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 4,901–5,000 of 5,240
Compositionality as a mechanism of intelligence in brains and machines
Zerroug, A. (2023) · theses.fr (ABES) · doi:10.70675/cd39b7d4z0eafz464az87caz2c10307cb623
Towards Out-of-Distribution Generalization for Image Manipulation Detection
Abecidan, R. (2024) · theses.fr (ABES) · doi:10.70675/ef72a576z3ff7z4702z80b4zc2b94ed8a046
Generative Regression Based Watch Time Prediction for Short-Video Recommendation
Ma, H., Tian, K., Zhang, T., Zhang, X., Zhou, H., Jin, C., Chen, C., Li, H., Guan, J., & Zhou, S. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792290 · Cited by 25
“To accelerate model convergence, we adopt curriculum learning [2] strategy during training to alleviate training-and-inference inconsistency, commonly known as exposure bias [8, 36].”
Task Specification and Generalization in Multi-Task Reinforcement Learning
Cachet, T. (2025) · theses.fr (ABES) · doi:10.70675/08eadb30z8466z4763zbba6za6a4dab9a29e
Li, G., Xu, S., Hao, Y., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6550386
Zhou, P., zhu, Z., Bing, W., et al. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6561548
Markudova, D., & Meo, M. (2026) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2026.3683265
“We also experimented with curriculum learning to investigate whether training could be improved by gradually exposing the agent to more diffi-cult traces instead of randomly sampling from the full dataset [37].”
Wang, X. (2026) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2026.104764
Jia, X., & Chen, X. D. (2026) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2026.3678824
“Curriculum learning is a machine learning concept where models are trained on data in a meaningful order, starting with simpler examples and gradually increasing in complexity [24], [25].”
Mo, X., Tan, S., Peng, R. C., Li, B., & Huang, J. (2026) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2026.3684843
“To address the final challenge, we employed curriculum learning [37] to facilitate the convergence of our agents’ learning process.”
Seelig, L. F. C., & Salles, R. M. (2026) · Ad Hoc Networks · doi:10.1016/j.adhoc.2026.104263
Alam, E., Littlefield, N. G., Shaban‐Nejad, A., & Moradi, H. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1793305 · Cited by 1
“This structured approach holds the potential to enhance model robustness, accuracy, and fairness (Bengio et al., 2009; Hacohen and Weinshall, 2019; Weinshall et al., 2018).”
Multi-View Stereo With Holistic Guidance of Collaborative Geometric Priors
Han, Y., Wang, C., Han, H., & Wu, F. (2026) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2026.3685186
“On the Tanks [42], IterMVS [43], GeoMVSNet [28], and ground truth on scan10, 34, 62, and 75 from the DTU dataset. and Temples dataset, we input 11 images with a resolution of 1920 × 1024, and the fused point clouds are submitted to the o ffi cial website for online evaluation.”
Mistake gating leads to energy and memory efficient continual learningPreprint
Pache, A., & van Rossum, M. C. W. (2026) · bioRxiv · doi:10.64898/2026.04.16.718919 · Cited by 1
“It also naturally combines with curriculum learning [Bengio et al., 2009, Wang et al., 2022], as well as class imbalanced datasets [Johnson and Khoshgoftaar, 2019], since rare classes would be classified incorrectly more often.”
Curriculum-Guided Reasoning Reranker for Multimodal Document Retrieval
Li, Y., Gao, H., & Wang, C. (2026) · 2026 5th International Symposium on Computer Applications and Information Technology (ISCAIT) · doi:10.1109/iscait69154.2026.11477488
“To mitigate the instability inherent in standard Reinforcement Learning , we integrate Group Relative Policy Optimization (GRPO) with a progressive dual-curriculum design [11] and a Rank-Calibrated DCG reward system.”
Liu, Y., Liu, F., Zhao, Z., Shen, Q., & Kuang, Y. (2025) · International Conference on Artificial Intelligence, Digital Media Technology and Interaction Design (ICADI) · doi:10.1145/3795926.3796025
“Curriculum learning [1] presents training examples in meaningful order from easy to hard.”
