mlsanityData to 7 Oct 2026

Interpreting Predictive Probabilities: Model Confidence or Human Label Variation?

arXiv 2402.16102 ↗First posted 25 Feb 2024cs.CLcited by 15

Not on the map. Why?

Abstract (arXiv)

With the rise of increasingly powerful and user-facing NLP systems, there is growing interest in assessing whether they have a good representation of uncertainty by evaluating the quality of their predictive distribution over outcomes. We identify two main perspectives that drive starkly different evaluation protocols. The first treats predictive probability as an indication of model confidence; the second as an indication of human label variation. We discuss their merits and limitations, and take the position that both are crucial for trustworthy and fair NLP systems, but that exploiting a single predictive distribution is limiting. We recommend tools and highlight exciting directions towards models with disentangled representations of uncertainty about predictions and uncertainty about human labels.

Key facts. Cited by 15 papers in the map (data through 7 Oct 2026). Builds on 46 of its 92 references that are in the map. First posted on arXiv 25 Feb 2024.

Builds on

46 of 92 references.

References · 46

  1. 2017On Calibration of Modern Neural Networkscited by 4,860
  2. 2019Aleatoric and Epistemic Uncertainty in Machine Learning: An Introduction to Concepts and Methodscited by 903
  3. 2020GoEmotions: A Dataset of Fine-Grained Emotionscited by 479
  4. 2019Measuring Calibration in Deep Learningcited by 443
  5. 2020How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answeringcited by 439
  6. 2020AmbigQA: Answering Ambiguous Open-domain Questionscited by 385
  7. 2019Beyond temperature scaling: Obtaining well-calibrated multiclass probabilities with Dirichlet calibrationcited by 359
  8. 2019Verified Uncertainty Calibrationcited by 307
  9. 2020Calibration of Pre-trained Transformerscited by 307
  10. 2021Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotationscited by 306
  11. 2020Effect of Confidence and Explanation on Accuracy and Trust Calibration in AI-Assisted Decision Makingcited by 281
  12. 2021Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotationscited by 253
34 more
  1. 2019Human uncertainty makes classification more robustcited by 231
  2. 2020Selective Question Answering under Domain Shiftcited by 228
  3. 2020Unsupervised Quality Estimation for Neural Machine Translationcited by 210
  4. 2021Deep Deterministic Uncertainty: A Simple Baselinecited by 200
  5. 2019Are We Modeling the Task or the Annotator? An Investigation of Annotator Bias in Natural Language Understanding Datasetscited by 173
  6. 2019Evaluating model calibration in classificationcited by 165
  7. 2020What Can We Learn from Collective Human Opinions on Natural Language Inference Data?cited by 148
  8. 2021MultiVerS: Improving scientific claim verification with weak supervision and full-document contextcited by 140
  9. 2023We're Afraid Language Models Aren't Modeling Ambiguitycited by 124
  10. 2020Calibration of Neural Networks using Splinescited by 97
  11. 2021Is GPT-3 Text Indistinguishable from Human Text? Scarecrow: A Framework for Scrutinizing Machine Textcited by 89
  12. 2019Calibration tests in multi-class classification: A unifying frameworkcited by 79
  13. 2022Stop Measuring Calibration When Humans Disagreecited by 70
  14. 2023Uncertainty in Natural Language Generation: From Theory to Applicationscited by 61
  15. 2022Annotation Error Detection: Analyzing the Past and Present for a More Coherent Futurecited by 45
  16. 2022Eliciting and Learning with Soft Labels from Every Annotatorcited by 43
  17. 2022QAMPARI: An Open-domain Question Answering Benchmark for Questions with Many Answers from Multiple Paragraphscited by 42
  18. 2021AmbiFC: Fact-Checking Ambiguous Claims with Evidencecited by 38
  19. 2020Uncertainty Estimation Using a Single Deep Deterministic Neural Networkcited by 37
  20. 2021Uncertainty-Aware Machine Translation Evaluationcited by 35
  21. 2018Confidence Modeling for Neural Semantic Parsingcited by 33
  22. 2022On the Calibration of Pre-trained Language Models using Mixup Guided by Area Under the Margin and Saliencycited by 31
  23. 2021Learning with Different Amounts of Annotation: From Zero to Many Labelscited by 28
  24. 2023Sources of Uncertainty in Supervised Machine Learning -- A Statisticians' Viewcited by 27
  25. 2022Uncalibrated Models Can Improve Human-AI Collaborationcited by 26
  26. 2022Exploring Predictive Uncertainty and Calibration in NLP: A Study on the Impact of Method & Data Scarcitycited by 24
  27. 2023A Diachronic Perspective on User Trust in AI under Uncertaintycited by 23
  28. 2020Parameter Space Factorization for Zero-Shot Learning across Tasks and Languagescited by 22
  29. 2021Embracing Ambiguity: Shifting the Training Target of NLI Modelscited by 22
  30. 2021Transformer Uncertainty Estimation with Hierarchical Stochastic Attentioncited by 21
  31. 2016Scalable Bayesian Learning of Recurrent Neural Networks for Language Modelingcited by 9
  32. 2023ActiveAED: A Human in the Loop Improves Annotation Error Detectioncited by 9
  33. 2022Calibration Meets Explanation: A Simple and Effective Approach for Model Confidence Estimatescited by 7
  34. 2023Understanding and Predicting Human Label Variation in Natural Language Inference through Explanationcited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 6; 2025: 6; 2026: 3.

15 papers.

Sort:

By area:

  1. 2024Conformal Prediction for Natural Language Processing: A Surveycited by 52
  2. 2025Specializing Large Language Models to Simulate Survey Response Distributions for Global Populationscited by 50
  3. 2025Token Weighting for Long-Range Language Modelingcited by 14
  4. 2024Improving Health Question Answering with Reliable and Time-Aware Evidence Retrievalcited by 12
  5. 2024Which Demographics do LLMs Default to During Annotation?cited by 10
  6. 2025Always Tell Me The Odds: Fine-grained Conditional Probability Estimationcited by 9
  7. 2024Benchmark Transparency: Measuring the Impact of Data on Evaluationcited by 4
  8. 2024Understanding The Effect Of Temperature On Alignment With Human Opinionscited by 4
  9. 2025Uncertainty Quantification for Evaluating Machine Translation Biascited by 2
  10. 2025Do Language Models Update their Forecasts with New Information?cited by 2
  11. 2026Uncertainty Decomposition via Cyclical SG-MCMC and Soft-label Learning for Subjective NLPcited by 2
  12. 2024The Craft of Selective Prediction: Towards Reliable Case Outcome Classification -- An Empirical Study on European Court of Human Rights Casescited by 1
3 more
  1. 2025Express Your Doubts -- Probabilistic World Modeling Should not be Based on Token logprobscited by 1
  2. 2026Clarify, Abstain or Answer? Strategising in Conversation with Belief-Augmented Generationcited by 0
  3. 2026The Role of Ambiguity in Error Prediction via Uncertainty Quantificationcited by 0