Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Proactive Gradient Conflict Mitigation in Multi-Task Learning: A Sparse Training Perspective

T0 review · 3 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Sparse training — updating only a fixed subset of encoder parameters chosen per neuron — cuts the incidence of gradient conflict in multi-task learning and improves all tested gradient-manipulation methods.

desk verdict The performance gains look real and broad, but the headline claim about reducing gradient conflict rests on an unstated metric choice that makes the mechanism unfalsifiable as written. read the letter →

arxiv 2411.18615 v1 pith:RTFNTMVH submitted 2024-11-27 cs.LG cs.AIcs.CV

Zhi Zhang , Jiayi Shen , Congfeng Cao , Gaole Dai , Shiji Zhou , Qizhe Zhang , Shanghang Zhang , Ekaterina Shutova This is my paper · ORCID
classification cs.LGcs.AIcs.CV
keywords multi-tasklearninggradientconflictsparsetrainingparameter-efficientfine-tuningmanipulationper-neuronselectionvisiontransformers
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Multi-task models often train several tasks through one shared encoder, and the tasks' gradients can point against each other — a phenomenon the paper calls gradient conflict. The paper argues that updating only a fixed subset of the encoder's parameters, chosen per neuron by magnitude, reduces how often such conflicts occur, and that this simple change improves multi-task performance across datasets and architectures. If true, this means a one-time mask applied before training is a cheap, architecture-preserving lever for multi-task learning, and it can be bolted onto existing gradient-manipulation methods such as PCGrad, CAGrad, GradDrop, MGDA, IMTL-G, and NashMTL to make them work better. The paper supports this with experiments on CNN and transformer backbones (MTAN, SAM, ViT, Swin) over dense-prediction and classification benchmarks.

What carries the argument

The central object is a fixed binary mask $M$ on the shared encoder, computed once before training by the parameter-selection rule the paper calls PSN (per-neuron selection). For each neuron, PSN keeps the top-$K$ input connections by magnitude, so every neuron stays trainable while most weights are frozen; the update is $\hat{\theta}_{sha} = \theta_{sha} - M \frac{1}{T}\sum_{t=1}^T \nabla_{\theta_{sha}} L_t$. The paper interprets this mask as projecting the averaged task gradient onto a lower-dimensional subspace, which reduces interference between tasks. The theoretical justification is a Lagrangian reformulation of the constrained objective into a regularized problem, citing the stability and generalization bound of parameter-efficient fine-tuning.

What would settle it

Re-run the NYU-v2/SAM setup and compute p% on the raw per-task gradients for both Joint Train and Joint Train with ST, with the mask applied only to the parameter update; if the gap between 31.89% and 26.33% vanishes or inverts, the conflict-reduction claim is an artifact of measuring on masked gradients.

Watch

Extended reading notes

Core claim

The paper's central claim is that sparse training (ST) — updating only the highest-magnitude incoming weights of each neuron in the shared encoder, with all other encoder parameters frozen — reduces the incidence of gradient conflict, defined as negative cosine similarity between per-task gradients. On the NYUv2 benchmark with the SAM backbone, the paper reports that ST lowers the average incidence of gradient conflict from 31.89% to 26.33% over all epochs, and from 35.85% to 29.14% over the last 50% of training. The same pattern holds when ST is combined with all six gradient-manipulation baselines, and the performance gains are consistent on CelebA, Clevr, SmallNORB, NYU-v2, and CityScapes. The paper also reports that larger pre-trained models exhibit more gradient conflict, making them the clearest beneficiaries of the strategy.

Load-bearing premise

The paper's central comparison assumes the gradient-conflict metric is measured on the same gradient quantity for both full and sparse training; Section 4.2 never states whether the incidence p% is computed on raw per-task gradients, post-manipulation gradients, or masked gradients.

