A Sensitivity Analysis of (and Practitioners' Guide to) Convolutional Neural Networks for Sentence Classification

TLDR

Convolutional neural networks achieve strong performance on sentence classification but require careful specification of architecture and hyperparameters. The study investigates how sensitive sentence‑classification performance is to changes in CNN architecture and hyperparameter settings. We conduct a sensitivity analysis of one‑layer CNNs, varying filter sizes, regularization, and other design choices to identify which components most affect performance. Our extensive experiments yield practical guidelines that help practitioners optimize one‑layer CNNs for real‑world sentence‑classification tasks.

Abstract

Convolutional Neural Networks (CNNs) have recently achieved remarkably strong performance on the practically important task of sentence classification (kim 2014, kalchbrenner 2014, johnson 2014). However, these models require practitioners to specify an exact model architecture and set accompanying hyperparameters, including the filter region size, regularization parameters, and so on. It is currently unknown how sensitive model performance is to changes in these configurations for the task of sentence classification. We thus conduct a sensitivity analysis of one-layer CNNs to explore the effect of architecture components on model performance; our aim is to distinguish between important and comparatively inconsequential design decisions for sentence classification. We focus on one-layer CNNs (to the exclusion of more complex models) due to their comparative simplicity and strong empirical performance, which makes it a modern standard baseline method akin to Support Vector Machine (SVMs) and logistic regression. We derive practical advice from our extensive empirical results for those interested in getting the most out of CNNs for sentence classification in real world settings.