mlsanityData to 7 Oct 2026

Uniform Sampling over Episode Difficulty

arXiv 2108.01662 ↗First posted 3 Aug 2021cs.LG cs.AI cs.CVcited by 6

Not on the map. Why?

Abstract (arXiv)

Episodic training is a core ingredient of few-shot learning to train models on tasks with limited labelled data. Despite its success, episodic training remains largely understudied, prompting us to ask the question: what is the best way to sample episodes? In this paper, we first propose a method to approximate episode sampling distributions based on their difficulty. Building on this method, we perform an extensive analysis and find that sampling uniformly over episode difficulty outperforms other sampling schemes, including curriculum and easy-/hard-mining. As the proposed sampling method is algorithm agnostic, we can leverage these insights to improve few-shot learning accuracies across many episodic training algorithms. We demonstrate the efficacy of our method across popular few-shot learning datasets, algorithms, network architectures, and protocols.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 44 of its 74 references that are in the map. First posted on arXiv 3 Aug 2021.

Builds on

44 of 74 references.

References · 44

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  4. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  5. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  6. 2016Matching Networks for One Shot Learningcited by 2,246
  7. 2013Describing Textures in the Wildcited by 2,110
  8. 2013Fine-Grained Visual Classification of Aircraftcited by 1,701
  9. 2016Optimization Methods for Large-Scale Machine Learningcited by 1,140
  10. 2015Prioritized Experience Replaycited by 1,070
  11. 2018On First-Order Meta-Learning Algorithmscited by 1,003
  12. 2020Meta-Learning in Neural Networks: A Surveycited by 944
32 more
  1. 2016Training Region-based Object Detectors with Online Hard Example Miningcited by 577
  2. 2019Meta-Learning with Implicit Gradientscited by 523
  3. 2019Generalizing from a Few Examples: A Survey on Few-Shot Learningcited by 462
  4. 2019Meta-Learning with Differentiable Convex Optimizationcited by 461
  5. 2018Meta-Learning for Semi-Supervised Few-Shot Classificationcited by 456
  6. 2018Conditional Neural Processescited by 421
  7. 2018TADAM: Task dependent adaptive metric for improved few-shot learningcited by 408
  8. 2018Meta-learning with differentiable closed-form solverscited by 395
  9. 2018Dynamic Few-Shot Visual Learning without Forgettingcited by 385
  10. 2019Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAMLcited by 359
  11. 2018Not All Samples Are Created Equal: Deep Learning with Importance Samplingcited by 323
  12. 2019Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examplescited by 312
  13. 2019A Baseline for Few-Shot Image Classificationcited by 298
  14. 2018Few-Shot Learning via Embedding Adaptation with Set-to-Set Functionscited by 293
  15. 2017Sampling Matters in Deep Embedding Learningcited by 285
  16. 2018DropBlock: A regularization method for convolutional networkscited by 261
  17. 2018Bayesian Model-Agnostic Meta-Learningcited by 194
  18. 2018Recasting Gradient-Based Meta-Learning as Hierarchical Bayescited by 180
  19. 2020When Do Curricula Work?cited by 105
  20. 2019Meta-Learning with Warped Gradient Descentcited by 98
  21. 2019Meta-Curvaturecited by 70
  22. 2020learn2learn: A Library for Meta-Learning Researchcited by 69
  23. 2019Few-Shot Learning with Embedded Class Models and Shot-Free Meta Trainingcited by 57
  24. 2020On Episodes, Prototypical Networks, and Few-shot Learningcited by 46
  25. 2016Importance Sampling for Minibatchescited by 36
  26. 2021Generalization of Model-Agnostic Meta-Learning Algorithms: Recurring and Unseen Taskscited by 36
  27. 2020Adaptive Task Sampling for Meta-Learningcited by 29
  28. 2019Meta-Transfer Learning through Hard Taskscited by 23
  29. 2014Accelerating Minibatch Stochastic Gradient Descent using Stratified Samplingcited by 20
  30. 2018Active Mini-Batch Sampling using Repulsive Point Processescited by 19
  31. 2020Revisiting Meta-Learning as Supervised Learningcited by 14
  32. 2020Expert Training: Task Hardness Aware Meta-Learning for Few-Shot Classificationcited by 6

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 1; 2023: 0; 2024: 4.

6 papers.

  1. 2022Continuous-Time Meta-Learning with Forward Mode Differentiationcited by 12
  2. 2024Data-Efficient and Robust Task Selection for Meta-Learningcited by 2
  3. 2024Task Attribute Distance for Few-Shot Learning: Theoretical Analysis and Applicationscited by 1
  4. 2021On Hard Episodes in Meta-Learningcited by 0
  5. 2024Meta Episodic learning with Dynamic Task Sampling for CLIP-based Point Cloud Classificationcited by 0
  6. 2024Oops, I Sampled it Again: Reinterpreting Confidence Intervals in Few-Shot Learningcited by 0