Editorial extensions

If this is right

  • Sparse training can be added to any gradient-manipulation method without changing that method's update rule, and the paper shows consistent performance gains across all six tested methods.
  • The conflict reduction grows in later training epochs, suggesting the mask stabilizes optimization trajectories rather than merely shrinking the effective learning rate.
  • Larger pre-trained models show more gradient conflict and benefit more from sparsity, making this a low-cost route for multi-task fine-tuning of foundation models.
  • Because the mask is fixed before training, the computational overhead is negligible compared to methods that update masks per iteration.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's conflict metric may be measured on masked gradients for sparse training; if so, part of the reported reduction would be baked into the measurement. A cleaner comparison would compute the incidence on raw per-task gradients for all methods.
  • The per-neuron top-$K$ rule is effectively a structured-pruning prior; testing adaptive masks refreshed every few epochs could reveal whether the conflict reduction is a property of sparsity itself or of the fixed mask.
  • The paper's regularization argument implies that sparse training should improve fine-tuning stability even in single-task settings, which the paper does not test.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper proposes sparse training (ST) as a proactive mechanism to reduce gradient conflict (GC) in multi-task learning. The method selects a fixed subset of encoder parameters by per-neuron top-magnitude (PSN) and updates only those parameters, leaving the rest frozen; it can be combined with gradient-manipulation methods such as PCGrad, CAGrad, GradDrop, MGDA, IMTL-G, and NashMTL. The authors report that ST lowers the average incidence of GC across multiple datasets (NYUv2, CityScapes, CelebA, Clevr, SmallNORB) and architectures (MTAN, SAM, ViT, Swin), and that it improves overall task performance as measured by relative task drop (∆m%). They also provide ablations on sparsity ratio, parameter-selection strategy, and model size.

Significance. If the central claim holds, the paper identifies a simple, architecture-preserving lever for multi-task optimization: freezing most shared-encoder parameters via a static mask reduces gradient conflicts and improves downstream performance. The empirical sweep is broad and internally consistent, covering several dense-prediction and classification benchmarks, six (plus FAMO) gradient-manipulation baselines, and both randomly initialized and pre-trained vision backbones. The supplementary material includes three-seed results for MTAN on NYUv2 and a clear statement of hyperparameters. The main weakness is that the GC-incidence metric is under-specified, which makes the headline mechanism claim currently ambiguous and potentially circular; the theoretical section also contains mislabeled inequalities. These issues are fixable without changing the overall experimental design, but they must be resolved before the mechanism claim can be accepted.

major comments (3)
  1. [§4.1, Eq. for p%; Tables 1, 6, 10, 13] The definition of p% does not state whether gradient conflicts are measured on raw per-task gradients, on gradients after gradient manipulation, or on the masked gradients that sparse training actually updates. This is load-bearing because the core claim is that ST reduces the incidence of GC: if p% for ST is computed on M·g_i and M·g_j while p% for baselines is computed on raw g_i and g_j, the reported reductions (e.g., Joint Train 31.89%→26.33% in Table 1) are partly an artifact of coordinate projection, not an independent property of the optimization landscape. Figure 2 explicitly describes ST as an orthographic/coordinate projection, so the metric choice is not a minor implementation detail. Please specify the gradient representation for every method in the p% tables, and re-report the incidence values on a common representation (e.g., the actual update vectors used by each method).
  2. [§3.3 and Appendix A] The theoretical analysis is mathematically mislabeled and does not support the GC-reduction claim. The chain in Appendix A shows L = min_Θ max_λ ... ≥ max_λ min_Θ ... ≥ min_Θ (L + ||(I−M)(θ_sha−θ_in_sha)||_2), so the regularized objective in Eq. (12) is a lower bound on the primal value L, not an upper bound. The main text says "This can be transformed to optimize the upper bound L of regularized problem," which inverts the direction of the inequality. Moreover, Eq. (12) is a penalized relaxation, not an equivalent reformulation of the constrained problem in Eq. (10), and no argument connects this regularized objective to the actual projected-gradient updates used by ST. Please correct the inequality/labeling and either provide a rigorous link between sparse training and GC incidence or present this material explicitly as intuition.
  3. [Tables 2, 3, 7, 10, 11, 12] The main performance tables lack error bars for most configurations. Table 7 (MTAN on NYUv2) reports mean±std over three seeds, but the flagship results on SAM (Table 2) and the summaries on CelebA, Clevr, SmallNORB, and CityScapes (Table 3) are single point estimates. The paper repeatedly claims that ST "consistently improves" all methods, but without repeated-seed statistics it is impossible to assess whether differences such as Joint Train Δm% 6.763→5.314 are meaningful. Please provide repeated-seed results for at least the main comparisons, or state the number of seeds and quantify uncertainty.
minor comments (6)
  1. [§4.2, first paragraph] The phrase "any gradient magnitude methods" appears twice and should be "any gradient manipulation methods."
  2. [§3.4, Definition 2] The mask M is defined as an |Θ|×|Θ| diagonal matrix; this works formally, but it is more natural to define a boolean vector m ∈ {0,1}^|Θ| and write the update as θ − α m ⊙ ∇L(θ). Consider using the vector notation to avoid confusion.
  3. [Figure 5c caption] "(C) Different sparse methods" should be "(c) Different sparse methods" for consistency with the other panels.
  4. [§4.2, Table 6] In the MTAN results (Table 6), the reduction for MGDA w/ ST is only 0.39% over all epochs and 0.85% in the last 50%; this is a much weaker effect than for other methods. The text should acknowledge this exception rather than claiming uniform effectiveness.
  5. [Appendix D.2] The sentence "we employed a batch size of 3 and searched for the optimal learning rate from the set{2e-4, 5e-5}, and then the best results are reported" is grammatically awkward; split into two sentences.
  6. [References] The Recon paper [29] is cited only for the distribution of gradient angles, but Recon also proposes proactive GC reduction by converting conflicting layers into task-specific layers. Given that the present paper addresses the same goal, a brief comparison/discussion of the relationship between ST and Recon would help position the contribution.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the sparse-training claims are evaluated on external test metrics, and the ambiguous gradient-conflict metric is a consistency concern rather than a circular reduction.

