mlsanityData to 14 Oct 2026

Beyond Losses Reweighting: Empowering Multi-Task Learning via the Generalization Perspective

arXiv 2211.13723 ↗First posted 24 Nov 2022cs.LG cs.AIcited by 15

Not on the map. Why?

Abstract (arXiv)

Multi-task learning (MTL) trains deep neural networks to optimize several objectives simultaneously using a shared backbone, which leads to reduced computational costs, improved data efficiency, and enhanced performance through cross-task knowledge sharing. Although recent gradient manipulation techniques aim to find a common descent direction that benefits all tasks, conventional empirical loss minimization still leaves models vulnerable to overfitting and gradient conflicts. To address this, we introduce a novel MTL framework that leverages weight perturbation to regulate gradient norms, thus improving generalization. By adaptively modulating weight perturbations, our approach harmonizes task-specific gradients, reducing conflicts and encouraging more robust learning across tasks. Theoretical insights reveal that controlling the gradient norm through weight perturbation directly contributes to better generalization. Extensive experiments across diverse applications demonstrate that our method significantly outperforms existing gradient-based MTL techniques in terms of task performance and overall model robustness.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Builds on 56 of its 86 references that are in the map. First posted on arXiv 24 Nov 2022.

Builds on

56 of 86 references.

References · 56

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2017On Calibration of Modern Neural Networkscited by 4,960
  3. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  4. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
  5. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
  6. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  7. 2015SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentationcited by 1,487
  8. 2020Sharpness-Aware Minimization for Efficiently Improving Generalizationcited by 1,417
  9. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,358
  10. 2016On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minimacited by 1,352
  11. 2019Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shiftcited by 1,287
  12. 2020Gradient Surgery for Multi-Task Learningcited by 1,256
38 more
  1. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,209
  2. 2017Visualizing the Loss Landscape of Neural Netscited by 1,130
  3. 2017GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networkscited by 834
  4. 2018Multi-Task Learning as Multi-Objective Optimizationcited by 821
  5. 2017An Overview of Multi-Task Learning in Deep Neural Networkscited by 803
  6. 2017Exploring Generalization in Deep Learningcited by 600
  7. 2018Predictive Uncertainty Estimation via Prior Networkscited by 557
  8. 2018End-to-End Multi-Task Learning with Attentioncited by 534
  9. 2017Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Datacited by 533
  10. 2019Fantastic Generalization Measures and Where to Find Themcited by 513
  11. 2017Sharp Minima Can Generalize For Deep Netscited by 466
  12. 2017Regularizing Neural Networks by Penalizing Confident Output Distributionscited by 464
  13. 2021Conflict-Averse Gradient Descent for Multi-task Learningcited by 459
  14. 2021SWAD: Domain Generalization by Seeking Flat Minimacited by 434
  15. 2016Entropy-SGD: Biasing Gradient Descent Into Wide Valleyscited by 359
  16. 2019Multi-Task Deep Neural Networks for Natural Language Understandingcited by 350
  17. 2021ASAM: Adaptive Sharpness-Aware Minimization for Scale-Invariant Learning of Deep Neural Networkscited by 293
  18. 2021When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentationscited by 260
  19. 2019Pareto Multi-Task Learningcited by 224
  20. 2022Multi-Task Learning as a Bargaining Gamecited by 215
  21. 2020Just Pick a Sign: Optimizing Deep Multitask Models with Gradient Sign Dropoutcited by 201
  22. 2020Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Modelscited by 169
  23. 2019Attentive Single-Tasking of Multiple Taskscited by 148
  24. 2022Generalized Federated Learning via Sharpness Aware Minimizationcited by 144
  25. 2021Efficient Sharpness-aware Minimization for Improved Training of Neural Networkscited by 137
  26. 2022Towards Efficient and Scalable Sharpness-Aware Minimizationcited by 129
  27. 2020Learning the Pareto Front with Hypernetworkscited by 125
  28. 2022Penalizing Gradient Norm for Efficiently Improving Generalization in Deep Learningcited by 122
  29. 2015On the properties of variational approximations of Gibbs posteriorscited by 121
  30. 2021Sharpness-Aware Minimization Improves Language Model Generalizationcited by 108
  31. 2021Reasonable Effectiveness of Random Weighting: A Litmus Test for Multi-Task Learningcited by 101
  32. 2022Improving Generalization in Federated Learning by Seeking Flat Minimacited by 98
  33. 2019Asymmetric Valleys: Beyond Sharp and Flat Local Minimacited by 90
  34. 2020Relative Flatness and Generalizationcited by 88
  35. 2022Understanding the Generalization Benefit of Normalization Layers: Sharpness Reductioncited by 86
  36. 2021RotoGrad: Gradient Homogenization in Multitask Learningcited by 83
  37. 2020Regularizing Neural Networks via Adversarial Model Perturbationcited by 78
  38. 2021Flattening Sharpness for Dynamic Gradient Projection Memory Benefits Continual Learningcited by 72

50 most cited of 56 All 56 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 4; 2025: 6; 2026: 3.

15 papers.

Sort:

By area:

  1. 2025Selective Task Group Updates for Multi-Task Optimizationcited by 17
  2. 2025Mitigating Parameter Interference in Model Merging via Sharpness-Aware Fine-Tuningcited by 16
  3. 2023Flat Seeking Bayesian Neural Networkscited by 10
  4. 2024Controllable Prompt Tuning For Balancing Group Distributional Robustnesscited by 10
  5. 2024Enhancing Domain Adaptation through Prompt Gradient Alignmentcited by 8
  6. 2025Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Modelscited by 5
  7. 2025Unveiling m-Sharpness Through the Structure of Stochastic Gradient Noisecited by 2
  8. 2023Sharpness & Shift-Aware Self-Supervised Learningcited by 1
  9. 2024Diversity-Aware Agnostic Ensemble of Sharpness Minimizerscited by 1
  10. 2026Sharpness-Aware Minimization in Logit Space Efficiently Enhances Direct Preference Optimizationcited by 1
  11. 2024DataFreeShield: Defending Adversarial Attacks without Training Datacited by 0
  12. 2025DGSAM: Domain Generalization via Individual Sharpness-Aware Minimizationcited by 0
3 more
  1. 2025BoundarySeg:Boundary-Aware Multi-Task Learning for Semi-Supervised 3D Medical Image Segmentation with Limited Unlabeled Datacited by 0
  2. 2026Flatness and Gradient Alignment Are Both Necessary: Spectral-Aware Gradient-Aligned Exploration for Multi-Distribution Learningcited by 0
  3. 2026SUM: Unified Geometric Surgery on Spatio-Temporal Adaptation Vectors for Federated Class Incremental Learningcited by 0