Henríquez, N., Benítez, J., Camacho, C., Meneses, C., & Prado-Romo, A. J. (2025) · IEEE CHILEAN Conference on Electrical, Electronics Engineering, Information and Communication Technologies (CHILECON) · doi:10.1109/chilecon66915.2025.11476082
“This staged learning process facilitates stable skill acquisition, reduces possibility of early convergence or policy collapse, and enhances the agent’s generalization capabilities across diverse fight scenarios [17].”
Robust Multiuser Tracking in Indoor IoT Spaces via Low-Resolution mmWave Radar
Tang, N., Zhang, J., Liu, L., et al. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3685254
“To stabilize optimization under increasing scene complexity, we adopt a progressive curriculum training strategy [36] aligned with the spectral-ambiguity decomposition objective.”
Hong, Z., Chen, R. R., Li, L., et al. (2026) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2026.110299
V2A-DPO: Omni-Preference Optimization for Video-To-Audio Generation
Chan, N., Gang, T., Wang, Y., Liang, Y., & Wang, D. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463624
“By incorporating a curriculum learning [20] paradigm into the DPO optimization process, our human preference-aligned V2A models significantly enhance the perceptual quality, semantic consistency, temporal alignment, and even aesthetic appeal of the generated audio.”
Online Test-Time Adaptation for Shadow Segmentation
Wu, D., Xia, H., Sun, L., et al. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462548
Du, Y., Zhang, J., Wang, B., Qiu, J., Huang, L., Ge, Y., Liu, X., Xiao, T., & Zhu, J. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464611
Liu, R., Pu, J., Chen, Y., Miao, Y., Liu, D., & Wang, Q. (2026) · Remote Sensing · doi:10.3390/rs18081250
Triage Knowledge Distillation for Speaker Verification
Kim, J.-H., Jung, Y., Yang, J.-Y., Roh, J., Han, C. W., & Cho, H. Y. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11461888
“We propose Triage KD (TRKD) , which instantiates the assess– prioritize–focus triage principle within the distillation objective, building on difficulty-aware curricula and selective supervision [10–12].”
Curriculum Learning with Contrastive Loss for Lightweight Speaker Verification
Li, J., Mak, M., Rohdin, J., & Plchot, O. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11462096
Region Growing Physics-Informed Neural Network for Wind Field Reconstruction from Sparse Data
Liu, L., Hu, Z., Zhao, H., et al. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11464117
Structsoup: A Unified Adaptive Framework for Structure-Aware Retrieval-Augmented Generation
Gao, M., Shi, C., Gao, Z., & Kong, Y. (2026) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp55912.2026.11463625
Peng, S., Han, Y., Liu, L., Li, J., Li, R., Gu, Z., Liu, J., & Chu, X. (2026) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2026.3683936 · Cited by 1
“The class weight ω y j is determined by the sample’s label according to: where the anomaly class weight ω 1 adapts dynamically with training progress: where implements a curriculum learning schedule [42] that gradually increases emphasis on anomaly samples from 1 .”
Song, Y., Lv, R., Niu, Z., & Chen, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6636734
CoDA: A Cognitive-Inspired Approach for Domain Adaptation
GEMİRTER, C. B., Korkmaz, E. E., & Goularas, D. (2026) · Applied Sciences · doi:10.3390/app16094115
Can curriculum learning overcome structural disparity in MP-GNNs?
Pareek, U., Pandey, R., Shah, K., Achary, B. S., Gupta, S., & Asthana, S. (2025) · ACM IKDD Conference on Data Science · doi:10.1145/3799830.3799846
“Curriculum learning (CL), which mimics the human process of learning data samples in a meaningful sequence, has demonstrated its effectiveness in improving model performance, generalization, robustness, and even convergence across various domains [2, 17].”
Niu, Z., Chen, S. S., & Zhang, C. (2025) · International Conference on Algorithms, Computing and Systems (ICACS) · doi:10.1145/3789418.3789423
“On ImageNet-1k selecting only 20 % of the classes uniformly causes a 3.8 % top-1 drop relative to stratified selection [57], because rare classes are under-represented and their gradients are estimated with high variance.”