full rationale

The central claims of the paper are empirical and externally grounded. The claim that sparse training (ST) reduces gradient-conflict incidence is supported by a measured statistic p% over training, and the performance improvements are reported on held-out test metrics such as mIoU, Abs Err, and Delta-m% relative to single-task learning; these do not reduce to the method's own definitions or fitted parameters. The main ambiguity is that Section 4.1 defines p% without stating whether conflicts are computed on raw per-task gradients or on the masked gradients used by ST, and Figure 2 describes ST as a coordinate projection of conflicting gradients. However, this is a measurement-consistency issue, not circularity: no equation defines p% as identical to the ST update, and masking does not by construction guarantee that the cosine similarity of two projected gradients becomes non-negative, so the reported reduction remains an empirical finding. The theoretical justification in Section 3.3 leans on an external result by Fu et al. [11], not on the authors' own prior work. Reference [40], the authors' own prior paper on gradient-based parameter selection, is cited only as related work motivating parameter selection and is not load-bearing for the paper's conclusions. No fitted parameter is renamed as a prediction, and no uniqueness theorem or ansatz is imported through self-citation. Therefore no circular step can be exhibited, and the paper should receive a no-circularity verdict, with the metric ambiguity noted as a correctness and consistency risk rather than circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on a small number of free hyperparameters (sparsity ratio, learning rate split) and on unstated assumptions about how gradient conflict is measured and how the regularization theory transfers from single-task fine-tuning to MTL. No new entities are introduced.

free parameters (2)
  • sparsity ratio (fraction of trainable encoder weights via per-neuron top-K) = 30.97% (SAM/NYU-v2), 37.60% (Swin/CelebA), 29.38% (ViT/Clevr), 62.19% (MTAN/NYU-v2), 76.02% (MTAN/CityScapes)
    Grid-searched on Joint Train (Appendix B, Table 4) and reused across gradient manipulation methods; the optimal ratio differs by model and dataset, so reported gains depend on this fitted hyperparameter.
  • learning rate split = 2e-4 for sparse conditions, 5e-5 for full-parameter conditions
    Chosen from {2e-4, 5e-5} and assigned asymmetrically to sparse vs full training, so part of the improvement may come from a more favorable LR rather than from sparsity per se.
assumptions (4)
  • standard math Lagrangian weak duality justifies the regularized objective in Eq. (12).
    The paper invokes Lagrangian duality but labels L_r an upper bound when weak duality gives the opposite direction; the math as written does not support the claimed regularization benefit.
  • domain assumption The stability and generalization benefits of parameter-efficient fine-tuning shown by Fu et al. [11] transfer to multi-task objectives with the PSN mask.
    The paper cites [11] for Eq. (12) but does not establish that the result holds for MTL losses or for the specific per-neuron magnitude mask.
  • ad hoc to paper The gradient-conflict incidence p% is measured on the same gradient representation for all compared methods.
    The paper never specifies whether p% uses raw, manipulated, or masked gradients; the reported ST reduction and the absence of reduction for manipulation methods depend on this unstated choice.
  • domain assumption Selecting the top-magnitude input connections of every neuron preserves each neuron's activation potential and is sufficient for effective MTL.
    The PSN selection rule is justified by intuition and an ablation (Fig. 5c), not by a formal analysis or a mechanistic study of MTL gradient geometry.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Proactive Gradient Conflict Mitigation in Multi-Task Learning: A Sparse Training Perspective." pith.science (2026). https://pith.science/paper/RTFNTMVH

