Curriculum Learning as Transport: Understanding Curricula with Wasserstein Geodesics

Abstract

Curriculum learning is governed by several coupled design choices---how difficulty is defined, how examples are ordered, how much exposure each level receives, and how quickly training moves across levels---making it hard to isolate what actually helps. We present Wasserstein curriculum paths, a simple transport-based framework that decouples these factors by representing curricula as trajectories of training distributions over discrete difficulty levels. Across a calibrated synthetic suite with 12 tasks and 33 difficulty axes, we use this framework to isolate the effects of ordering, matched exposure, endpoint smoothness, and pacing under fixed training budgets. We find that curriculum effects are strongly context-dependent: no single strategy dominates across tasks, difficulty axes, and budgets, and curricula mainly change where a fixed budget is spent most effectively. Within this framework, easy-to-hard ordering improves hard-level performance relative to exposure-matched static sampling, showing that the benefit is not explained by cumulative exposure alone. We further show that endpoint smoothness and pacing substantially affect where along the difficulty spectrum a curriculum is effective. Finally, we show that the same transport view naturally supports extensions to learned pacing through geometry and to structured difficulty spaces beyond one-dimensional orderings.

The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.

More in Machine Learning

• arXiv preprint · 8 September 2026

  Risk-Conditioned Fine-Tuning of Large Language Models

  Zixuan Liu, Fangzheng Wu, Brian Summa and 1 other
• arXiv preprint · 9 September 2026

  A Statistical Approach to Estimating Sample Size of Machine Learning Models

  Dat Phan-Trong, Sunil Gupta and Svetha Venkatesh
• arXiv preprint · 8 September 2026

  A Gradient-based yet Spike-Timing-Dependent Solution to the Feedback Learning Problem in Neural Microcircuits

  Xiangnan Zhang, Jingxin Liu, Ranqi Lu and 6 others
• arXiv preprint · 9 September 2026

  TEFM: Token-Efficient Faithful Modeling for Structured Data

  Zhichao Hou, Lingdao Sha, Xueyu Mao and 3 others
• arXiv preprint · 8 September 2026

  A Machine Learning Framework for Predicting Restaurant Food Waste to Support Sustainable Food Management

  Md Mehedi Hasan Naeem, Md Ashraful Islam, Moumita Barua and 2 others

All new papers in Machine Learning