通过调整令牌空间解决任务间梯度冲突，提升Transformer多任务学习效果
Resolving Token-Space Gradient Conflicts: Token Space Manipulation for Transformer-Based Multi-Task Learning
- 在令牌空间识别并分类梯度冲突，按类型自适应调节
- 实验显示性能提升显著，计算开销极低
- 适合需要高效多任务学习的场景，如大规模语言模型
多任务学习（MTL）使多个任务在共享网络中同时学习，但不同任务目标差异会导致负迁移，即一个任务的学习损害另一任务表现。尽管预训练Transformer显著提升了MTL性能，其固定的网络容量和僵化结构限制了适应性。此前动态网络架构尝试解决此问题，但效率低下，因直接将共享参数转化为特定任务参数。我们提出动态令牌调制与扩展（DTME-MTL），适用于任何基于Transformer的MTL架构。该方法通过识别令牌空间中的梯度冲突，并根据冲突类型应用自适应解决方案，增强模型适应性并减少过拟合。不同于以往通过复制参数缓解负迁移的方法，DTME-MTL完全在令牌空间操作，实现高效适应且无需大量参数增长。大量实验表明，DTME-MTL在保持极小计算开销的同时持续提升多任务性能，为增强Transformer-based MTL模型提供可扩展、有效的解决方案。
原文摘要 · Abstract (English)
Multi-Task Learning (MTL) enables multiple tasks to be learned within a shared network, but differences in objectives across tasks can cause negative transfer, where the learning of one task degrades another task's performance. While pre-trained transformers significantly improve MTL performance, their fixed network capacity and rigid structure limit adaptability. Previous dynamic network architectures attempt to address this but are inefficient as they directly convert shared parameters into task-specific ones. We propose Dynamic Token Modulation and Expansion (DTME-MTL), a framework applicable to any transformer-based MTL architecture. DTME-MTL enhances adaptability and reduces overfitting by identifying gradient conflicts in token space and applying adaptive solutions based on conflict type. Unlike prior methods that mitigate negative transfer by duplicating network parameters, DTME-MTL operates entirely in token space, enabling efficient adaptation without excessive parameter growth. Extensive experiments demonstrate that DTME-MTL consistently improves multi-task performance with minimal computational overhead, offering a scalable and effective solution for enhancing transformer-based MTL models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。