@misc{pith2026241118615,
  author       = {Pith},
  title        = {Pith review of: Proactive Gradient Conflict Mitigation in Multi-Task Learning: A Sparse Training Perspective},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RTFNTMVH}},
  note         = {Machine review of arXiv:2411.18615}
}
read the original abstract

Advancing towards generalist agents necessitates the concurrent processing of multiple tasks using a unified model, thereby underscoring the growing significance of simultaneous model training on multiple downstream tasks. A common issue in multi-task learning is the occurrence of gradient conflict, which leads to potential competition among different tasks during joint training. This competition often results in improvements in one task at the expense of deterioration in another. Although several optimization methods have been developed to address this issue by manipulating task gradients for better task balancing, they cannot decrease the incidence of gradient conflict. In this paper, we systematically investigate the occurrence of gradient conflict across different methods and propose a strategy to reduce such conflicts through sparse training (ST), wherein only a portion of the model's parameters are updated during training while keeping the rest unchanged. Our extensive experiments demonstrate that ST effectively mitigates conflicting gradients and leads to superior performance. Furthermore, ST can be easily integrated with gradient manipulation techniques, thus enhancing their effectiveness.

Figures

Figures reproduced from arXiv: 2411.18615 by the authors.

Figure 1
Figure 1. The average occurrence percentage of gradient conflict over epochs (all epochs/last 50% epochs) during training on the SAM [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Visualization of gradients change for different methods. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. PSN. Top-1 highest-magnitude parameter among all in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: The incidence of GC between tasks during training SAM [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Ablation study for Joint Train with NYU-v2 dataset. (a) The average incidence of GC during joint training on different sizes of Swin transformers. Please see the numerical statics for all epochs in Tab. 8 in Appendix F.4. (b) The different number of trainable parameter…
Figure 6
Figure 6. Figure 6: The distribution of selected trainable parameters for different sparse training methods over different blocks. The experiments are [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The number of occurrence gradient conflictions between tasks during training SAM on NYUv2 dataset. [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: The average occurrence percentage of gradient conflict over epochs (all epochs/last 50% epochs) during training on MTAN model [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: The number of occurrence gradient conflictions between tasks during tuning MTAN on NYUv2 dataset. [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: The number of occurrence gradient conflictions between tasks during tuning Swin on CelebA dataset. [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: The average occurrence percentage of gradient conflict over epochs (all epochs/last 50% epochs) during training on Swin model [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]
Figure 12
Figure 12. Figure 12: The number of occurrence gradient conflictions between tasks during tuning MTAN on CityScapes dataset. [PITH_FULL_IMAGE:figures/full_fig_p022_12.png]
Figure 13
Figure 13. Figure 13: The average occurrence percentage of gradient conflict over epochs (all epochs/last 50% epochs) during training on MTAN [PITH_FULL_IMAGE:figures/full_fig_p022_13.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Unified Gradient Projection: Language-Balanced Continual Learning for Multilingual Low-Resource ASR

    cs.CL 2026-07 conditional novelty 5.0 of 10

    Language-balanced gradient projection plus experience replay yields near-zero average forgetting when adapting Whisper-large-v3 to low-resource languages while preserving target plasticity.

