Supervised Learning of Universal Sentence Representations from Natural Language Inference Data
This work addresses the need for better sentence embeddings in NLP, offering a supervised approach that could serve as a foundational resource similar to ImageNet in computer vision, though it is incremental in leveraging existing inference data.
The paper tackles the problem of learning effective sentence representations by using supervised data from the Stanford Natural Language Inference datasets, resulting in universal sentence representations that consistently outperform unsupervised methods like SkipThought vectors on various transfer tasks.
Many modern NLP systems rely on word embeddings, previously trained in an unsupervised manner on large corpora, as base features. Efforts to obtain embeddings for larger chunks of text, such as sentences, have however not been so successful. Several attempts at learning unsupervised representations of sentences have not reached satisfactory enough performance to be widely adopted. In this paper, we show how universal sentence representations trained using the supervised data of the Stanford Natural Language Inference datasets can consistently outperform unsupervised methods like SkipThought vectors on a wide range of transfer tasks. Much like how computer vision uses ImageNet to obtain features, which can then be transferred to other tasks, our work tends to indicate the suitability of natural language inference for transfer learning to other NLP tasks. Our encoder is publicly available.