mlsanityData to 7 Oct 2026

Sample-Level Weighting for Multi-Task Learning with Auxiliary Tasks

arXiv 2306.04519 ↗First posted 7 Jun 2023cs.LG cs.AIcited by 6

Not on the map. Why?

Abstract (arXiv)

Multi-task learning (MTL) can improve the generalization performance of neural networks by sharing representations with related tasks. Nonetheless, MTL can also degrade performance through harmful interference between tasks. Recent work has pursued task-specific loss weighting as a solution for this interference. However, existing algorithms treat tasks as atomic, lacking the ability to explicitly separate harmful and helpful signals beyond the task level. To this end, we propose SLGrad, a sample-level weighting algorithm for multi-task learning with auxiliary tasks. Through sample-specific task weights, SLGrad reshapes the task distributions during training to eliminate harmful auxiliary signals and augment useful task signals. Substantial generalization performance gains are observed on (semi-) synthetic datasets and common supervised multi-task problems.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 35 of its 63 references that are in the map. First posted on arXiv 7 Jun 2023.

Builds on

35 of 63 references.

References · 35

  1. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  2. 2018Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentationcited by 3,210
  3. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,345
  4. 2020Gradient Surgery for Multi-Task Learningcited by 1,210
  5. 2020Meta-Learning in Neural Networks: A Surveycited by 944
  6. 2017GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networkscited by 816
  7. 2018Multi-Task Learning as Multi-Objective Optimizationcited by 809
  8. 2017An Overview of Multi-Task Learning in Deep Neural Networkscited by 791
  9. 2017Dynamic Routing Between Capsulescited by 689
  10. 2018Learning to Reweight Examples for Robust Deep Learningcited by 670
  11. 2017MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labelscited by 652
  12. 2021Deep Learning on a Data Diet: Finding Important Examples Early in Trainingcited by 529
23 more
  1. 2021Conflict-Averse Gradient Descent for Multi-task Learningcited by 441
  2. 2021Opacus: User-Friendly Differential Privacy Library in PyTorchcited by 375
  3. 2018Not All Samples Are Created Equal: Deep Learning with Importance Samplingcited by 323
  4. 2018What is the Effect of Importance Weighting in Deep Learning?cited by 250
  5. 2021Efficiently Identifying Task Groupings for Multi-Task Learningcited by 218
  6. 2020Just Pick a Sign: Optimizing Deep Multitask Models with Gradient Sign Dropoutcited by 199
  7. 2015Cost Sensitive Learning of Deep Feature Representations from Imbalanced Datacited by 167
  8. 2020Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Modelscited by 167
  9. 2021Reasonable Effectiveness of Random Weighting: A Litmus Test for Multi-Task Learningcited by 100
  10. 2018Adapting Auxiliary Losses Using Gradient Similaritycited by 98
  11. 2020Rethinking Importance Weighting for Deep Learning under Distribution Shiftcited by 92
  12. 2020A Second-Order Approach to Learning with Instance-Dependent Label Noisecited by 90
  13. 2021RotoGrad: Gradient Homogenization in Multitask Learningcited by 82
  14. 2021Learning Auxiliary Monocular Contexts Helps Monocular 3D Object Detectioncited by 66
  15. 2018Deep Auxiliary Learning for Visual Localization and Odometrycited by 59
  16. 2021Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learningcited by 49
  17. 2019MultiNet++: Multi-Stream Feature Aggregation and Geometric Loss Strategy for Multi-Task Learningcited by 45
  18. 2021Understanding the role of importance weighting for deep learningcited by 38
  19. 2020Dynamic Task Weighting Methods for Multi-task Networks in Autonomous Driving Systemscited by 34
  20. 2022MetaBalance: Improving Multi-Task Recommendations via Adapting Gradient Magnitudes of Auxiliary Taskscited by 34
  21. 2022LibMTL: A Python Library for Multi-Task Learningcited by 25
  22. 2017Multi-Task Convolutional Neural Network for Pose-Invariant Face Recognitioncited by 24
  23. 2021Instance-Level Task Parameters: A Robust Multi-task Weighting Frameworkcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 3; 2026: 2.

6 papers.

  1. 2025Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoningcited by 6
  2. 2024Using dynamic loss weighting to boost improvements in forecast stabilitycited by 4
  3. 2025Controlled Data Rebalancing in Multi-Task Learning for Real-World Image Super-Resolutioncited by 0
  4. 2025Generating Auxiliary Tasks with Reinforcement Learningcited by 0
  5. 2026When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretrainingcited by 0
  6. 2026Multi-Task Learning for Heterogeneous Prediction from Video Game State with Transfer Learningcited by 0