Introduction

This is an update to the implementation of EntNet done by jimfleming to let it take Children Book Test as input.

Recurrent Entity Networks

This repository contains an independent TensorFlow implementation of recurrent entity networks from Tracking the World State with Recurrent Entity Networks. This paper introduces the first method to solve all of the bAbI tasks using 10k training examples. The author's original Torch implementation is now available here.

Results

Accuracy of EntNet when run on CBT according to the paper. Actual results with this repo will be posted later.

Model	Named Entities	Common Nouns
EntNet (general)	0.484	0.540
EntNet (simple)	0.616	0.588

Setup

1. Download the dataset CBTest.tgz from here and extract it to a folder called CBT.
2. Run prep_data.py which will convert the datasets into TFRecords.
3. Run python -m entity_networks.main adding the required arguments. --help will show them.

Major Dependencies

• TensorFlow v1.4.0

(For additional dependencies see requirements.txt)