InfluentialGenetic Feature Selection and Multistage Deep Learning for Bridge Substructure Condition Prediction
Awuku, B., & Asa, E. (2026) · Journal of Performance of Constructed Facilities · doi:10.1061/jpcfev.cfeng-5198
Learned Optimal Visual Time-of-Flight Imaging With Fisher Information Guidance
Wang, K., Li, J., Li, Z., Wang, K., Kong, X., Yan, F., Yue, T., & Hu, X. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3688946
Periodic Motion Generation with Curriculum Learning: A Case Study on Fashion Show Catwalk Sequencing
Yan, Z., Hu, Z., Chang, Y., Zhong, T., Zhang, Y., & Deng, Y. (2025) · China Automation Congress (CAC) · doi:10.1109/cac67268.2025.11487785
“Curriculum learning is a training strategy that gradually increases the difficulty of the learning task to improve model performance and generalization [11].”
Zhang, J. F., Li, Q., & Wan, H. (2025) · China Automation Congress (CAC) · doi:10.1109/cac67268.2025.11487875
“To address the training difficulties encountered in PONNs, the concept of curriculum learning can be incorporated into the training process [25].”
Cho, H., Jeong, Y., & Yoon, K. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00583
“B, providing an overview of the available images, ground truths, and additional data characteristics such as stereo disparity, intermediate flow, and distance maps”
Debiased Curriculum Adaptation for Safe Transfer Learning in Chest X-Ray Classification
Liu, M., Chen, X., Yang, S., Li, X., Guan, W., & Nie, L. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02099
“To facilitate safe knowledge transfer, inspired by the idea of curriculum learning [4, 12, 38], we propose an easy-to-hard way to transfer from source to target with curriculum adaptation.”
CECKG: A Credible Entity Classification Method for Knowledge Graph
Yu, S., Li, Q., Liu, M., Xu, H., & Wang, Z. (2026) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2026.3688922
“For example, Bengio et al. [23] suggested that simple instances are more beneficial than difficult ones, as difficult instances might confuse the classifier, leading to the establishment of unreasonable decision boundaries and overfitting problem.”
Xiong, S., Chen, S., Liu, Y., et al. (2026) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-026-18199-y
“Progressive distillation addresses capac - ity gaps through intermediate models and curriculum learning [2].”
Bandara, N. S., Kandappu, T., & Misra, A. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00141 · Cited by 1
“Across Graph Heterogeneity [5] with progressive training stages (Algorithm 1).”
A Dataset and Framework for Learning State-invariant Object Representations
Sarkar, R., & Kak, A. C. (2026) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61042.2026.00458 · Cited by 2
“(B) Curriculum Learning: As introduced in [2, 19], curriculum learning trains models by gradually increasing the data complexity from easy to hard cases.”
Two-scale neural networks for singularly perturbed dynamical systems with multiple parameters
Zhuang, Q., Wang, T., Wanjiku, R., et al. (2026) · Results in Applied Mathematics · doi:10.1016/j.rinam.2026.100766
Islam, M. S., Hasan, M. R., & Khatun, M. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11490569
“Additionally, while common regularizers like dropout and label smoothing are widely used, advanced techniques such as stochastic depth [19] and curriculum learning [20] remain underutilized in Bangla captioning pipelines.”
Advancing Intelligent Energy Harvesting in IoT Networks Using Agentic and GenAI Approaches
Majumder, A. J. A., Kuchi, A., Gibb, J., et al. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11490533
“Agentic AI will enable the system to make real-time adjustments, prioritize tasks, and coordinate energy sharing among nodes without external intervention [11].”
Research on UAV Path Planning and Obstacle Avoidance Integrating PPO and DWA
Wang, Z., & Shen, S. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-7254-0_21
Research on Key Algorithms for Text Auto-Correction Based on DeepSeek
Wei, Y., Zhang, B. H., Li, Q., et al. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-7254-0_31
Liu, S., Mei, Y., & Zhang, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-23604-3_7
Dual pathway architecture in songbirds enables robust sensorimotor learningPreprint
Sankar, R., Suryawanshi, A., Rougier, N. P., & Leblois, A. (2026) · eLife · doi:10.64898/2026.05.07.723469 · Cited by 1
Bridge-SQL: Bridging Single- and Multi-Turn Text-to-SQL via Preference-Aligned Question Rewriting
Zhu, X., Tang, X., Wu, S., Tang, Y., Wang, H., Yao, C., Xia, R., & Chen, G. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-0372-7_8 · Cited by 1
Construction and optimization of interpretable models for QT500-7 laser hardening process parameters
Liang, Q., Chen, H. Y., Jin, F., et al. (2026) · Optics & Laser Technology · doi:10.1016/j.optlastec.2026.115459
Wang, H., Xi, L., Zhang, R., Li, H., & Xie, X. (2026) · Lecture notes in electrical engineering · doi:10.1007/978-981-95-8329-4_13
Yang, W., Shang, W., Wang, L., et al. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3691776
“A carefully designed three-phase training strategy accommodates the unsupervised learning of pore aspect ratio ( α ), with adaptive physics weight scheduling to ensure stable convergence [27].”
