.1145/3715069 · Cited by 1
“According to the principle of curriculum learning, the optimization difficulty gradually intensifies from the objective 𝐽 ( 𝜃 1 ) to 𝐽 ( 𝜃 𝐿 ) [4].”
Node Fragility Reward Shaping Based on Reinforcement Learning in Intranet Penetration Scenario
Xu, Y., Tan, H., Deng, J., et al. (2024) · IEEE International Conference on Data Science in Cyberspace · doi:10.1109/dsc63484.2024.00050
“Bengio et al. [3] emphasize that the essence of curriculum learning lies in adjusting the distribution of learning samples gradually.”
Tao, Z. (2024) · International Conference on Algorithms, Computing and Systems (ICACS) · doi:10.1145/3708597.3708615
“The curriculum learning concept was introduced in 2009 [26], and has been used to guide agents from simple to complex tasks, thereby improving their ability to learn and generalize [25].”
UnSP: Improving event-to-image reconstruction with uncertainty guided self-paced learning
Yang, J., Zhang, X., Wang, S., Sun, Y., & Yin, B. (2025) · Displays · doi:10.1016/j.displa.2025.102985
Curriculum Reinforcement Learning for Tokamak Control
Kerboua-Benlarbi, S., Nouailletas, R., Faugeras, B., & Moreau, P. J. (2025) · Lecture notes in computer science · doi:10.1007/978-981-97-9536-9_3 · Cited by 1
Reinforcement Learning for Improved Alignment in Radiology Reports Generation
Shao, Y., Xu, R., Tan, C., Liu, G., Fang, T., & Huo, H. (2024) · China Automation Congress (CAC) · doi:10.1109/cac63892.2024.10864815
“This approach is inspired by the concept of Curriculum Learning by [22] and the methodology proposed by [23].”
Meta-Data-Guided Robust Deep Neural Network Classification with Noisy Label
Lu, J., Wang, Y., Shi, A., Ma, J., & Jin, Q. (2025) · Applied Sciences · doi:10.3390/app15042080
Lan-Bridge’s Submission to CCMT 2024 Translation Evaluation Task
Luo, M., Hu, G., & Wei, J. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-2292-4_11
CL4CO: A Curriculum Training Framework for Graph-Based Neural Combinatorial Optimization
Liu, Y., Zhou, C. D., Zhang, P., et al. (2024) · Proceedings · doi:10.1109/icdm59182.2024.00092
ralphi: a deep reinforcement learning framework for haplotype assemblyPreprint
Battistella, E., Maheshwari, A., Ekim, B., Berger, B., & Popic, V. (2025) · bioRxiv · doi:10.1101/2025.02.17.638151 · Cited by 4
“We used curriculum learning [6] for ONT training, which shows progressively harder examples to the agent as it trains, by ordering our graphs from the smallest and sparsest to the largest and most dense.”
Moment-Aware Video Retrieval for Video Corpus Moment Retrieval
Choi, Y., Ahn, D., & Choi, J. (2025) · IEEE Access · doi:10.1109/access.2025.3542720
“To address this challenge of temporal specificity, we introduce a curriculum-based hard negative sample mining strategy, inspired by the principles of curriculum learning [28], [33], [34], [35].”
Veranyurt, O., & Sakar, C. O. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06270-2
A Domain-Adapted Brain-Computer Interface for Implicit Emotion Tagging of Multimedia Content
Cui, Y., Liang, W., Yang, Y., et al. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-2373-0_29
Bărbălau, A. (2024) · International Conference on Artificial Intelligence, Robotics, and Communication (ICAIRC) · doi:10.1109/icairc64177.2024.10899939 · Cited by 1
“Curriculum learning [3, 14, 16] is a widely adopted strategy in deep learning, serving to improve model generalization capabilities and convergence time, by means of incrementally increasing training difficulty, starting from the easiest samples, and gradually moving to harder ones as the learning…”
Curriculum-Guided Adversarial Learning for Enhanced Robustness in 3D Object Detection
Huang, J., Xie, Y., Chen, Z., & Ye, S. (2025) · Sensors · doi:10.3390/s25061697 · Cited by 1
“By synthesizing the strengths of both curriculum learning and adversarial training [14,15], a novel framework named CGAL is constructed.”
ClingTP: Curriculum Learning based Multi-style Title Prefix Generation
Wang, Y., Li, D., Shen, J., Xu, Y., Zhong, S., & Xu, M. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890491
“Curriculum learning is a training approach that draws inspiration from the structured sequence in which humans learn, organizing examples in a specific order for more effective learning [23], [24].”
CPL: Curriculum Pseudo Labeling for Weakly Supervised Temporal Forgery Localization
Zhang, D., Fang, M., Lu, Z., & Xie, H. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10887844
“Curriculum learning [22]–[24] emulates the human learning process by presenting data sequentially from easy to hard, effectively handling noisy data and improving model convergence and generalization.”
