Amobee at SemEval-2018 Task 1: GRU Neural Network with a CNN Attention Mechanism for Sentiment Classification

TL;DR

This paper presents a sentiment analysis system using a GRU with CNN attention, achieving top results in SemEval 2018's English and Spanish sub-tasks.

Contribution

It introduces a novel combination of GRU and CNN attention mechanisms for sentiment classification and employs task-specific embeddings and ensemble methods.

Findings

Achieved 1st place in Spanish valence sub-task

Achieved 3rd place in English valence ordinal classification

Demonstrated effectiveness of CNN attention in sentiment analysis

Abstract

This paper describes the participation of Amobee in the shared sentiment analysis task at SemEval 2018. We participated in all the English sub-tasks and the Spanish valence tasks. Our system consists of three parts: training task-specific word embeddings, training a model consisting of gated-recurrent-units (GRU) with a convolution neural network (CNN) attention mechanism and training stacking-based ensembles for each of the sub-tasks. Our algorithm reached 3rd and 1st places in the valence ordinal classification sub-tasks in English and Spanish, respectively.