mlsanityData to 7 Oct 2026

MetaNODE: Prototype Optimization as a Neural ODE for Few-Shot Learning

arXiv 2103.14341 ↗First posted 26 Mar 2021cs.CVcited by 11

Not on the map. Why?

Abstract (arXiv)

Few-Shot Learning (FSL) is a challenging task, i.e., how to recognize novel classes with few examples? Pre-training based methods effectively tackle the problem by pre-training a feature extractor and then predicting novel classes via a cosine nearest neighbor classifier with mean-based prototypes. Nevertheless, due to the data scarcity, the mean-based prototypes are usually biased. In this paper, we attempt to diminish the prototype bias by regarding it as a prototype optimization problem. To this end, we propose a novel meta-learning based prototype optimization framework to rectify prototypes, i.e., introducing a meta-optimizer to optimize prototypes. Although the existing meta-optimizers can also be adapted to our framework, they all overlook a crucial gradient bias issue, i.e., the mean-based gradient estimation is also biased on sparse data. To address the issue, we regard the gradient and its flow as meta-knowledge and then propose a novel Neural Ordinary Differential Equation (ODE)-based meta-optimizer to polish prototypes, called MetaNODE. In this meta-optimizer, we first view the mean-based prototypes as initial prototypes, and then model the process of prototype optimization as continuous-time dynamics specified by a Neural ODE. A gradient flow inference network is carefully designed to learn to estimate the continuous gradient flow for prototype dynamics. Finally, the optimal prototypes can be obtained by solving the Neural ODE. Extensive experiments on miniImagenet, tieredImagenet, and CUB-200-2011 show the effectiveness of our method.

Key facts. Cited by 11 papers in the map (data through 7 Oct 2026). Builds on 40 of its 55 references that are in the map. First posted on arXiv 26 Mar 2021.

Builds on

40 of 55 references.

References · 40

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  4. 2018Neural Ordinary Differential Equationscited by 3,733
  5. 2017Prototypical Networks for Few-shot Learningcited by 3,001
  6. 2016Matching Networks for One Shot Learningcited by 2,246
  7. 2015Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)cited by 1,061
  8. 2019A Closer Look at Few-shot Classificationcited by 679
  9. 2019Meta-Learning with Implicit Gradientscited by 523
  10. 2018Meta-Learning with Latent Embedding Optimizationcited by 476
  11. 2020Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need?cited by 435
  12. 2019Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAMLcited by 359
28 more
  1. 2019A Baseline for Few-Shot Image Classificationcited by 298
  2. 2019Cross Attention Network for Few-shot Classificationcited by 218
  3. 2018Learning to Propagate Labels: Transductive Propagation Network for Few-shot Learningcited by 208
  4. 2020CrossTransformers: spatially-aware few-shot transfercited by 174
  5. 2021Free Lunch for Few-shot Learning: Distribution Calibrationcited by 165
  6. 2018Task-Agnostic Meta-Learning for Few-shot Learningcited by 151
  7. 2020Negative Margin Matters: Understanding Margin in Few-shot Classificationcited by 129
  8. 2019Finding Task-Relevant Features for Few-Shot Learning by Category Traversalcited by 101
  9. 2019Meta-Learning with Warped Gradient Descentcited by 98
  10. 2019Prototype Rectification for Few-Shot Learningcited by 96
  11. 2020Laplacian Regularized Few-Shot Learningcited by 90
  12. 2020Embedding Propagation: Smoother Manifold for Few-Shot Classificationcited by 79
  13. 2020DPGN: Distribution Propagation Graph Network for Few-shot Learningcited by 79
  14. 2020Meta-Learning with Adaptive Hyperparameterscited by 79
  15. 2021Partial Is Better Than All: Revisiting Fine-tuning Strategy for Few-shot Learningcited by 72
  16. 2019Transductive Episodic-Wise Adaptive Metric for Few-Shot Learningcited by 67
  17. 2019Associative Alignment for Few-shot Image Classificationcited by 58
  18. 2020Instance Credibility Inference for Few-Shot Learningcited by 57
  19. 2020Prototype Completion with Primitive Knowledge for Few-Shot Learningcited by 50
  20. 2020Empirical Bayes Transductive Meta-Learning with Synthetic Gradientscited by 48
  21. 2019Few-Shot Learning with Global Class Representationscited by 44
  22. 2020Transductive Information Maximization For Few-Shot Learningcited by 44
  23. 2020One-Shot Image Classification by Learning to Restore Prototypescited by 14
  24. 2020MRI Image Reconstruction via Learning Optimization Using Neural ODEscited by 14
  25. 2020A Dynamical View on Optimization Algorithms of Overparameterized Neural Networkscited by 12
  26. 2020Cold-start Sequential Recommendation via Meta Learnercited by 11
  27. 2020Uncertainty-Aware Few-Shot Image Classificationcited by 9
  28. 2020Implicit Euler ODE Networks for Single-Image Dehazingcited by 6

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 4; 2023: 3; 2024: 1; 2025: 2.

11 papers.

  1. 2022A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunitiescited by 162
  2. 2023MetaDiff: Meta-Learning with Conditional Diffusion for Few-Shot Learningcited by 40
  3. 2022Continuous-Time Meta-Learning with Forward Mode Differentiationcited by 12
  4. 2021Prototype Completion for Few-Shot Learningcited by 7
  5. 2023Bayesian Gaussian Process ODEs via Double Normalizing Flowscited by 7
  6. 2022Robust Prototypical Few-Shot Organ Segmentation with Regularized Neural-ODEscited by 5
  7. 2022MetaDT: Meta Decision Tree with Class Hierarchy for Interpretable Few-Shot Learningcited by 2
  8. 2023Compositional Prototypical Networks for Few-Shot Classificationcited by 1
  9. 2024Learning and Current Prediction of PMSM Drive via Differential Neural Networkscited by 1
  10. 2025HyperFlow: Gradient-Free Emulation of Few-Shot Fine-Tuningcited by 1
  11. 2025Unbiased Max-Min Embedding Classification for Transductive Few-Shot Learning: Clustering and Classification Are All You Needcited by 0