Learning Distributed Representations of Sentences from Unlabelled Data

NAACL 2016  ·  Felix Hill, Kyunghyun Cho, Anna Korhonen ·

Unsupervised methods for learning distributed representations of words are ubiquitous in today's NLP research, but far less is known about the best ways to learn distributed phrase or sentence representations from unlabelled data. This paper is a systematic comparison of models that learn such representations. We find that the optimal approach depends critically on the intended application. Deeper, more complex models are preferable for representations to be used in supervised systems, but shallow log-linear models work best for building representation spaces that can be decoded with simple spatial distance metrics. We also propose two new unsupervised representation-learning objectives designed to optimise the trade-off between training time, domain portability and performance.

PDF Abstract NAACL 2016 PDF NAACL 2016 Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Subjectivity Analysis SUBJ SDAE Accuracy 90.8 # 16

Methods


No methods listed for this paper. Add relevant methods here