Reference graph

Works this paper leans on

40 extracted references · 25 canonical work pages · cited by 1 Pith paper

  1. [1]

    Segnet: A deep convolutional encoder-decoder architecture for image segmentation

    Vijay Badrinarayanan, Alex Kendall, and Roberto Cipolla. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pattern anal- ysis and machine intelligence , 39(12):2481–2495, 2017. 2, 5

    work page 2017

  2. [2]

    Sparse multi-task reinforcement learning

    Daniele Calandriello, Alessandro Lazaric, and Marcello Restelli. Sparse multi-task reinforcement learning. Advances in neural information processing systems, 27, 2014. 2

    work page 2014

  3. [3]

    Sam fails to segment anything? – sam- adapter: Adapting sam in underperformed scenes: Camou- flage, shadow, and more, 2023

    Tianrun Chen, Lanyun Zhu, Chaotao Ding, Runlong Cao, Shangzhan Zhang, Yan Wang, Zejian Li, Lingyun Sun, Papa Mao, and Ying Zang. Sam fails to segment anything? – sam- adapter: Adapting sam in underperformed scenes: Camou- flage, shadow, and more, 2023. 2

    work page 2023

  4. [4]

    Just pick a sign: Optimizing deep multitask models with gra- dient sign dropout

    Zhao Chen, Jiquan Ngiam, Yanping Huang, Thang Luong, Henrik Kretzschmar, Yuning Chai, and Dragomir Anguelov. Just pick a sign: Optimizing deep multitask models with gra- dient sign dropout. Advances in Neural Information Process- ing Systems, 33:2039–2050, 2020. 1, 2, 3, 5

    work page 2020

  5. [5]

    The cityscapes dataset for semantic urban scene understanding

    Marius Cordts, Mohamed Omran, Sebastian Ramos, Timo Rehfeld, Markus Enzweiler, Rodrigo Benenson, Uwe Franke, Stefan Roth, and Bernt Schiele. The cityscapes dataset for semantic urban scene understanding. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 3213–3223, 2016. 5

    2016

  6. [6]

    Indoor semantic segmentation using depth in- formation, 2013

    Camille Couprie, Cl ´ement Farabet, Laurent Najman, and Yann LeCun. Indoor semantic segmentation using depth in- formation, 2013. 5

    work page 2013

  7. [7]

    Imagenet: A large-scale hierarchical im- age database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical im- age database. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009. 5

    work page 2009

  8. [8]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 2, 5

    arXiv 2010

Show all 40 references
  1. [9]

    On interpretability of artificial neural networks: A survey

    Fenglei Fan, Jinjun Xiong, Mengzhou Li, and Ge Wang. On interpretability of artificial neural networks: A survey. IEEE Transactions on Radiation and Plasma Medical Sciences, 5: 741–760, 2020. 4

    2020

  2. [10]

    The lottery ticket hy- pothesis: Finding sparse, trainable neural networks

    Jonathan Frankle and Michael Carbin. The lottery ticket hy- pothesis: Finding sparse, trainable neural networks. arXiv preprint arXiv:1803.03635, 2018. 4

    2018 arXiv

  3. [11]

    On the effectiveness of parameter-efficient fine-tuning

    Zihao Fu, Haoran Yang, Anthony Man-Cho So, Wai Lam, Lidong Bing, and Nigel Collier. On the effectiveness of parameter-efficient fine-tuning. In Proceedings of the AAAI Conference on Artificial Intelligence , pages 12799–12807,

  4. [12]

    Learn- ing to branch for multi-task learning

    Pengsheng Guo, Chen-Yu Lee, and Daniel Ulbricht. Learn- ing to branch for multi-task learning. InInternational confer- ence on machine learning, pages 3854–3863. PMLR, 2020. 2

    2020

  5. [13]

    Learn- ing both weights and connections for efficient neural net- work

    Song Han, Jeff Pool, John Tran, and William Dally. Learn- ing both weights and connections for efficient neural net- work. Advances in neural information processing systems , 28, 2015. 4

    2015

  6. [14]

    Multi-task learning using uncertainty to weigh losses for scene geome- try and semantics

