ChatPaper.aiChatPaper

Rep-MTL:釋放表徵層級任務顯著性在多任務學習中的潛力

Rep-MTL: Unleashing the Power of Representation-level Task Saliency for Multi-Task Learning

July 28, 2025
作者: Zedong Wang, Siyuan Li, Dan Xu
cs.AI

摘要

儘管多任務學習在利用跨任務的互補知識方面展現了巨大潛力,現有的多任務優化(MTO)技術仍專注於通過以優化器為中心的損失縮放和梯度操控策略來解決衝突,卻未能帶來一致的性能提升。本文主張,在任務交互自然發生的共享表示空間中,蘊含著豐富的信息和潛力,可進行與現有優化器互補的操作,特別是在促進任務間互補性方面,這在MTO中鮮有探索。這一洞見催生了Rep-MTL,它利用表示層面的任務顯著性來量化任務特定優化與共享表示學習之間的交互。通過基於熵的懲罰和樣本級跨任務對齊來引導這些顯著性,Rep-MTL旨在通過維持個別任務的有效訓練而非單純解決衝突來減輕負遷移,同時明確促進互補信息的共享。實驗在涵蓋任務偏移和領域偏移場景的四個具有挑戰性的MTL基準上進行。結果表明,即使搭配基本的等權重策略,Rep-MTL也能實現競爭性的性能提升,並具有良好的效率。除了標準的性能指標外,冪律指數分析證明了Rep-MTL在平衡任務特定學習與跨任務共享方面的有效性。項目頁面可於此處訪問。
English
Despite the promise of Multi-Task Learning in leveraging complementary knowledge across tasks, existing multi-task optimization (MTO) techniques remain fixated on resolving conflicts via optimizer-centric loss scaling and gradient manipulation strategies, yet fail to deliver consistent gains. In this paper, we argue that the shared representation space, where task interactions naturally occur, offers rich information and potential for operations complementary to existing optimizers, especially for facilitating the inter-task complementarity, which is rarely explored in MTO. This intuition leads to Rep-MTL, which exploits the representation-level task saliency to quantify interactions between task-specific optimization and shared representation learning. By steering these saliencies through entropy-based penalization and sample-wise cross-task alignment, Rep-MTL aims to mitigate negative transfer by maintaining the effective training of individual tasks instead pure conflict-solving, while explicitly promoting complementary information sharing. Experiments are conducted on four challenging MTL benchmarks covering both task-shift and domain-shift scenarios. The results show that Rep-MTL, even paired with the basic equal weighting policy, achieves competitive performance gains with favorable efficiency. Beyond standard performance metrics, Power Law exponent analysis demonstrates Rep-MTL's efficacy in balancing task-specific learning and cross-task sharing. The project page is available at HERE.
PDF374July 29, 2025