Chainer implementation of Commonsense Knowledge Base Completion

Chainer-based implementation of Commonsense Knowledge Base Completion.

See "Commonsense Knowledge Base Completion", Xiang Li; Aynaz Taheri; Lifu Tu; Kevin Gimpel, ACL 2016.

Development Environment

• Ubuntu 16.04
• Python 3.5.2
• Chainer 3.1.0
• numpy 1.13.3
• cupy 2.1.0
• nltk
• progressbar
• and their dependencies

Dataset

Dataset is derived from the original implementation. Now you have seven files:

• new_omcs100.txt: 100k facts which are used for training.
• new_omcs300.txt: 300k facts which are used for training.
• new_omcs600.txt: 600k facts which are used for training.
• new_omcs_dev1.txt: 1.2k facts which are used for learning a threshold to classify true facts and false facts.
• new_omcs_dev2.txt: 1.2k facts which are used for measuring generalization ability during training.
• new_omcs_test.txt: 2.4k facts which are used for measuring generalization ability.
• rel.txt: A list of vocabulary of relations.

How to Run

First, you need to prepare vocabulary of concepts.

$ python preprocess.py <path/to/training-data> --n-vocab <number of vocabulary>

Then, let's start training.

$ python train.py <path/to/training-data> <path/to/concept-vocabulary> <path/to/relation-vocabulary> --validation1 <path/to/validation-data1> --validation2 <path/to/validation-data2> --embedding <path/to/pretrained-word-embedding> -g <gpu-id>

See command line help for other options.