Rebalancing Curriculum Learning via In-Batch Difficulty Reallocation for Neural Machine Translation
Eo, S., & Park, C. (2026) · Mathematics · doi:10.3390/math14101607
Li, X.-Z., Hsu, H.-H., & Chao, W.-H. (2025) · International Conference on Education and E-Learning (ICEEL) · doi:10.1145/3789859.3789890
TSA-CA: Tumor-Scale Aware Curriculum Augmentation for Post-Treatment Glioma Segmentation
Omri, A., ElBedoui, K., Cirrincione, G., et al. (2026) · IEEE International Conference on Advanced Systems and Emergent Technologies (IC_ASET) · doi:10.1109/ic_aset69920.2026.11502521
“Curriculum learning [12] [13] helps models generalize better by arranging training examples from easy to hard, thereby providing stronger gradient signals early in training [14].”
Neural Network Training Inspired By Advance Organizers
Khupse, V., Sawarkar, A. D., Mapari, R., et al. (2025) · doi:10.1109/rcsm67767.2025.11506699
“Curriculum Learning (CL) trains models from “easy to hard” samples to improve convergence and generalization [8]– [10].”
Wang, X., Liu, X., Zhu, S., Li, Z., & Gao, L. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3691833 · Cited by 1
“This approach aligns with the challenges that Curriculum Learning aims to address [53].”
Thahniyath, G., Benaka Santhosha, S., Sardar, T. H., Sridhar, S. K., Dilip, K., & Naitik, S. T. (2026) · International Conference on Inventive Computation Technologies (ICICT) · doi:10.1109/icict68280.2026.11510970
“…and becomes more complex over time, we not only see better end-of-training performance, but also more stable training dynamics, fewer cases of loss spikes and gradient explosion that may arise when domain-specific jargon is presented to a general-purpose model in a single instance [15], [17].”
Receptor-Anchored Olfaction Representation through Perception-Consistent Metric LearningPreprint
Tian, C., Wang, J., Hou, J., Liu, W., Luo, Y., Wang, Y., Yang, L., & Lin, W. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.05.08.723701
“A curriculum learning strategy[33] was employed to ensure stable training and strong generalization.”
Adaptive Semantic Image Transmission With Object Prioritization and Over-the-Air Validation
Ding, J., Jiang, P., Wen, C.-K., & Jin, S. (2026) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2026.3692647 · Cited by 1
“Prior to the introduction of semantic feature remapping, the training process is divided into three stages [39], [40], which can be interpreted as a block-wise optimization strategy for a highly non-convex neural network.”
Autonomous Maneuver Decision‐Making for Unmanned Aerial Vehicles via Transfer Reinforcement Learning
Huang, Y., Long, Y., Cheng, Z., Shen, A., & Liu, M. H. (2026) · International Journal of Robust and Nonlinear Control · doi:10.1002/rnc.70586
Li, Z., Feng, Y., Cao, X., Ji, Z., Wang, Y., Liu, Y., Chen, X., Wei, R., Hu, T., Feng, J., et al. (2026) · Journal of Innovative Optical Health Sciences · doi:10.1142/s1793545826500173
CRUNet-MR-Univ: A Foundation Model for Diverse Cardiac MRI Reconstruction
Lyu, D., Staring, M., Lamb, H. J., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-17734-6_29
Wang, H., Zhang, Z., Chen, J., Fan, Z., Jiang, R., Yuan, W., Shibasaki, R., & Xuan, S. O. N. G. (2026) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2026.3685281
“The main idea of curriculum learning, first proposed by [11], is to imitate human learning and start the curriculum from easy to hard.”
