mlsanityData to 7 Oct 2026

Learning Compositional Representations for Few-Shot Recognition

arXiv 1812.09213 ↗First posted 21 Dec 2018cs.CVcited by 50

Not on the map. Why?

Abstract (arXiv)

One of the key limitations of modern deep learning approaches lies in the amount of data required to train them. Humans, by contrast, can learn to recognize novel categories from just a few examples. Instrumental to this rapid learning ability is the compositional structure of concept representations in the human brain --- something that deep learning models are lacking. In this work, we make a step towards bridging this gap between human and machine learning by introducing a simple regularization technique that allows the learned representation to be decomposable into parts. Our method uses category-level attribute annotations to disentangle the feature space of a network into subspaces corresponding to the attributes. These attributes can be either purely visual, like object parts, or more abstract, like openness and symmetry. We demonstrate the value of compositional representations on three datasets: CUB-200-2011, SUN397, and ImageNet, and show that they require fewer examples to learn classifiers for novel categories.

Key facts. Cited by 50 papers in the map (data through 7 Oct 2026). First posted on arXiv 21 Dec 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 16; 2022: 16; 2023: 11; 2024: 5; 2025: 1; 2026: 1.

50 papers.

Sort:

By area:

  1. 2022Disentangled Representation Learningcited by 111
  2. 2021Beyond Max-Margin: Class Margin Equilibrium for Few-shot Object Detectioncited by 80
  3. 2021Detecting Human-Object Interaction via Fabricated Compositional Learningcited by 79
  4. 2022Compositional Generalization in Unsupervised Compositional Representation Learning: A Study on Disentanglement and Emergent Languagecited by 42
  5. 2022Generating Representative Samples for Few-Shot Classificationcited by 39
  6. 2022Do Vision-Language Pretrained Models Learn Composable Primitive Concepts?cited by 31
  7. 2022BatchFormer: Learning to Explore Sample Relationships for Robust Representation Learningcited by 29
  8. 2022Probabilistic Compositional Embeddings for Multimodal Image Retrievalcited by 29
  9. 2023MMG-Ego4D: Multi-Modal Generalization in Egocentric Action Recognitioncited by 26
  10. 2023Prompting Language-Informed Distribution for Compositional Zero-Shot Learningcited by 24
  11. 2021Learning Compositional Representation for 4D Captures with Neural ODEcited by 22
  12. 2023Use Your Head: Improving Long-Tail Video Recognitioncited by 22
38 more
  1. 2022Attribute Prototype Network for Any-Shot Learningcited by 19
  2. 2024When does compositional structure yield compositional generalization? A kernel theorycited by 19
  3. 2022Contributions of Shape, Texture, and Color in Visual Recognitioncited by 16
  4. 2022H4D: Human 4D Modeling by Learning Neural Compositional Representationcited by 15
  5. 2023Simple Semantic-Aided Few-Shot Learningcited by 15
  6. 2024A Survey on Compositional Learning of AI Models: Theoretical and Experimental Practicescited by 15
  7. 2021Disentangled Feature Representation for Few-shot Image Classificationcited by 12
  8. 2021Learnable Expansion-and-Compression Network for Few-shot Class-Incremental Learningcited by 11
  9. 2022vMFNet: Compositionality Meets Domain-generalised Segmentationcited by 11
  10. 2021MFNet: Multi-class Few-shot Segmentation Network with Pixel-wise Metric Learningcited by 10
  11. 2022Hierarchical Compositional Representations for Few-shot Action Recognitioncited by 10
  12. 2021Meta Learning on a Sequence of Imbalanced Domains with Difficulty Awarenesscited by 8
  13. 2021Learning Single/Multi-Attribute of Object with Symmetry and Groupcited by 8
  14. 2021Prototype Completion for Few-Shot Learningcited by 7
  15. 2021Information Symmetry Matters: A Modal-Alternating Propagation Network for Few-Shot Learningcited by 6
  16. 2023Mitigating the Effect of Incidental Correlations on Part-based Learningcited by 6
  17. 2021Contextualizing Meta-Learning via Learning to Decomposecited by 5
  18. 2021Learning Compositional Representation for Few-shot Visual Question Answeringcited by 3
  19. 2021Generating Compositional Color Representations from Textcited by 3
  20. 2022Reference-Limited Compositional Zero-Shot Learningcited by 3
  21. 2021Shaping Visual Representations with Attributes for Few-Shot Recognitioncited by 2
  22. 2023ENInst: Enhancing Weakly-supervised Low-shot Instance Segmentationcited by 2
  23. 2023Compositionally Equivariant Representation Learningcited by 2
  24. 2021Image interpretation by iterative bottom-up top-down processingcited by 1
  25. 2021Compositional Fine-Grained Low-Shot Learningcited by 1
  26. 2022A Unified Framework with Meta-dropout for Few-shot Learningcited by 1
  27. 2023Compositional Prototypical Networks for Few-Shot Classificationcited by 1
  28. 2023Logarithm-transform aided Gaussian Sampling for Few-Shot Learningcited by 1
  29. 2025Data-Agnostic Cardinality Learning from Imperfect Workloadscited by 1
  30. 2022Fine-grained Few-shot Recognition by Deep Object Parsingcited by 0
  31. 2022Learning Primitive-aware Discriminative Representations for Few-shot Learningcited by 0
  32. 2022Interpretable Few-shot Learning with Online Attribute Selectioncited by 0
  33. 2023Concept Discovery for Fast Adapatationcited by 0
  34. 2023Compositional Representation Learning for Brain Tumour Segmentationcited by 0
  35. 2024Enhanced Data Transfer Cooperating with Artificial Triplets for Scene Graph Generationcited by 0
  36. 2024Enhancing Cross-Modal Medical Image Segmentation through Compositionalitycited by 0
  37. 2024A Feature Generator for Few-Shot Learningcited by 0
  38. 2026Decomposed Vision-Language Alignment for Fine-Grained Open-Vocabulary Segmentationcited by 0