Learning Markup Language Model for Composite Relationships Extraction
Lu, F., Duan, J., & Liu, J. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10890070
“To mitigate the risk of generating broken entity and relation spans or misused special tokens, MarkET trains a model with curriculum learning [20].”
Mitigating Gender Bias in Information Retrieval Systems
Seyedsalehi, S. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-88720-8_36 · Cited by 1
Leveraging LLMs for Energy Forecasting: The AcegasApsAmga Case Study
Roitero, K., Zancola, A., Mea, V. D., & Mizzaro, S. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-88720-8_25
Ethical Considerations and Challenges of Leveraging IoT and Machine Learning in Smart Urban Planning
Mohtashim, M. (2025) · IGI Global eBooks · doi:10.4018/979-8-3693-9030-6.ch007 · Cited by 1
Keitel, D., Modafferi, L. M., & Tenorio, R. (2025) · Springer Series in Astrophysics and Cosmology · doi:10.1007/978-981-96-1737-1_12
An overview of vulnerabilities and mitigation strategies in facial recognition systems
Garg, U. V., Rani, A., & Jaiswal, A. K. (2025) · AIP conference proceedings · doi:10.1063/5.0262020
Weaken Noisy Feature: Boosting Semi-supervised Learning by Noise Estimation
Liu, J., Sun, D., Xia, H., Bai, J., & Fan, X. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-5815-2_20
Real-Time Online Training Framework with Lesson-Based Curriculum for Minecraft Battle Bots
Yeh, F.-H., Ji, J., Chen, J.-L., Hung, M., Kao, M.-H., & Chen, C. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-5887-9_12
Learning Sensorimotor Synchronization Through Reinforcement in Neural NetworksPreprint
Ommi, Y., Yousefabadi, M., & Cannon, J. J. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5233815
Virtual Intelligent Autonomous Agents
Afolabi, A. S., & Akinola, O. A. (2025) · Advances in electronic government, digital divide, and regional development book series · doi:10.4018/979-8-3693-7832-8.ch014
Physics of Language Models: Part 4.1, Architecture Design and the Magic of Canon LayersPreprint
Allen-Zhu, Z., & Alfarano, A. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5240330 · Cited by 41
Liu, X., Zhi, X., Luo, J., He, Z., Mei, Q., Zhou, T., & Liu, M. (2025) · GeoInformatica · doi:10.1007/s10707-025-00548-w
“Adaptive loss functions, such as dynamic weight loss and curriculum learning [44], enable models to better adapt to complex data distributions by dynamically adjusting the loss form or training strategy.”
Intra-Class Patch Swap for Self-Distillation
Choi, H., Jeon, E. S., Shukla, A., & Turaga, P. K. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130408
“We explain the observed empirical phenomenon by drawing inspiration from the principles of curriculum learning [71].”
UBG: An Unreal BattleGround Benchmark With Object-Aware Hierarchical Proximal Policy Optimization
Niu, L., Li, B., Fan, X., Fang, H., Li, J., Xing, J., Wan, J., & Lei, Z. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3567001
“The game data logging feature also suggests the potential for applying reinforcement, imitation, and curriculum learning [49] algorithms.”
Many-Objective Neuroevolution for Testing Games
Feldmeier, P., Schmelz, K., & Fraser, G. (2025) · IEEE International Conference on Software Testing, Verification and Validation (ICST) · doi:10.1109/icst62969.2025.10988927 · Cited by 3
“This approach allows the test generator to employ a Curriculum Learning [13] strategy such that sophisticated gameplay can be evolved based on previously learned gameplay strategies.”
Cooperative Formation Control of USVs and UAVs Based on Reinforcement Learning
Wu, T., Ye, L., Li, X., & Peng, Y. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-5373-7_58 · Cited by 1
Guo, F., Zhai, J., Jia, J., Liu, J., Zhao, K., Yu, B., & Shi, C. (2025) · Design, Automation, and Test in Europe · doi:10.23919/date64628.2025.10992959 · Cited by 1
“Curriculum learning (CL) [28] is a training strategy that trains ML models from easier samples to harder samples, which imitates the meaningful learning order in human curricula.”
Zhang, X., Wang, S., Feng, F., & Liu, J. (2025) · The Visual Computer · doi:10.1007/s00371-025-03932-7
“It simulates the way humans learn and advocates having the model start with easy samples and progress to complex samples and knowledge [49], thus helping the model to better adapt to complex features [50].”
Distilling Gaming Strategy through Explainability in Tetris
Eleftheria, L., Georgios, L., Kosti, M. V., et al. (2025) · doi:10.5753/imxw.2025.4388
PDFT: parameter-diminish fine-tuning for transformer-based models
Zhang, M., Meng, W., Jia, M., Gu, J., Shao, Y., Wang, C., Xu, R., Ma, Z., & Zhang, X. (2025) · The Visual Computer · doi:10.1007/s00371-025-03973-y · Cited by 1
“Consequently, we design a strategy that dynamically adjusts the replacing rate at each optional position, drawing inspiration from contemporary pedagogical approaches [24].”
