mlsanityData to 14 Oct 2026

Auxiliary Learning as an Asymmetric Bargaining Game

arXiv 2301.13501 ↗First posted 31 Jan 2023cs.LGcited by 12

Not on the map. Why?

Abstract (arXiv)

Auxiliary learning is an effective method for enhancing the generalization capabilities of trained models, particularly when dealing with small datasets. However, this approach may present several difficulties: (i) optimizing multiple objectives can be more challenging, and (ii) how to balance the auxiliary tasks to best assist the main task is unclear. In this work, we propose a novel approach, named AuxiNash, for balancing tasks in auxiliary learning by formalizing the problem as generalized bargaining game with asymmetric task bargaining power. Furthermore, we describe an efficient procedure for learning the bargaining power of tasks based on their contribution to the performance of the main task and derive theoretical guarantees for its convergence. Finally, we evaluate AuxiNash on multiple multi-task benchmarks and find that it consistently outperforms competing methods.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Builds on 44 of its 67 references that are in the map. First posted on arXiv 31 Jan 2023.

Builds on

44 of 67 references.

References · 44

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,444
  3. 2015SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentationcited by 1,487
  4. 2020Gradient Surgery for Multi-Task Learningcited by 1,256
  5. 2017GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networkscited by 834
  6. 2018Multi-Task Learning as Multi-Objective Optimizationcited by 821
  7. 2017An Overview of Multi-Task Learning in Deep Neural Networkscited by 803
  8. 2018Speech Commands: A Dataset for Limited-Vocabulary Speech Recognitioncited by 752
  9. 2017Dynamic Routing Between Capsulescited by 689
  10. 2018End-to-End Multi-Task Learning with Attentioncited by 534
  11. 2019Meta-Learning with Implicit Gradientscited by 525
  12. 2016Cross-stitch Networks for Multi-task Learningcited by 467
32 more
  1. 2021Conflict-Averse Gradient Descent for Multi-task Learningcited by 459
  2. 2018Bilevel Programming for Hyperparameter Optimization and Meta-Learningcited by 405
  3. 2019Which Tasks Should Be Learned Together in Multi-task Learning?cited by 375
  4. 2019S4L: Self-Supervised Semi-Supervised Learningcited by 335
  5. 2019Optimizing Millions of Hyperparameters by Implicit Differentiationcited by 304
  6. 2016Hyperparameter optimization with approximate gradientcited by 233
  7. 2022Multi-Task Learning as a Bargaining Gamecited by 215
  8. 2020Just Pick a Sign: Optimizing Deep Multitask Models with Gradient Sign Dropoutcited by 201
  9. 2020Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Modelscited by 169
  10. 2021Investigating Bi-Level Optimization for Learning and Vision from a Unified Perspective: A Survey and Beyondcited by 153
  11. 2019Attentive Single-Tasking of Multiple Taskscited by 148
  12. 2017A Review on Bilevel Optimization: From Classical to Evolutionary Approaches and Applicationscited by 129
  13. 2015Instance-aware Semantic Segmentation via Multi-task Network Cascadescited by 121
  14. 2019Self-Tuning Networks: Bilevel Optimization of Hyperparameters using Structured Best-Response Functionscited by 105
  15. 2018Adapting Auxiliary Losses Using Gradient Similaritycited by 100
  16. 2020Self-Supervised Learning for Domain Adaptation on Point-Cloudscited by 99
  17. 2019Self-Supervised Generalisation with Meta Auxiliary Learningcited by 88
  18. 2021RotoGrad: Gradient Homogenization in Multitask Learningcited by 83
  19. 2022Auto-Lambda: Disentangling Dynamic Task Relationshipscited by 73
  20. 2015Scalable Gradient-Based Tuning of Continuous Regularization Hyperparameterscited by 71
  21. 2018Reviving and Improving Recurrent Back-Propagationcited by 69
  22. 2016A Joint Many-Task Model: Growing a Neural Network for Multiple NLP Taskscited by 62
  23. 2021Part-aware Panoptic Segmentationcited by 56
  24. 2020Auxiliary Learning by Implicit Differentiationcited by 45
  25. 2019Ray Interference: a Source of Plateaus in Deep Reinforcement Learningcited by 40
  26. 2019Bi-level Actor-Critic for Multi-agent Coordinationcited by 39
  27. 2020Self-supervised Auxiliary Learning with Meta-paths for Heterogeneous Graphscited by 38
  28. 2016Learning to Push by Grasping: Using multiple tasks for effective learningcited by 37
  29. 2022On Implicit Bias in Overparameterized Bilevel Optimizationcited by 30
  30. 2020Auxiliary Task Reweighting for Minimum-data Learningcited by 23
  31. 2020Teaching with Commentariescited by 16
  32. 2022AANG: Automating Auxiliary Learningcited by 12

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 8; 2025: 3; 2026: 1.

12 papers.

  1. 2024LLM-Rubric: A Multidimensional, Calibrated Approach to Automated Evaluation of Natural Language Textscited by 87
  2. 2024Bayesian Uncertainty for Gradient Aggregation in Multi-Task Learningcited by 11
  3. 2024Enhancing Domain Adaptation through Prompt Gradient Alignmentcited by 8
  4. 2024Using dynamic loss weighting to boost improvements in forecast stabilitycited by 4
  5. 2025Expert Merging in Sparse Mixture of Experts with Nash Bargainingcited by 3
  6. 2025Model Merging via Multi-Teacher Knowledge Distillationcited by 2
  7. 2024Enhancing Molecular Property Prediction with Auxiliary Learning and Task-Specific Adaptationcited by 1
  8. 2024Rethinking Teacher-Student Curriculum Learning through the Cooperative Mechanics of Experiencecited by 1
  9. 2024Learning Representation for Multitask learning through Self Supervised Auxiliary learningcited by 1
  10. 2026Not All Objectives Are Born Equal: Priority-Constrained Descent for Hierarchical Multi-Objective Optimizationcited by 1
  11. 2024Two-Stage Multi-task Self-Supervised Learning for Medical Image Segmentationcited by 0
  12. 2025Generating Auxiliary Tasks with Reinforcement Learningcited by 0