fufrank5 / relatedness

Implementation for the relatedness paper.

relatedness

We provide the implementation for Learning Relatedness between Types with Prototypes for Relation Extraction

Environment

python==2.7 tensorflow==1.4

Dataset

ACE: ACE2005 (https://catalog.ldc.upenn.edu/LDC2006T06)

ERE: LDC2015E29, LDC2015E68, LDC2015E78, LDC2015R26, LDC2016E31, LDC2016E73

Both datasets are available through LDC.

We processed the LDC data and transformed them into intermediate files. An example of the format is located at data/ERE/ere_directional_example.txt. These are the input files to our implementation.

Implementation

The model implementation is available in this code repository. It also contains redundant codes that were used for various experiments, but not necessarily for this paper.

An example command: python multi.py --expt=ace05_ere --joint=joint_ksupport --debug

Note: the input processed data files are not provided in the repository.