Gorshenin, A. K., & Antonenko, S. V. (2026) · Pattern Recognition and Image Analysis · doi:10.1134/s1054661825701263
“In a sense, this approach is close to the so-called curriculum learning [5, 30], as it involves a transition from a simpler, low-variance dataset to a more complex and diverse one.”
Scale-Aware Curriculum Learning for Data-Efficient Lung Nodule Detection with YOLOV11
Luo, Y., Guo, Y., Hooshangnejad, H., et al. (2026) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi61048.2026.11515949
“This approach systematically organizes training samples based on their difficulty levels, allowing neural networks to first master easier examples before tackling more challenging ones [13].”
Yang, Z., & Wu, Y. (2026) · International Journal on Document Analysis and Recognition (IJDAR) · doi:10.1007/s10032-026-00597-3
Curriculum Learning on Image-Denoising Autoencoders
dos Santos, F. P., & Schiezaro, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-23176-5_10
Incremental Learning and Reward Shaping Strategies for Deep Reinforcement Learning upon CVRP
Alagözlü, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21477-5_24
Physics-Informed Machine Learning for Life Assessment of Aerospace Structures Under Fatigue
Kunal, & Chakraborty, P. (2026) · IEEE Aerospace Conference (AERO) · doi:10.1109/aero66936.2026.11520167
“In reality, curriculum learning techniques are occasionally added to the training process to gradually introduce progressively more complex residuals and boundary constraints, further empowering the model to learn stable solutions [17], [18], [19].”
MXPose: Multiplex interactive learning for multi-view 3D Human Pose Estimation
Zhang, W., Guan, Z., Liu, M., et al. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.113993
Self-evolving Embodied Intelligence
Wang, X., Feng, T., Liu, H., & Zhu, W. (2026) · Springer eBooks · doi:10.1007/978-981-95-7750-7_4 · Cited by 1
R 2 GCurL: Reinforced Robust Knowledge Tracing via Dynamic Graph Curriculum Learning
Peng, T., Pu, Y., Zhang, Y., Li, Y., Ren, J., Luo, J., Zhang, K., Yuan, H., Wang, S., Yin, D., et al. (2026) · ACM Transactions on Information Systems · doi:10.1145/3818614
“We therefore introduce CL [5] as a complementary data-centric mechanism for KT under noisy and heterogeneous training conditions.”
Jiang, L., Gao, X., Liao, J., Fan, C., & Jia, H. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6834939
A Robust Framework Based on Embedding Projection Learning for Intelligent Process Monitoring
Yang, J., Wang, Y., Huo, J., et al. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3697096
“Curriculum learning (CL) is a learning paradigm inspired by the human learning process, in which training starts with simpler instances and gradually progresses to more challenging ones by increasing the di ffi culty level [34].”
Tan, C. J., Hayashi, E., Mowshowitz, A., & Lim, W. S. (2026) · Robotics · doi:10.3390/robotics15050098
“Stage-dependent reward multipliers, inspired by the principles of curriculum learning [30], are applied to progressively shift selective pressure from basic garbage pickup toward energy management, humanoid interaction, and delivery behaviors.”
Li, C., Wang, C., Liu, B., Zhang, Y., Cai, L., & Zhao, B. (2026) · Communications in computer and information science · doi:10.1007/978-981-95-7299-1_26
Fine-Tuning Vision Language Models for Medical Visual Question Answering
Teslaru, V., Pojoga, G., & Achirei, Ș.-D. (2026) · World Congress on Medical Physics and Biomedical Engineering, September 7 - 12, 2009, Munich, Germany · doi:10.1007/978-3-032-24724-7_32
ClickIRST: Toward Interactive Segmentation for Infrared Small Target
Lu, P., Luo, Y., Jia, H., Zhang, X., Liu, Y., Shen, W., Xia, S., & Liu, Y. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3698933 · Cited by 1
“To tackle this challenge, researchers have developed curriculum learning (CL) [35], [36]—a training paradigm that mimics the human educational process by progressively introducing data from easier to more di ffi cult concepts.”
