{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,19]],"date-time":"2026-07-19T02:51:06Z","timestamp":1784429466484,"version":"3.55.0"},"reference-count":68,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.00326","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"3413-3423","source":"Crossref","is-referenced-by-count":1,"title":["Rep-MTL: Unleashing the Power of Representation-Level Task Saliency for Multi-Task Learning"],"prefix":"10.1109","author":[{"given":"Zedong","family":"Wang","sequence":"first","affiliation":[{"name":"The Hong Kong University of Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Siyuan","family":"Li","sequence":"additional","affiliation":[{"name":"Zhejiang University"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dan","family":"Xu","sequence":"additional","affiliation":[{"name":"The Hong Kong University of Science and Technology"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3534678.3539442"},{"key":"ref2","article-title":"Fair resource allocation in multi-task learning","volume-title":"Forty-first International Conference on Machine Learning","author":"Ban","year":"2024"},{"key":"ref3","article-title":"Sparks of artificial general intelligence: Early experiments with gpt-4","author":"Bubeck","year":"2023","journal-title":"arXiv preprint"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1023\/A:1007379606734"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1802.02611"},{"key":"ref6","first-page":"1597","article-title":"A simple framework for contrastive learning of visual representations","volume-title":"International conference on machine learning","author":"Chen","year":"2020"},{"key":"ref7","first-page":"22243","article-title":"Big self-supervised models are strong semi-supervised learners","volume":"33","author":"Chen","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref8","article-title":"GradNorm: Gradient normalization for adaptive loss balancing in deep multitask networks","volume-title":"International Conference on Machine Learning","author":"Chen","year":"2018"},{"key":"ref9","article-title":"Just pick a sign: Optimizing deep multitask models with gradient sign dropout","author":"Chen","year":"2020","journal-title":"Neural Information Processing Systems"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2019.00159"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.350"},{"key":"ref12","first-page":"496","article-title":"Improvable gap balancing for multi-task learning","author":"Dai","year":"2023","journal-title":"Uncertainty in Artificial Intelligence"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1016\/j.crma.2012.03.014"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/iccv.2017.226"},{"key":"ref15","article-title":"Mitigating gradient bias in multi-objective learning: A provably convergent approach","volume-title":"International Conference on Learning Representations","author":"Devaka Fernando","year":"2023"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.5555\/3495724.3497510"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00975"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1145\/3485447.3512093"},{"key":"ref20","article-title":"Position: The platonic representation hypothesis","volume-title":"Forty-first International Conference on Machine Learning","author":"Huh","year":"2024"},{"key":"ref21","article-title":"Rotograd: Gradient homogenization in multitask learning","volume-title":"International Conference on Learning Representations","author":"Javaloy","year":"2021"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00781"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00371"},{"key":"ref24","first-page":"12169","volume":"35","author":"Kurin","year":"2022","journal-title":"defense of the unitary scalarization for deep multi-task learning"},{"key":"ref25","article-title":"Unveiling the backbone-optimizer coupling bias in visual representation learning","author":"Li","year":"2024","journal-title":"arXiv preprint"},{"key":"ref26","author":"Li","year":"2024","journal-title":"Openmixup: Open mixup toolbox and benchmark for visual representation learning"},{"issue":"1","key":"ref27","first-page":"9999","article-title":"Libmtl: A python library for deep multi-task learning","volume":"24","author":"Lin","year":"2023","journal-title":"The Journal of Machine Learning Research"},{"key":"ref28","article-title":"Reasonable effectiveness of random weighting: A litmus test for multi-task learning","author":"Lin","year":"2022","journal-title":"Transactions on Machine Learning Research"},{"key":"ref29","article-title":"Reasonable effectiveness of random weighting: A litmus test for multi-task learning","author":"Lin","year":"2022","journal-title":"Transactions on Machine Learning Research"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2025.108317"},{"key":"ref31","article-title":"Smooth tchebycheff scalarization for multi-objective optimization","volume-title":"Forty-first International Conference on Machine Learning","author":"Lin","year":"2024"},{"key":"ref32","article-title":"Conflict-averse gradient descent for multi-task learning","author":"Liu","year":"2021","journal-title":"Neural Information Processing Systems"},{"key":"ref33","article-title":"Famo: Fast adaptive multitask optimization","author":"Liu","year":"2024","journal-title":"Advances in Neural Information Processing Systems, 36"},{"key":"ref34","article-title":"Towards impartial multi-task learning","volume-title":"International Conference on Learning Representations","author":"Liu","year":"2021"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00197"},{"key":"ref36","article-title":"Auto-lambda: Disentangling dynamic task relationships","author":"Liu","year":"2022","journal-title":"Transactions on Machine Learning Research"},{"key":"ref37","article-title":"Unified-io: A unified model for vision, language, and multi-modal tasks","volume-title":"The Eleventh International Conference on Learning Representations","author":"Lu","year":"2022"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02497"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1145\/3219819.3220007"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1145\/3209978.3210104"},{"key":"ref41","article-title":"Traditional and heavy tailed self regularization in neural network models","volume-title":"International Conference on Machine Learning","author":"Mahoney","year":"2019"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.837"},{"key":"ref43","article-title":"Robust analysis of multi-task learning on a complex vision system","author":"Mao","year":"2024","journal-title":"arXiv preprint"},{"issue":"165","key":"ref44","first-page":"1","article-title":"Implicit selfregularization in deep neural networks: Evidence from random matrix theory and implications for learning","volume":"22","author":"Martin","year":"2021","journal-title":"Journal of Machine Learning Research"},{"issue":"165","key":"ref45","first-page":"1","article-title":"Implicit selfregularization in deep neural networks: Evidence from random matrix theory and implications for learning","volume":"22","author":"Martin","year":"2021","journal-title":"Journal of Machine Learning Research"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-021-24025-8"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.433"},{"key":"ref48","article-title":"Can optimization trajectories explain multi-task transfer?","author":"Mueller","year":"2024","journal-title":"arXiv preprint"},{"key":"ref49","first-page":"16428","article-title":"Multitask learning as a bargaining game","volume-title":"International Conference on Machine Learning","author":"Navon","year":"2022"},{"issue":"8","key":"ref50","article-title":"Language models are unsupervised multitask learners","volume":"1","author":"Radford","year":"2019","journal-title":"OpenAI blog"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33014822"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-15561-1_16"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.74"},{"key":"ref54","article-title":"Multi-task learning as multi-objective optimization","author":"Sener","year":"2018","journal-title":"Advances in neural information processing systems, 31"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01923"},{"key":"ref56","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","author":"Shazeer","year":"2017","journal-title":"arXiv preprint"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.52202\/079017-3537"},{"key":"ref58","article-title":"Recon: Reducing conflicting gradients from the root for multi-task learning","volume":"abs\/2302.11289","author":"Shi","year":"2023","journal-title":"ArXiv"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-33715-4_54"},{"key":"ref60","first-page":"9120","article-title":"Which tasks should be learned together in multi-task learning?","volume-title":"International conference on machine learning","author":"Standley","year":"2020"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.1145\/3383313.3412236"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.572"},{"key":"ref63","article-title":"Gradient vaccine: Investigating and improving multi-task optimization in massively multilingual models","volume-title":"International Conference on Learning Representations","author":"Wang","year":"2021"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0988"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i4.20352"},{"key":"ref66","first-page":"5824","article-title":"Gradient surgery for multi-task learning","volume":"33","author":"Yu","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/JAS.2022.106004"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2021.3070203"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11444514.pdf?arnumber=11444514","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T06:17:31Z","timestamp":1777529851000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11444514\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":68,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.00326","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}