Baseline Needs More Love: On Simple Word-Embedding-Based Models and Associated Pooling Mechanisms

ACL 2018 Dinghan ShenGuoyin WangWenlin WangMartin Renqiang MinQinliang SuYizhe ZhangChunyuan LiRicardo HenaoLawrence Carin

Many deep learning architectures have been proposed to model the compositionality in text sequences, requiring a substantial number of parameters and expensive computations. However, there has not been a rigorous evaluation regarding the added value of sophisticated compositional functions... (read more)

PDF Abstract

Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
Text Classification AG News SWEM-concat Error 7.34 # 8
Named Entity Recognition CoNLL 2003 (English) SWEM-CRF F1 86.28 # 20
Text Classification DBpedia SWEM-concat Error 1.43 # 16
Sentiment Analysis MR SWEM-concat Accuracy 78.2 # 7
Sentiment Analysis SST-2 Binary classification SWEM-concat Accuracy 84.3 # 21
Sentiment Analysis SST-5 Fine-grained classification SWEM-concat Accuracy 46.1 # 12
Subjectivity Analysis SUBJ SWEM-concat Accuracy 93.0 # 7
Text Classification TREC-6 SWEM-aver Error 7.8 # 9
Text Classification Yahoo! Answers SWEM-concat Accuracy 73.53 # 4
Sentiment Analysis Yelp Binary classification SWEM-hier Error 4.19 # 11
Sentiment Analysis Yelp Fine-grained classification SWEM-hier Error 36.21 # 11