Adaptive Path Tracking Control Method for Autonomous Vehicles Based on Deep Reinforcement Learning
Xu, Y., Tang, X., Chen, Z. H., & Rao, X. (2026) · IET Intelligent Transport Systems · doi:10.1049/itr2.70223
Nguyen, P.-A., Le, T.-S., Le, D.-T., & Nguyen, C.-V. T. (2026) · Neural Computing and Applications · doi:10.1007/s00521-026-12160-6 · Cited by 1
“Inspired by the structured progression of knowledge acquisition in human cognition, Curriculum Learning (CL) [48, 49] has emerged as a training paradigm that organizes the learning process by introducing samples in an incremental order of complex-ity—from simpler to more challenging examples.”
Wang, T., Di, Y., Feng, S., Haohao, C., & Liu, Q. (2026) · Algorithms · doi:10.3390/a19060451
Jin, Y., Li, D., Li, Z., & Wu, N. (2026) · Applied Sciences · doi:10.3390/app16115426
“…widely studied in transfer learning [43]—we employ a two-stage training strategy: Curriculum Learning: The model is trained on data with gradually increasing complexity [44], starting from regions with clear accident patterns and progressing to more challenging cases with sparse or noisy labels.”
InfluentialSun, Y., Dai, J., Liu, M., Cao, Y., Wang, F., Wang, X. G., Li, P., & Su, S. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3699679 · Cited by 1
“In contrast, curriculum learning arranges tasks from easy to hard [29], yet a fixed task sequence that strictly follows an easy-to-hard order may overfit to a specific curriculum, and the simple tasks that appear early may be overly emphasized.”
Gao, P., Wang, P., Wang, F., Fujita, H., & Aljuaid, H. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-00549-1
Chen, Z., Zhu, S., Sun, S., et al. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3699687
“1) Three-Stage Curriculum Strategy: Inspired by cognitive learning principles [20], the training process is structured to expose the network to progressively more challenging geospatial distributions based on altitude.”
Jiao, C., Hu, C., Tang, X., Zhu, H., & Jiao, L. (2026) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2026.3700042
Progressive Learning-to-Learn Knowledge Distillation for Robust Out-of-Domain Generalization
Xiang, K., & Fujii, A. (2026) · International Journal of Semantic Computing · doi:10.1142/s1793351x26420031
Hierarchical multi-policy adversarial reinforcement learning
Wang, X., Wang, S., Zhang, J., & (384284), Y. M. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.115297
Li, Y., Zhou, Z., Li, H., Sun, Y., Wu, Y., Lai, Y., Wang, Y., Gong, Y., & Mei, J. (2026) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2026.3700665
“To facilitate the learning of scalable coordination strategies and mitigate the sparse reward problem, a progressive training curriculum [34] is employed.”
Muley, A., Behera, M., Mukherjee, R., et al. (2026) · doi:10.1109/aimlcps68702.2026.11542325
“Initially defined by Bengio et al. [6], Curriculum Learning indicates that models perform better when they encounter examples in a thoughtfully arranged sequence of increasing difficulty.”
Exploring neural networks and machine learning in predicting personal loan defaults
HOANG, H. T. T., Nguyen, N. M., & Ngo, D. (2026) · Risk Governance and Control Financial Markets & Institutions · doi:10.22495/rgcv16i2p11
PallasGNN: Curriculum-Based Pattern Mining for Robust GNNs
Xia, K., Wu, H., Xie, M., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-1465-5_42
LLM-guided graph neural coordination framework for cooperative multi-agent reinforcement learning
Kuang, Z., Xu, Z. S., Wan, L., & Zhang, C. (2026) · Complex & Intelligent Systems · doi:10.1007/s40747-026-02356-7
“Following the curriculum-learning perspective [62] of starting from an easier,highlyguidedregimeandthenrelaxingexternalsuper-vision, we allocate the initial 0.”
A Survey of Hierarchical Reinforcement Learning for Combinatorial Optimization in Logistics
Chevva, P. (2025) · International Conference on Data Analytics for Business and Industry (ICDABI) · doi:10.1109/icdabi67967.2025.11547569 · Cited by 1
“Curriculum learning, where agents are exposed to progressively harder problem instances, helps policies learn transferable skills and improves convergence [35].”
Van, D. C., Tam, T. D., Hung, D. M., et al. (2025) · doi:10.1109/defi67526.2025.11551638
Yang, Y. P., & Ricks, K. (2026) · Research Square · doi:10.21203/rs.3.rs-8960801/v1
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.