mlsanityData to 14 Oct 2026

On Missing Labels, Long-tails and Propensities in Extreme Multi-label Classification

arXiv 2207.13186 ↗First posted 26 Jul 2022cs.LG cs.IRcited by 15

Not on the map. Why?

Abstract (arXiv)

The propensity model introduced by Jain et al. 2016 has become a standard approach for dealing with missing and long-tail labels in extreme multi-label classification (XMLC). In this paper, we critically revise this approach showing that despite its theoretical soundness, its application in contemporary XMLC works is debatable. We exhaustively discuss the flaws of the propensity-based approach, and present several recipes, some of them related to solutions used in search engines and recommender systems, that we believe constitute promising alternatives to be followed in XMLC.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Builds on 15 of its 46 references that are in the map. First posted on arXiv 26 Jul 2022.

Builds on

15 of 46 references.

References · 15

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2019Class-Balanced Loss Based on Effective Number of Samplescited by 1,245
  4. 2018Learning from positive and unlabeled data: a surveycited by 219
  5. 2016Unbiased Learning-to-Rank with Biased Feedbackcited by 157
  6. 2018AttentionXML: Label Tree-based Attention-Aware Deep Model for High-Performance Extreme Multi-Label Text Classificationcited by 102
  7. 2015Inferring Networks of Substitutable and Complementary Productscited by 92
  8. 2019Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedbackcited by 79
  9. 2016DiSMEC - Distributed Sparse Machines for Extreme Multi-label Classificationcited by 71
  10. 2019Bonsai -- Diverse and Shallow Trees for Extreme Multi-label Classificationcited by 61
  11. 2019Extreme Classification in Log Memory using Count-Min Sketch: A Case Study of Amazon Search with 50M Productscited by 36
  12. 2020Learning Optimal Tree Models Under Beam Searchcited by 26
3 more
  1. 2014Conditional Probability Tree Estimation Analysis and Algorithmscited by 10
  2. 2021Propensity-scored Probabilistic Label Treescited by 8
  3. 2021Unbiased Loss Functions for Multilabel Classification with Missing Labelscited by 5

Built on by

Cited by year

Citing papers per year: 2022: 2; 2023: 4; 2024: 9.

15 papers.

Sort:

  1. 2022CascadeXML: Rethinking Transformers for End-to-end Multi-resolution Training in Extreme Multi-label Classificationcited by 12
  2. 2022A Survey on Extreme Multi-label Learningcited by 7
  3. 2023Towards Memory-Efficient Training for Extremely Large Output Spaces -- Learning with 500k Labels on a Single Commodity GPUcited by 6
  4. 2023Generalized test utilities for long-tail performance in extreme multi-label classificationcited by 6
  5. 2024Consistent algorithms for multi-label classification with macro-at-k metricscited by 4
  6. 2023MatchXML: An Efficient Text-label Matching Framework for Extreme Multi-label Text Classificationcited by 3
  7. 2024Learning label-label correlations in Extreme Multi-label Classification via Label Featurescited by 3
  8. 2024Navigating Extremes: Dynamic Sparsity in Large Output Spacescited by 3
  9. 2024Labels in Extremes: How Well Calibrated are Extreme Multi-label Classifiers?cited by 2
  10. 2023Single-sample versus case-control sampling scheme for Positive Unlabeled data: the story of two scenarioscited by 1
  11. 2024UniDEC : Unified Dual Encoder and Classifier Training for Extreme Multi-Label Classificationcited by 1
  12. 2024Determined Multi-Label Learning via Similarity-Based Promptcited by 0
3 more
  1. 2024A Survey on Incomplete Multi-label Learning: Recent Advances and Future Trendscited by 0
  2. 2024Augmented prediction of a true class for Positive Unlabeled data under selection biascited by 0
  3. 2024On the Necessity of World Knowledge for Mitigating Missing Labels in Extreme Classificationcited by 0