Strawman: an Ensemble of Deep Bag-of-Ngrams for Sentiment Analysis

TL;DR

Strawman is an ensemble of deep bag-of-ngrams designed for sentence-level sentiment analysis, serving as a baseline for testing the robustness of NLP systems in the Build It, Break It shared task.

Contribution

It introduces 'strawman', a novel ensemble method using deep bag-of-ngrams for sentiment analysis, tailored for robustness testing in NLP.

Findings

Provides a strong baseline for sentiment analysis

Facilitates identification of minimal sentence differences that break the analyzer

Enhances understanding of model robustness in NLP systems

Abstract

This paper describes a builder entry, named "strawman", to the sentence-level sentiment analysis task of the "Build It, Break It" shared task of the First Workshop on Building Linguistically Generalizable NLP Systems. The goal of a builder is to provide an automated sentiment analyzer that would serve as a target for breakers whose goal is to find pairs of minimally-differing sentences that break the analyzer.