mlsanityData to 7 Oct 2026

End-to-End Weak Supervision

arXiv 2107.02233 ↗First posted 5 Jul 2021cs.LG cs.AI stat.MLcited by 33

Not on the map. Why?

Abstract (arXiv)

Aggregating multiple sources of weak supervision (WS) can ease the data-labeling bottleneck prevalent in many machine learning applications, by replacing the tedious manual collection of ground truth labels. Current state of the art approaches that do not use any labeled training data, however, require two separate modeling steps: Learning a probabilistic latent variable model based on the WS sources -- making assumptions that rarely hold in practice -- followed by downstream model training. Importantly, the first step of modeling does not consider the performance of the downstream model. To address these caveats we propose an end-to-end approach for directly learning the downstream model by maximizing its agreement with probabilistic labels generated by reparameterizing previous probabilistic posteriors with a neural network. Our results show improved performance over prior work in terms of end model performance on downstream test sets, as well as in terms of improved robustness to dependencies among weak supervision sources.

Key facts. Cited by 33 papers in the map (data through 7 Oct 2026). Builds on 27 of its 42 references that are in the map. First posted on arXiv 5 Jul 2021.

Builds on

27 of 42 references.

References · 27

  1. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  2. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  3. 2016Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filteringcited by 517
  4. 2017Snorkel: Rapid Training Data Creation with Weak Supervisioncited by 370
  5. 2019Bias in Bios: A Case Study of Semantic Representation Bias in a High-Stakes Settingcited by 346
  6. 2021Understanding self-supervised Learning Dynamics without Contrastive Pairscited by 254
  7. 2016Data Programming: Creating Large Training Sets, Quicklycited by 246
  8. 2018Training Complex Models with Multi-Task Weak Supervisioncited by 106
  9. 2017Deep learning from crowdscited by 105
  10. 2020Fast and Three-rious: Speeding Up Weak Supervision with Triplet Methodscited by 98
  11. 2017Who Said What: Modeling Individual Labelers Improves Classificationcited by 77
  12. 2014Spectral Methods meet EM: A Provably Optimal Algorithm for Crowdsourcingcited by 63
15 more
  1. 2020Learning from Rules Generalizing Labeled Exemplarscited by 59
  2. 2018Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scalecited by 52
  3. 2020Interactive Weak Supervision: Learning Useful Heuristics for Data Labelingcited by 52
  4. 2017Learning the Structure of Generative Models without Labeled Datacited by 50
  5. 2017Learning From Noisy Singly-labeled Datacited by 50
  6. 2021Self-Training with Weak Supervisioncited by 49
  7. 2019Learning Dependency Structures for Weak Supervision Modelscited by 32
  8. 2019Cross-Modal Data Programming Enables Rapid Medical Machine Learningcited by 32
  9. 2019Max-MIG: an Information Theoretic Approach for Joint Learning from Crowdscited by 26
  10. 2017Inferring Generative Model Structure with Static Analysiscited by 18
  11. 2017Using Rule-Based Labels for Weak Supervised Learning: A ChemNet for Transferable Chemical Property Predictioncited by 12
  12. 2018Deep Probabilistic Logic: A Unifying Framework for Indirect Supervisioncited by 9
  13. 2019Data Programming using Continuous and Quality-Guided Labeling Functionscited by 9
  14. 2021Comparing the Value of Labeled and Unlabeled Data in Method-of-Moments Latent Variable Estimationcited by 7
  15. 2021Dependency Structure Misspecification in Multi-Source Weak Supervision Modelscited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 15; 2023: 9; 2024: 5; 2025: 3; 2026: 1.

33 papers.

Sort:

  1. 2022A Survey on Programmatic Weak Supervisioncited by 66
  2. 2024Theoretical Analysis of Weak-to-Strong Generalizationcited by 46
  3. 2022Shoring Up the Foundations: Fusing Model Embeddings and Weak Supervisioncited by 18
  4. 2022Losses over Labels: Weakly Supervised Learning via Direct Loss Constructioncited by 11
  5. 2023Transferring Annotator- and Instance-dependent Transition Matrix for Learning from Crowdscited by 11
  6. 2022Nemo: Guiding and Contextualizing Weak Supervision for Interactive Data Programmingcited by 10
  7. 2026CARE: Confounder-Aware Aggregation for Reliable LLM Evaluationcited by 10
  8. 2023Multi-annotator Deep Learning: A Probabilistic Framework for Classificationcited by 8
  9. 2022MiDAS: Multi-integrated Domain Adaptive Supervision for Fake News Detectioncited by 7
  10. 2022Adaptive Ranking-based Sample Selection for Weakly Supervised Class-imbalanced Text Classificationcited by 7
  11. 2022Leveraging Instance Features for Label Aggregation in Programmatic Weak Supervisioncited by 6
  12. 2022WeCheck: Strong Factual Consistency Checker via Weakly Supervised Learningcited by 6
21 more
  1. 2022Ground Truth Inference for Weakly Supervised Entity Matchingcited by 5
  2. 2023Pareto Optimal Learning for Estimating Large Language Model Errorscited by 5
  3. 2022SepLL: Separating Latent Class Labels from Weak Supervision Noisecited by 4
  4. 2023Analyzing the impact of climate change on critical infrastructure from the scientific literature: A weakly supervised NLP approachcited by 4
  5. 2023Making Binary Classification from Multiple Unlabeled Datasets Almost Free of Supervisioncited by 4
  6. 2023Weak Supervision Performance Evaluation via Partial Identificationcited by 4
  7. 2022Evaluating Generalizability of Fine-Tuned Models for Fake News Detectioncited by 3
  8. 2024Annot-Mix: Learning with Noisy Class Labels from Multiple Annotators via a Mixup Extensioncited by 3
  9. 2024dopanim: A Dataset of Doppelganger Animals with Noisy Annotations from Multiple Humanscited by 3
  10. 2022ULF: Unsupervised Labeling Function Correction using Cross-Validation for Weak Supervisioncited by 1
  11. 2022XPASC: Measuring Generalization in Weak Supervision by Explainability and Associationcited by 1
  12. 2022Time-Aware Datasets are Adaptive Knowledgebases for the New Normalcited by 1
  13. 2023Fusing Conditional Submodular GAN and Programmatic Weak Supervisioncited by 1
  14. 2024Convergence Behavior of an Adversarial Weak Supervision Methodcited by 1
  15. 2024Mislabeled examples detection viewed as probing machine learning models: concepts, survey and extensive benchmarkcited by 1
  16. 2025Medifact at PerAnsSumm 2025: Leveraging Lightweight Models for Perspective-Specific Summarization of Clinical Q&A Forumscited by 1
  17. 2025Learning from Interval Targetscited by 1
  18. 2022Generative Modeling Helps Weak Supervision (and Vice Versa)cited by 0
  19. 2023Weakly Supervised Label Learning Flowscited by 0
  20. 2023Local Boosting for Weakly-Supervised Learningcited by 0
  21. 2025Controllable Pareto Trade-off between Fairness and Accuracycited by 0