    Alex Kendall, Yarin Gal, and Roberto Cipolla. Multi-task learning using uncertainty to weigh losses for scene geome- try and semantics. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 7482–7491,

  7. [15]

    Segment any- thing

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment any- thing. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 4015–4026, 2023. 5

    2023

  8. [16]

    Block pruning for faster transformers

    Franc ¸ois Lagunas, Ella Charlaix, Victor Sanh, and Alexan- der M Rush. Block pruning for faster transformers. arXiv preprint arXiv:2109.04838, 2021. 2

    2021 arXiv

  9. [17]

    Conflict-averse gradient descent for multi-task learn- ing

    Bo Liu, Xingchao Liu, Xiaojie Jin, Peter Stone, and Qiang Liu. Conflict-averse gradient descent for multi-task learn- ing. Advances in Neural Information Processing Systems , 34:18878–18890, 2021. 1, 2, 3, 5, 8

    2021

  10. [18]

    Famo: Fast adaptive multitask optimization, 2023

    Bo Liu, Yihao Feng, Peter Stone, and Qiang Liu. Famo: Fast adaptive multitask optimization, 2023. 1, 2, 8

    2023

  11. [19]

    Towards impartial multi-task learning

    Liyang Liu, Yi Li, Zhanghui Kuang, J Xue, Yimin Chen, Wenming Yang, Qingmin Liao, and Wayne Zhang. Towards impartial multi-task learning. In International Conference on Learning Representations, 2021. 1, 2, 3, 5

    2021

  12. [20]

    End- to-end multi-task learning with attention

    Shikun Liu, Edward Johns, and Andrew J Davison. End- to-end multi-task learning with attention. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 1871–1880, 2019. 2, 5

    2019

  13. [21]

    Deep learning face attributes in the wild

    Ziwei Liu, Ping Luo, Xiaogang Wang, and Xiaoou Tang. Deep learning face attributes in the wild. In Proceedings of the IEEE international conference on computer vision, pages 3730–3738, 2015. 5

    2015

  14. [22]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, pages 10012–10022, 2021. 2, 5

    2021

  15. [23]

    Attentive single-tasking of multiple tasks

    Kevis-Kokitsi Maninis, Ilija Radosavovic, and Iasonas Kokkinos. Attentive single-tasking of multiple tasks. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1851–1860, 2019. 5

    2019

  16. [24]

    Cross-stitch networks for multi-task learning

    Ishan Misra, Abhinav Shrivastava, Abhinav Gupta, and Mar- tial Hebert. Cross-stitch networks for multi-task learning. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 3994–4003, 2016. 2

    2016

  17. [25]

    Parameter efficient train- ing of deep convolutional neural networks by dynamic sparse reparameterization

    Hesham Mostafa and Xin Wang. Parameter efficient train- ing of deep convolutional neural networks by dynamic sparse reparameterization. In International Conference on Machine Learning, pages 4646–4655. PMLR, 2019. 2, 4

    2019

  18. [26]

    Multi- task learning as a bargaining game, 2022

    Aviv Navon, Aviv Shamsian, Idan Achituve, Haggai Maron, Kenji Kawaguchi, Gal Chechik, and Ethan Fetaya. Multi- task learning as a bargaining game, 2022. 1, 2, 3, 5

    2022

  19. [27]

    Movement pruning: Adaptive sparsity by fine-tuning

    Victor Sanh, Thomas Wolf, and Alexander Rush. Movement pruning: Adaptive sparsity by fine-tuning. Advances in neu- ral information processing systems, 33:20378–20389, 2020. 2, 4 9

    2020

  20. [28]

    Multi-task learning as multi-objective optimization

    Ozan Sener and Vladlen Koltun. Multi-task learning as multi-objective optimization. Advances in Neural Informa- tion Processing Systems, 31, 2018. 1, 2, 3, 5

    2018

  21. [29]

    Recon: Reducing conflicting gradi- ents from the root for multi-task learning

    Guangyuan Shi, Qimai Li, Wenlong Zhang, Jiaxin Chen, and Xiao-Ming Wu. Recon: Reducing conflicting gradi- ents from the root for multi-task learning. arXiv preprint arXiv:2302.11289, 2023. 1, 6

    2023 arXiv

  22. [30]

