mlsanityData to 7 Oct 2026

When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysis

arXiv 2308.05017 ↗First posted 9 Aug 2023cs.LGcited by 21

Not on the map. Why?

Abstract (arXiv)

Novel Class Discovery (NCD) aims at inferring novel classes in an unlabeled set by leveraging prior knowledge from a labeled set with known classes. Despite its importance, there is a lack of theoretical foundations for NCD. This paper bridges the gap by providing an analytical framework to formalize and investigate when and how known classes can help discover novel classes. Tailored to the NCD problem, we introduce a graph-theoretic representation that can be learned by a novel NCD Spectral Contrastive Loss (NSCL). Minimizing this objective is equivalent to factorizing the graph's adjacency matrix, which allows us to derive a provable error bound and provide the sufficient and necessary condition for NCD. Empirically, NSCL can match or outperform several strong baselines on common benchmark datasets, which is appealing for practical usage while enjoying theoretical guarantees.

Key facts. Cited by 21 papers in the map (data through 7 Oct 2026). Builds on 32 of its 52 references that are in the map. First posted on arXiv 9 Aug 2023.

Builds on

32 of 52 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  4. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  5. 2016CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoningcited by 1,445
  6. 2022Out-of-Distribution Detection with Deep Nearest Neighborscited by 596
  7. 2019A Theoretical Analysis of Contrastive Unsupervised Representation Learningcited by 530
  8. 2021ReAct: Out-of-distribution Detection With Rectified Activationscited by 486
  9. 2021Provable Guarantees for Self-Supervised Deep Learning with Spectral Contrastive Losscited by 311
  10. 2019Learning to Discover Novel Visual Categories via Deep Transfer Clusteringcited by 223
  11. 2022Delving into Out-of-Distribution Detection with Vision-Language Representationscited by 215
  12. 2021A Unified Objective for Novel Class Discoverycited by 161
20 more
  1. 2020Predicting What You Already Know Helps: Provable Self-Supervised Learningcited by 152
  2. 2020Contrastive learning, multi-view redundancy, and linear modelscited by 146
  3. 2022Contrastive and Non-Contrastive Self-Supervised Learning Recover Global and Local Spectral Embedding Methodscited by 146
  4. 2020Automatically Discovering and Learning New Visual Categories with Ranking Statisticscited by 141
  5. 2017Learning to cluster in order to transfer across domains and taskscited by 128
  6. 2022How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?cited by 128
  7. 2021Neighborhood Contrastive Learning for Novel Class Discoverycited by 119
  8. 2021Novel Visual Category Discovery with Dual Ranking Statistics and Mutual Knowledge Distillationcited by 104
  9. 2019Multi-class Classification without Multi-class Labelscited by 100
  10. 2018SpectralNet: Spectral Clustering using Deep Neural Networkscited by 94
  11. 2020OpenMix: Reviving Known Knowledge for Discovering Novel Visual Categories in An Open Worldcited by 93
  12. 2021Supporting Clustering with Contrastive Learningcited by 90
  13. 2022Connect, Not Collapse: Explaining Contrastive Learning for Unsupervised Domain Adaptationcited by 66
  14. 2021Joint Representation Learning and Novel Category Discovery on Single- and Multi-modal Datacited by 55
  15. 2020Contrastive estimation reveals topic posterior information to linear modelscited by 54
  16. 2022Beyond Separability: Analyzing the Linear Transferability of Contrastive Representations to Related Subpopulationscited by 44
  17. 2022OpenCon: Open-world Contrastive Learningcited by 33
  18. 2021Meta Discovery: Learning to Discover Novel Classes given Very Limited Datacited by 30
  19. 2023The Trade-off between Universality and Label Efficiency of Representations from Contrastive Learningcited by 27
  20. 2022A Closer Look at Novel Class Discovery from the Labeled Setcited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 5; 2024: 11; 2025: 4; 2026: 1.

21 papers.

Sort:

By area:

  1. 2024ARGS: Alignment as Reward-Guided Searchcited by 99
  2. 2024Process Reward Model with Q-Value Rankingscited by 78
  3. 2023Learn to Categorize or Categorize to Learn? Self-Coding for Generalized Category Discoverycited by 39
  4. 2024Do Large Language Models Have Compositional Ability? An Investigation into Limitations and Scalabilitycited by 39
  5. 2023Diversified Outlier Exposure for Out-of-Distribution Detection via Informative Extrapolationcited by 38
  6. 2024Towards Few-Shot Adaptation of Foundation Models via Multitask Finetuningcited by 31
  7. 2024PointAD: Comprehending 3D Anomalies from Points and Pixels for Zero-shot 3D Anomaly Detectioncited by 29
  8. 2023A Graph-Theoretic Framework for Understanding Open-World Semi-Supervised Learningcited by 22
  9. 2024Differential Privacy Mechanisms in Neural Tangent Kernel Regressioncited by 14
  10. 2024On the Learnability of Out-of-distribution Detectioncited by 8
  11. 2024When and How Does In-Distribution Label Help Out-of-Distribution Detection?cited by 6
  12. 2024DFA-RAG: Conversational Semantic Router for Large Language Model with Definite Finite Automatoncited by 5
9 more
  1. 2025Category Discovery: An Open-World Perspectivecited by 5
  2. 2025Open-World Deepfake Attribution via Confidence-Aware Asymmetric Learningcited by 5
  3. 2024Labeled Data Selection for Category Discoverycited by 4
  4. 2024Bridging OOD Detection and Generalization: A Graph-Theoretic Viewcited by 3
  5. 2023A Practical Approach to Novel Class Discovery in Tabular Datacited by 2
  6. 2025MCL-AD: Multimodal Collaboration Learning for Zero-Shot 3D Anomaly Detectioncited by 2
  7. 2023Detecting and Learning Out-of-Distribution Data in the Open world: Algorithm and Theorycited by 1
  8. 2025PointAD+: Learning Hierarchical Representations for Zero-shot 3D Anomaly Detectioncited by 1
  9. 2026The Devil Is in Gradient Entanglement: Energy-Aware Gradient Coordinator for Robust Generalized Category Discoverycited by 0