An Empirical Evaluation of doc2vec with Practical Insights into Document Embedding Generation

TL;DR

This paper provides a comprehensive empirical evaluation of doc2vec for document embedding, highlighting its robustness with large corpora and pre-trained embeddings, along with practical hyper-parameter recommendations.

Contribution

It offers the first thorough empirical analysis of doc2vec's performance across multiple tasks and compares it with other embedding methods, including practical guidelines.

Findings

doc2vec performs well with large external corpora

Pre-trained word embeddings enhance doc2vec performance

Provides hyper-parameter recommendations for general use

Abstract

Recently, Le and Mikolov (2014) proposed doc2vec as an extension to word2vec (Mikolov et al., 2013a) to learn document-level embeddings. Despite promising results in the original paper, others have struggled to reproduce those results. This paper presents a rigorous empirical evaluation of doc2vec over two tasks. We compare doc2vec to two baselines and two state-of-the-art document embedding methodologies. We found that doc2vec performs robustly when using models trained on large external corpora, and can be further improved by using pre-trained word embeddings. We also provide recommendations on hyper-parameter settings for general purpose applications, and release source code to induce document embeddings using our trained doc2vec models.