NLP with cognitive data

This repository contains all code used in the experiments described in the following paper:

Nora Hollenstein, Maria Barrett, Marius Troendle, Francesco Bigiolli, Nicolas Langer & Ce Zhang. Advancing NLP with Cognitive Language Processing Data. 2019. https://arxiv.org/abs/1904.02682

Data

The original Zurich Cognitive Language Processing Corpus (ZuCo) corpus can be found here: https://osf.io/q3zws/

Named Entity Recognition

The named entity recognition system enhanced with eye-tracking and EEG features is in named-entity-recognition/. This code contains the experiments described in Section 4.1 of the paper.

Relation Classification

The code in relation-classification/ contains the experiments described in Section 4.2 of the paper.

Sentiment Analysis

The code in sentiment-analysis/ contains the experiments described in Section 4.3 of the paper. The configuration can be set to binary or ternary classification.

Multi-task learning

The code in multi-task-learning/ contains the experiments described in Section 6 of the paper.