mlsanityData to 7 Oct 2026

A Universal Representation Transformer Layer for Few-Shot Image Classification

arXiv 2006.11702 ↗First posted 21 Jun 2020cs.LG cs.CV stat.MLcited by 65

Not on the map. Why?

Abstract (arXiv)

Few-shot classification aims to recognize unseen classes when presented with only a small number of samples. We consider the problem of multi-domain few-shot image classification, where unseen classes and examples come from diverse data sources. This problem has seen growing interest and has inspired the development of benchmarks such as Meta-Dataset. A key challenge in this multi-domain setting is to effectively integrate the feature representations from the diverse set of training domains. Here, we propose a Universal Representation Transformer (URT) layer, that meta-learns to leverage universal features for few-shot classification by dynamically re-weighting and composing the most appropriate domain-specific representations. In experiments, we show that URT sets a new state-of-the-art result on Meta-Dataset. Specifically, it achieves top-performance on the highest number of data sources compared to competing methods. We analyze variants of URT and present a visualization of the attention score heatmaps that sheds light on how the model performs cross-domain generalization. Our code is available at https://github.com/liulu112601/URT.

Key facts. Cited by 65 papers in the map (data through 7 Oct 2026). First posted on arXiv 21 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 17; 2022: 23; 2023: 10; 2024: 11; 2025: 4.

65 papers.

Sort:

By area:

  1. 2021Transformers in Vision: A Surveycited by 1,009
  2. 2021FedProto: Federated Prototype Learning across Heterogeneous Clientscited by 349
  3. 2022A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunitiescited by 162
  4. 2022Federated Learning from Pre-Trained Models: A Contrastive Learning Approachcited by 141
  5. 2022Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Differencecited by 123
  6. 2022Head2Toe: Utilizing Intermediate Representations for Better Transfer Learningcited by 83
  7. 2021Cross-domain Few-shot Learning with Task-specific Adapterscited by 79
  8. 2022PSTNet: Point Spatio-Temporal Convolution on Point Cloud Sequencescited by 68
  9. 2023Advances and Challenges in Meta-Learning: A Technical Reviewcited by 65
  10. 2021Learning a Universal Template for Few-shot Dataset Generalizationcited by 61
  11. 2021Anomaly Detection in Dynamic Graphs via Transformercited by 57
  12. 2021Universal Representation Learning from Multiple Domains for Few-shot Classificationcited by 56
38 more
  1. 2022HyperTransformer: Model Generation for Supervised and Semi-Supervised Few-Shot Learningcited by 55
  2. 2021Shallow Bayesian Meta Learning for Real-World Few-Shot Recognitioncited by 48
  3. 2022Point Primitive Transformer for Long-Term 4D Point Cloud Video Understandingcited by 35
  4. 2023Meta-Album: Multi-domain Meta-Dataset for Few-Shot Image Classificationcited by 29
  5. 2021Ranking Distance Calibration for Cross-Domain Few-Shot Learningcited by 27
  6. 2021Impact of Aliasing on Generalization in Deep Convolutional Networkscited by 25
  7. 2022Universal Representations: A Unified Look at Multiple Task and Domain Learningcited by 23
  8. 2022Self-Promoted Supervision for Few-Shot Transformercited by 22
  9. 2021TOHAN: A One-step Approach towards Few-shot Hypothesis Adaptationcited by 17
  10. 2021Improving Zero-shot Generalization in Offline Reinforcement Learning using Generalized Similarity Functionscited by 17
  11. 2021Which images to label for few-shot medical landmark detection?cited by 16
  12. 2022Few-Shot Learning Meets Transformer: Unified Query-Support Transformers for Few-Shot Classificationcited by 16
  13. 2025Reliable Few-shot Learning under Dual Noisescited by 16
  14. 2022Tree Structure-Aware Few-Shot Image Classification via Hierarchical Aggregationcited by 14
  15. 2023DETA: Denoised Task Adaptation for Few-Shot Learningcited by 13
  16. 2021A State-of-the-art Survey of Object Detection Techniques in Microorganism Image Analysis: From Classical Methods to Deep Learning Approachescited by 12
  17. 2022Mask-guided Vision Transformer (MG-ViT) for Few-Shot Learningcited by 10
  18. 2022CAD: Co-Adapting Discriminative Features for Improved Few-Shot Classificationcited by 9
  19. 2021On the Importance of Firth Bias Reduction in Few-Shot Classificationcited by 7
  20. 2024Cross-Domain Few-Shot Learning via Adaptive Transformer Networkscited by 7
  21. 2024Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learningcited by 7
  22. 2021Supervised Momentum Contrastive Learning for Few-Shot Classificationcited by 6
  23. 2024Chameleon: A Data-Efficient Generalist for Dense Visual Prediction in the Wildcited by 6
  24. 2025OSLoPrompt: Bridging Low-Supervision Challenges and Open-Set Domain Generalization in CLIPcited by 6
  25. 2021Omni-Training: Bridging Pre-Training and Meta-Training for Few-Shot Learningcited by 5
  26. 2021MDFM: Multi-Decision Fusing Model for Few-Shot Learningcited by 5
  27. 2022GCT: Graph Co-Training for Semi-Supervised Few-Shot Learningcited by 5
  28. 2022Transformers Meet Visual Learning Understanding: A Comprehensive Reviewcited by 5
  29. 2024FedCCL: Federated Dual-Clustered Feature Contrast Under Domain Heterogeneitycited by 5
  30. 2022Beyond Simple Meta-Learning: Multi-Purpose Models for Multi-Domain, Active and Continual Few-Shot Learningcited by 4
  31. 2022Category-Agnostic 6D Pose Estimation with Conditional Neural Processescited by 4
  32. 2022CEIP: Combining Explicit and Implicit Priors for Reinforcement Learning with Demonstrationscited by 4
  33. 2022Enhancing Few-shot Image Classification with Cosine Transformercited by 4
  34. 2023DAC-MR: Data Augmentation Consistency Based Meta-Regularization for Meta-Learningcited by 4
  35. 2023Neural Fine-Tuning Search for Few-Shot Learningcited by 4
  36. 2023Diversity-enhancing Generative Network for Few-shot Hypothesis Adaptationcited by 4
  37. 2024MOKD: Cross-domain Finetuning for Few-shot Classification via Maximizing Optimized Kernel Dependencecited by 4
  38. 2022NeurIPS'22 Cross-Domain MetaDL competition: Design and baseline resultscited by 3

50 most cited of 65 All 65 (JSON)