mlsanityData to 7 Oct 2026

Robust Learning Through Cross-Task Consistency

arXiv 2006.04096 ↗First posted 7 Jun 2020cs.CV cs.GR cs.LGcited by 105

Not on the map. Why?

Abstract (arXiv)

Visual perception entails solving a wide set of tasks, e.g., object detection, depth estimation, etc. The predictions made for multiple tasks from the same image are not independent, and therefore, are expected to be consistent. We propose a broadly applicable and fully computational method for augmenting learning with Cross-Task Consistency. The proposed formulation is based on inference-path invariance over a graph of arbitrary tasks. We observe that learning with cross-task consistency leads to more accurate predictions and better generalization to out-of-distribution inputs. This framework also leads to an informative unsupervised quantity, called Consistency Energy, based on measuring the intrinsic consistency of the system. Consistency Energy correlates well with the supervised error (r=0.67), thus it can be employed as an unsupervised confidence metric as well as for detection of out-of-distribution inputs (ROC-AUC=0.95). The evaluations are performed on multiple datasets, including Taskonomy, Replica, CocoDoom, and ApolloScape, and they benchmark cross-task consistency versus various baselines including conventional multi-task learning, cycle consistency, and analytical consistency.

Key facts. Cited by 105 papers in the map (data through 7 Oct 2026). First posted on arXiv 7 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 28; 2022: 22; 2023: 30; 2024: 11; 2025: 11; 2026: 3.

105 papers.

Sort:

By area:

  1. 2021Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scanscited by 336
  2. 2021UniT: Multimodal Multitask Learning with a Unified Transformercited by 195
  3. 2021No-Reference Image Quality Assessment via Transformers, Relative Ranking, and Self-Consistencycited by 165
  4. 2021Geometry Uncertainty Projection Network for Monocular 3D Object Detectioncited by 155
  5. 2024Rethinking Inductive Biases for Surface Normal Estimationcited by 122
  6. 20223D Common Corruptions and Data Augmentationcited by 120
  7. 20234M: Massively Multimodal Masked Modelingcited by 118
  8. 2022Auto-Lambda: Disentangling Dynamic Task Relationshipscited by 73
  9. 2021Weakly Supervised Learning of Rigid 3D Scene Flowcited by 60
  10. 2021Shape from Polarization for Complex Scenes in the Wildcited by 53
  11. 2022Hierarchical Normalization for Robust Monocular Depth Estimationcited by 51
  12. 2022MulT: An End-to-End Multitask Learning Transformercited by 47
38 more
  1. 2021Deep Learning for HDR Imaging: State-of-the-Art and Future Trendscited by 45
  2. 2023TaskExpert: Dynamically Assembling Multi-Task Representations with Memorial Mixture-of-Expertscited by 45
  3. 2023StyleGAN knows Normal, Depth, Albedo, and Morecited by 44
  4. 2021Learning Multiple Dense Prediction Tasks from Partially Annotated Datacited by 42
  5. 2023Single-view 3D Scene Reconstruction with High-fidelity Shape and Texturecited by 37
  6. 2021Discrete Representations Strengthen Vision Transformer Robustnesscited by 34
  7. 2021CrossLoc: Scalable Aerial Localization Assisted by Multimodal Synthetic Datacited by 28
  8. 2023Learning to Ground Instructional Articles in Videos through Narrationscited by 28
  9. 2021Towards a Collective Agenda on AI for Earth Science Data Analysiscited by 27
  10. 2022MGNet: Monocular Geometric Scene Understanding for Autonomous Drivingcited by 27
  11. 2021Robustness via Cross-Domain Ensemblescited by 23
  12. 2021Cerberus Transformer: Joint Semantic, Affordance and Attribute Parsingcited by 22
  13. 2022Egocentric Video Task Translationcited by 19
  14. 2024DiffusionMTL: Learning Multi-Task Denoising Diffusion Model from Partially Annotated Datacited by 18
  15. 2023XVO: Generalized Visual Odometry via Cross-Modal Self-Trainingcited by 17
  16. 2024Semantic-CC: Boosting Remote Sensing Image Change Captioning via Foundational Knowledge and Semantic Guidancecited by 17
  17. 2022Consistency Learning via Decoding Path Augmentation for Transformers in Human Object Interaction Detectioncited by 16
  18. 2023Relational Multi-Task Learning: Modeling Relations between Data and Taskscited by 15
  19. 2021Generative Modeling for Multi-task Visual Learningcited by 14
  20. 2021Towards a Unified Foundation Model: Jointly Pre-Training Transformers on Unpaired Images and Textcited by 14
  21. 2023Vision Transformer Adapters for Generalizable Multitask Learningcited by 14
  22. 2024Supervised Fine-tuning in turn Improves Visual Foundation Modelscited by 13
  23. 2021Environment Predictive Coding for Embodied Agentscited by 12
  24. 2021Channel Exchanging Networks for Multimodal and Multitask Dense Image Predictioncited by 11
  25. 2022Seg4Reg+: Consistency Learning between Spine Segmentation and Cobb Angle Regressioncited by 11
  26. 2023Semantic Human Parsing via Scalable Semantic Transfer over Multiple Label Domainscited by 11
  27. 2022Beyond RGB: Scene-Property Synthesis with Neural Radiance Fieldscited by 10
  28. 2022GOOD: Exploring Geometric Cues for Detecting Objects in an Open Worldcited by 10
  29. 2023Rapid Network Adaptation: Learning to Adapt Neural Networks Using Test-Time Feedbackcited by 10
  30. 2023Human-oriented Representation Learning for Robotic Manipulationcited by 10
  31. 2022Minkowski Tracker: A Sparse Spatio-Temporal R-CNN for Joint Object Detection and Trackingcited by 9
  32. 2022ProCC: Progressive Cross-primitive Compatibility for Open-World Compositional Zero-Shot Learningcited by 9
  33. 2022Robust Perception through Equivariancecited by 9
  34. 2021See Yourself in Others: Attending Multiple Tasks for Own Failure Detectioncited by 8
  35. 2021PlaneRecNet: Multi-Task Learning with Cross-Task Consistency for Piece-Wise Plane Detection and Reconstruction from a Single RGB Imagecited by 8
  36. 2023Efficient Computation Sharing for Multi-Task Visual Scene Understandingcited by 8
  37. 2023Self-supervised Hypergraphs for Learning Multiple World Interpretationscited by 8
  38. 2023Multi-task learning with cross-task consistency for improved depth estimation in colonoscopycited by 8

50 most cited of 105 All 105 (JSON)