mlsanityData to 7 Oct 2026

What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?

arXiv 1703.04977 ↗First posted 15 Mar 2017cs.CVcited by 2,335

Not on the map. Why?

Cited by 45 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

There are two major types of uncertainty one can model. Aleatoric uncertainty captures noise inherent in the observations. On the other hand, epistemic uncertainty accounts for uncertainty in the model -- uncertainty which can be explained away given enough data. Traditionally it has been difficult to model epistemic uncertainty in computer vision, but with new Bayesian deep learning tools this is now possible. We study the benefits of modeling epistemic vs. aleatoric uncertainty in Bayesian deep learning models for vision tasks. For this we present a Bayesian deep learning framework combining input-dependent aleatoric uncertainty together with epistemic uncertainty. We study models under the framework with per-pixel semantic segmentation and depth regression tasks. Further, our explicit uncertainty formulation leads to new loss functions for these tasks, which can be interpreted as learned attenuation. This makes the loss more robust to noisy data, also giving new state-of-the-art results on segmentation and depth regression benchmarks.

Key facts. Cited by 2,335 papers in the map (data through 7 Oct 2026). First posted on arXiv 15 Mar 2017.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2018: 2; 2019: 2; 2020: 4; 2021: 411; 2022: 345; 2023: 387; 2024: 322; 2025: 396; 2026: 466.

2,335 papers.

Sort:

By area:

  1. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problemscited by 1,873
  2. 2025VGGT: Visual Geometry Grounded Transformercited by 1,475
  3. 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 795
  4. 2021A Survey of Uncertainty in Deep Neural Networkscited by 675
  5. 2018Scalable agent alignment via reward modeling: a research directioncited by 438
  6. 2021Score-based diffusion models for accelerated MRIcited by 308
  7. 2021What Are Bayesian Neural Network Posteriors Really Like?cited by 307
  8. 2022Diffusion Models for Medical Image Analysis: A Comprehensive Surveycited by 269
  9. 2023A Survey of Confidence Estimation and Calibration in Large Language Modelscited by 253
  10. 2023Visual Geometry Grounded Deep Structure From Motioncited by 225
  11. 2024MegaSaM: Accurate, Fast, and Robust Structure and Motion from Casual Dynamic Videoscited by 224
  12. 2022Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labelscited by 218
38 more
  1. 2021Deep Deterministic Uncertainty: A Simple Baselinecited by 200
  2. 2021Back to the Feature: Learning Robust Camera Localization from Pixels to Posecited by 199
  3. 2021Diffusion Models for Implicit Image Segmentation Ensemblescited by 198
  4. 2021Decoupling Zero-Shot Semantic Segmentationcited by 198
  5. 2022BEVerse: Unified Perception and Prediction in Birds-Eye-View for Vision-Centric Autonomous Drivingcited by 191
  6. 2021Objects are Different: Flexible Monocular 3D Object Detectioncited by 184
  7. 2021ST++: Make Self-training Work Better for Semi-supervised Semantic Segmentationcited by 184
  8. 2021On Hallucination and Predictive Uncertainty in Conditional Language Generationcited by 176
  9. 2023FengWu: Pushing the Skillful Global Medium-range Weather Forecast beyond 10 Days Leadcited by 170
  10. 2022Trusted Multi-View Classification with Dynamic Evidential Fusioncited by 167
  11. 2022BinsFormer: Revisiting Adaptive Bins for Monocular Depth Estimationcited by 166
  12. 2023Speculative Decoding with Big Little Decodercited by 163
  13. 2021Geometry Uncertainty Projection Network for Monocular 3D Object Detectioncited by 155
  14. 2021SegmentMeIfYouCan: A Benchmark for Anomaly Segmentationcited by 150
  15. 2021Self-supervised Augmentation Consistency for Adapting Semantic Segmentationcited by 149
  16. 2023Shifting Attention to Relevance: Towards the Predictive Uncertainty Quantification of Free-Form Large Language Modelscited by 147
  17. 2021Deep Learning Through the Lens of Example Difficultycited by 139
  18. 2021Delving into Localization Errors for Monocular 3D Object Detectioncited by 138
  19. 2022Quantifying Aleatoric and Epistemic Uncertainty in Machine Learning: Are Conditional Entropy and Mutual Information Appropriate Measures?cited by 124
  20. 2021Bayesian Neural Network Priors Revisitedcited by 121
  21. 2021Epistemic Neural Networkscited by 120
  22. 2022On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networkscited by 119
  23. 2021SMD-Nets: Stereo Mixture Density Networkscited by 118
  24. 2021Estimating and Exploiting the Aleatoric Uncertainty in Surface Normal Estimationcited by 116
  25. 2024Language Model Cascades: Token-level uncertainty and beyondcited by 115
  26. 2021Learning Accurate Dense Correspondences and When to Trust Themcited by 113
  27. 2021Pixel-wise Anomaly Detection in Complex Driving Scenescited by 109
  28. 2021Mutual Consistency Learning for Semi-supervised Medical Image Segmentationcited by 107
  29. 2022StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberationcited by 105
  30. 2021Semi-supervised Left Atrium Segmentation with Mutual Consistency Trainingcited by 104
  31. 2021Recent advances and clinical applications of deep learning in medical image analysiscited by 104
  32. 2022Uncertainty Quantification with Pre-trained Language Models: A Large-Scale Empirical Analysiscited by 104
  33. 2023A Comprehensive Survey on Source-free Domain Adaptationcited by 104
  34. 2024Back to Basics: Revisiting REINFORCE Style Optimization for Learning from Human Feedback in LLMscited by 103
  35. 2021DEUP: Direct Epistemic Uncertainty Predictioncited by 102
  36. 2022CARD: Classification and Regression Diffusion Modelscited by 102
  37. 2022Plex: Towards Reliability using Pretrained Large Model Extensionscited by 102
  38. 2023Probabilistic Concept Bottleneck Modelscited by 101

50 most cited of 2,335 All 2,335 (JSON)