mlsanityData to 14 Oct 2026

Hacking Task Confounder in Meta-Learning

arXiv 2312.05771 ↗First posted 10 Dec 2023cs.LG stat.MLcited by 8

Not on the map. Why?

Abstract (arXiv)

Meta-learning enables rapid generalization to new tasks by learning knowledge from various tasks. It is intuitively assumed that as the training progresses, a model will acquire richer knowledge, leading to better generalization performance. However, our experiments reveal an unexpected result: there is negative knowledge transfer between tasks, affecting generalization performance. To explain this phenomenon, we conduct Structural Causal Models (SCMs) for causal analysis. Our investigation uncovers the presence of spurious correlations between task-specific causal factors and labels in meta-learning. Furthermore, the confounding factors differ across different batches. We refer to these confounding factors as "Task Confounders". Based on these findings, we propose a plug-and-play Meta-learning Causal Representation Learner (MetaCRL) to eliminate task confounders. It encodes decoupled generating factors from multiple tasks and utilizes an invariant-based bi-level optimization mechanism to ensure their causality for meta-learning. Extensive experiments on various benchmark datasets demonstrate that our work achieves state-of-the-art (SOTA) performance.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 34 of its 54 references that are in the map. First posted on arXiv 10 Dec 2023.

Builds on

34 of 54 references.

References · 34

  1. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  2. 2017Prototypical Networks for Few-shot Learningcited by 3,015
  3. 2016Matching Networks for One Shot Learningcited by 2,255
  4. 2019Invariant Risk Minimizationcited by 1,854
  5. 2017Learning to Compare: Relation Network for Few-Shot Learningcited by 1,143
  6. 2020Meta-Learning in Neural Networks: A Surveycited by 950
  7. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 941
  8. 2017Learning to Generalize: Meta-Learning for Domain Generalizationcited by 763
  9. 2018Meta-Learning for Semi-Supervised Few-Shot Classificationcited by 456
  10. 2017Meta-SGD: Learning to Learn Quickly for Few-Shot Learningcited by 379
  11. 2019Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAMLcited by 359
  12. 2020Causal Intervention for Weakly-Supervised Semantic Segmentationcited by 229
22 more
  1. 2021Contrastive Learning Inverts the Data Generating Processcited by 221
  2. 2020How Much Position Information Do Convolutional Neural Networks Encode?cited by 172
  3. 2019A Meta-Transfer Objective for Learning to Disentangle Causal Mechanismscited by 158
  4. 2018Task-Agnostic Meta-Learning for Few-shot Learningcited by 151
  5. 2020Interventional Few-Shot Learningcited by 112
  6. 2018Meta-Learning Probabilistic Inference For Predictioncited by 110
  7. 2018Robustly Disentangled Causal Mechanisms: Validating Deep Representations for Interventional Robustnesscited by 108
  8. 2020Learning to Learn with Variational Information Bottleneck for Domain Generalizationcited by 102
  9. 2020Offline Meta-Reinforcement Learning with Advantage Weightingcited by 93
  10. 2019Meta-Learning without Memorizationcited by 92
  11. 2020Deconfounded Image Captioning: A Causal Retrospectcited by 65
  12. 2020Meta-Learning Requires Meta-Augmentationcited by 57
  13. 2020Improving Generalization in Meta-learning via Task Augmentationcited by 55
  14. 2019The Omniglot challenge: a 3-year progress reportcited by 54
  15. 2020When is invariance useful in an Out-of-Distribution Generalization problem ?cited by 52
  16. 2021Bridging Multi-Task Learning and Meta-Learning: Towards Efficient Training and Effective Adaptationcited by 50
  17. 2022Improving Multi-Task Generalization via Regularizing Spurious Correlationcited by 28
  18. 2019Variational Metric Scaling for Metric-Based Meta-Learningcited by 17
  19. 2020Meta Learning for Causal Directioncited by 16
  20. 2022MetAug: Contrastive Learning via Meta Feature Augmentationcited by 15
  21. 2020Measuring and Harnessing Transference in Multi-Task Learningcited by 11
  22. 2022Interventional Contrastive Learning with Meta Semantic Regularizercited by 11

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 5; 2025: 2; 2026: 1.

8 papers.

  1. 2025Task Arithmetic in Trust Region: A Training-Free Model Merging Approach to Navigate Knowledge Conflictscited by 13
  2. 2024Towards the Causal Complete Cause of Multi-Modal Representation Learningcited by 5
  3. 2024Rethinking Meta-Learning from a Learning Lenscited by 1
  4. 2024DiffDesign: Controllable Diffusion with Meta Prior for Efficient Interior Design Generationcited by 1
  5. 2026Compositional Meta-Learning for Mitigating Task Heterogeneity in Physics-Informed Neural Networkscited by 1
  6. 2024Neuromodulated Meta-Learningcited by 0
  7. 2024Spatio-Temporal Fuzzy-oriented Multi-Modal Meta-Learning for Fine-grained Emotion Recognitioncited by 0
  8. 2025Causal Prompt Calibration Guided Segment Anything Model for Open-Vocabulary Multi-Entity Segmentationcited by 0