    Learning sparse sharing architectures for multiple tasks

    Tianxiang Sun, Yunfan Shao, Xiaonan Li, Pengfei Liu, Hang Yan, Xipeng Qiu, and Xuanjing Huang. Learning sparse sharing architectures for multiple tasks. In Proceedings of the AAAI conference on artificial intelligence , pages 8936– 8943, 2020. 2

    2020

  23. [31]

    Multi-task learning for dense prediction tasks: A survey

    Simon Vandenhende, Stamatios Georgoulis, Wouter Van Gansbeke, Marc Proesmans, Dengxin Dai, and Luc Van Gool. Multi-task learning for dense prediction tasks: A survey. IEEE transactions on pattern analysis and machine intelligence, 44(7):3614–3633, 2021. 2

    2021

  24. [32]

    Pac-bayes information bottleneck

    Zifeng Wang, Shao-Lun Huang, Ercan E Kuruoglu, Jimeng Sun, Xi Chen, and Yefeng Zheng. Pac-bayes information bottleneck. arXiv preprint arXiv:2109.14509, 2021. 4

    2021 arXiv

  25. [33]

    Pad-net: Multi-tasks guided prediction-and-distillation net- work for simultaneous depth estimation and scene parsing

    Dan Xu, Wanli Ouyang, Xiaogang Wang, and Nicu Sebe. Pad-net: Multi-tasks guided prediction-and-distillation net- work for simultaneous depth estimation and scene parsing. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 675–684, 2018. 2

    2018

  26. [34]

    Raise a child in large language model: Towards effective and generalizable fine-tuning

    Runxin Xu, Fuli Luo, Zhiyuan Zhang, Chuanqi Tan, Baobao Chang, Songfang Huang, and Fei Huang. Raise a child in large language model: Towards effective and generalizable fine-tuning. arXiv preprint arXiv:2109.05687, 2021. 2, 4

    2021 arXiv

  27. [35]

    Gradient surgery for multi-task learning

    Tianhe Yu, Saurabh Kumar, Abhishek Gupta, Sergey Levine, Karol Hausman, and Chelsea Finn. Gradient surgery for multi-task learning. Advances in Neural Information Pro- cessing Systems, 33:5824–5836, 2020. 1, 2, 3, 5, 8

    2020

  28. [36]

    A large-scale study of representation learning with the visual task adaptation benchmark

    Xiaohua Zhai, Joan Puigcerver, Alexander Kolesnikov, Pierre Ruyssen, Carlos Riquelme, Mario Lucic, Josip Djo- longa, Andre Susano Pinto, Maxim Neumann, Alexey Doso- vitskiy, et al. A large-scale study of representation learning with the visual task adaptation benchmark. arXiv ...

    1910 arXiv

  29. [37]

    A survey on multi-task learning

    Yu Zhang and Qiang Yang. A survey on multi-task learning. IEEE Transactions on Knowledge and Data Engineering, 34 (12):5586–5609, 2021. 1, 2

    2021

  30. [38]

    Joint task-recursive learning for semantic segmentation and depth estimation

    Zhenyu Zhang, Zhen Cui, Chunyan Xu, Zequn Jie, Xiang Li, and Jian Yang. Joint task-recursive learning for semantic segmentation and depth estimation. In Proceedings of the European Conference on Computer Vision (ECCV) , pages 235–251, 2018. 2

    2018

  31. [39]

    Pattern-affinitive propagation across depth, surface normal and semantic segmentation

    Zhenyu Zhang, Zhen Cui, Chunyan Xu, Yan Yan, Nicu Sebe, and Jian Yang. Pattern-affinitive propagation across depth, surface normal and semantic segmentation. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4106–4115, 2019. 2

    2019

  32. [40]

    Gradient- based parameter selection for efficient fine-tuning

    Zhi Zhang, Qizhe Zhang, Zijun Gao, Renrui Zhang, Ekate- rina Shutova, Shiji Zhou, and Shanghang Zhang. Gradient- based parameter selection for efficient fine-tuning. arXiv preprint arXiv:2312.10136, 2023. 2, 4 10 Proactive Gradient Conflict Mitigation in Multi-Task Learning: A...

    2023 arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.