arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.00944cs.LGcs.CL
收藏

多任务学习中信息迁移的理解与改善

Understanding and Improving Information Transfer in Multi-Task Learning

Stanford University(斯坦福大学);University of Pennsylvania(宾夕法尼亚大学)

Sen Wu, Hongyang R. Zhang, Christopher Ré

摘要:

我们研究多任务学习方法,使用共享的功能表示的所有任务。为了更好地理解任务信息的传输,我们研究了一种架构,该架构具有所有任务的共享模块和每个任务的单独输出模块。我们在线性和ReLU激活模型上研究了这种设置的理论。我们的主要观察是,任务的数据是否对齐可以显着影响多任务学习的性能。我们发现,任务数据之间的不对齐会导致负迁移(或损害性能),并提供了充分的条件,为正迁移。受理论见解的启发,我们表明,对齐任务的嵌入层可以在GLUE基准测试和情感分析任务上实现多任务训练和迁移学习的性能提升;例如,使用我们的对齐方法,5个GLUE任务的GLUE得分平均提高了2.35%。我们还设计了一个基于SVD的任务权重调整方案,并表明它提高了多标签图像数据集上多任务训练的鲁棒性。

英文摘要:

We investigate multi-task learning approaches that use a shared feature representation for all tasks. To better understand the transfer of task information, we study an architecture with a shared module for all tasks and a separate output module for each task. We study the theory of this setting on linear and ReLU-activated models. Our key observation is that whether or not tasks' data are well-aligned can significantly affect the performance of multi-task learning. We show that misalignment between task data can cause negative transfer (or hurt performance) and provide sufficient conditions for positive transfer. Inspired by the theoretical insights, we show that aligning tasks' embedding layers leads to performance gains for multi-task training and transfer learning on the GLUE benchmark and sentiment analysis tasks; for example, we obtain a 2.35% GLUE score average improvement on 5 GLUE tasks over BERT-LARGE using our alignment method. We also design an SVD-based task reweighting scheme and show that it improves the robustness of multi-task training on a multi-label image dataset.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