A Curriculum Learning Framework to Boost Object Detection of Unmanned Aerial Vehicles
Aslan, E. S., Naddeo, K., Bouhsine, T., Robi Polikar, A., & Ditzler, G. (2025) · doi:10.1109/cisdb64969.2025.11010781 · Cited by 1
“[4].”
FMW-Net: a first-order meta-weight-net approach for sample weighting
Zhou, Y., Shu, J., Liu, J., & Meng, D. (2025) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-025-02681-2
“Inspired by the curriculum learning [44], we followed the approach in [22] and selected the 10 images with the lowest loss for each class as the meta-dataset in each epoch iteration.”
Curriculum Learning for Assessing Thyroid Nodules Using Imbalanced Ultrasound Images
Keatmanee, C., Thongsreejun, S., Nakkrasae, S., Mahaweerawat, A., Nilnumpetch, C., Klabwong, S., & Nakaguro, Y. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-1747-0_37
Wan, D., Yang, X., Tang, W., & Zhang, W. (2024) · Asian Conference on Artificial Intelligence Technology (ACAIT) · doi:10.1109/acait63902.2024.11022199
“Unlike curriculum learning [19], self-paced learning does not require a predefined difficulty score, but uses the model to calculate the loss value of each sample as the corresponding difficulty score.”
Efficient Routing-Based Synthesis for Digital Microfluidic Biochips via Reinforcement Learning
Liu, B., Jiang, C., Xu, Q., Yao, H., Ho, T.-Y., & Yuan, B. (2025) · IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems · doi:10.1109/tcad.2025.3579327 · Cited by 1
“Curriculum Learning: CL is a machine learning paradigm, inspired by human learning, to master complex tasks through starting solving with easy tasks and progressing to complex ones [33].”
Khant, K. S., Lin, H. Y., & Thongtanunam, P. (2025) · International Conference on Mining Software Repositories (MSR) · doi:10.1109/msr66628.2025.00044 · Cited by 2
“By comparison, curriculum learning (CL), a difficulty-based training strategy inspired by pedagogical learning principles [7] has demonstrated promising results in domains such as natural language processing and computer vision [8].”
How Minds Take Shape: Graph-ESN Reveals How Neural Ensembles Engineer Stable RepresentationsPreprint
Pottore, K. V., & Ponnambalam, A. R. (2025) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2025.06.11.659086
Choose Your Words Wisely: Domain-Adaptive Masking Makes Language Models Learn Faster
Kohli, V. S., Monis, A., & Mamidi, R. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-95841-0_40
Curriculum Learning with Difficulty Division for Metaphor Detection
Hu, B., Wang, Y., Fang, F., Zhang, L., & Yin, P. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-7030-7_21
Haplo-caller: A deep learning method for haplotype identification from mixed clonal samplesPreprint
Lim, M. G., Chuenchoksan, C., Long, Y., Harrott, A. J. R., Ariani, C. V., & Hamilton, W. L. (2025) · doi:10.12688/verixiv.1060.1
Toward a General Transformer-Based Framework for Multi-agent Path Finding
Sun, S., Liu, H., Zhuang, L., & Li, H. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-6954-7_7
OAKD: Objective-Oriented Adaptive Knowledge Distillation via Curriculum-Guided Dynamic Weighting
Li, C., Tan, L., Teng, X., Tang, G., Ding, Y., & Lan, L. (2025) · Communications in computer and information science · doi:10.1007/978-981-96-6954-7_18
Large-Scale Contextual Market Equilibrium Computation Through Deep Learning
Ma, Y., Bian, Y., Xu, H. Y., et al. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-8312-3_27
Metadata
- Authors
- Bengio, Y., Louradour, J., Collobert, R., & Weston, J.
- Affiliations
- Université de Montréal (CA)
- NEC Laboratories America, Princeton, NJ
- ORCID
- Bengio, Yoshua
- Venue
- Proceedings of the 26th Annual International Conference on Machine Learning
- Publisher
- ACM
- Published
- 2009-06-14
- Pages
- 41-48
- OpenAlex
- W2296073425
- Semantic Scholar
- 8de174ab5419b9d3127695405efd079808e956e8
- Cited by
- 5,366
- References
- 40
- Citations by year
- 2026: 526 · 2025: 636 · 2024: 551 · 2023: 591 · 2022: 536 · 2021: 711 · 2020: 565 · 2019: 442 · 2018: 293 · 2017: 198 · 2016: 103 · 2015: 63 · 2014: 41 · 2013: 40 · 2012: 33
- FWCI
- 16.17
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- International Conference on Machine Learning (publishedVersion)
- Queen Mary Research Online (Queen Mary University of London) (submittedVersion)
- Funders
- Canadian Institute for Advanced Research
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Stochastic Gradient Optimization Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- Keywords
- curriculum learning, training order, non-convex optimization, deep neural network, stochastic neural networks, deterministic neural networks, generalization, convergence rate, numerical continuation, global optimization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.