Neural DrugNet

TL;DR

Neural DrugNet is an ensemble system combining pretrained word-based LSTMs and character tri-gram LSTMs with attention, achieving high accuracy in classifying social media posts about medication intake.

Contribution

The paper introduces Neural DrugNet, a novel ensemble of two LSTM models with different input representations and attention, for improved medication-related social media classification.

Findings

Ranked 2nd in the shared task on medication intake classification

Effective combination of pretrained word models and character tri-gram models

Demonstrates the utility of ensemble methods in health-related social media analysis

Abstract

In this paper, we describe the system submitted for the shared task on Social Media Mining for Health Applications by the team Light. Previous works demonstrate that LSTMs have achieved remarkable performance in natural language processing tasks. We deploy an ensemble of two LSTM models. The first one is a pretrained language model appended with a classifier and takes words as input, while the second one is a LSTM model with an attention unit over it which takes character tri-gram as input. We call the ensemble of these two models: Neural-DrugNet. Our system ranks 2nd in the second shared task: Automatic classification of posts describing medication intake.