Abstract: Modern deep transfer learning approaches have mainly focused on learning generic feature vectors from one task that are transferable to other tasks, such as word embeddings in language and pre-trained convolutional features in vision. However, these approaches usually transfer unary features and largely ignore more structured graphical representations. This work explores the possibility of learning generic latent relational graphs that capture dependencies between pairs of data units (e.g., words or pixels) from large-scale unlabeled data and transferring the graphs to downstream tasks. Our proposed transfer learning framework improves performance on various tasks including question answering, natural language inference, sentiment analysis, and image classification. We also show that the learned graphs are generic enough to be transferred to different embeddings on which the graphs have not been trained (including GloVe embeddings, ELMo embeddings, and task-specific RNN hidden units), or embedding-free units such as image pixels.
现代深度迁移学习方法主要侧重于从一个任务中学习通用特征向量，这些向量可迁移到其他任务，例如语言中的词嵌入和视觉中预先训练的卷积特征。但是，这些方法通常迁移的是一元特征，并且在很大程忽略了更加结构化的图表示。该工作探索了学习通用隐含关系图的可能性，该关系图捕获来自大规模未标记数据的数据单元对（例如，单词或像素）之间的依赖性并将关系图迁移到下游任务。我们提出的迁移学习框架可以提高各种任务的性能，包括问答，自然语言推理，情感分析和图像分类。我们还表明，学到的图足够通用，可以迁移到图尚未训练的不同嵌入（包括GloVe嵌入，ELMo嵌入和任务特定的RNN隐藏单元），或无嵌入的单元，如图像像素。