From 1c38b5437bc8ffd7a58610caeb3958412528a1b6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Lo=C3=AFc=20Vial?= Date: Mon, 14 Oct 2019 11:35:43 +0200 Subject: [PATCH] Update README.md --- README.md | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/README.md b/README.md index da5b440..53132bc 100644 --- a/README.md +++ b/README.md @@ -32,6 +32,16 @@ To install **UFSAC**, simply: Once the dependencies are installed, please run `./java/compile.sh` to compile the Java code. +## Sense mappings + +We provide the two sense mappings used in our paper as standalone files in the directory `sense_mappings`. + +The files consist of 117659 lines (one line by synset): the left-hand ID is the original synset ID, and the right-hand is the ID of the associated group of synsets. + +The file `hypernyms_mapping.txt` results from the sense compression method through hypernyms. + +The file `all_relations_mapping.txt` results from the method through all relationships. + ## Using pre-trained models We are currently providing one of our best model trained on the SemCor and the WordNet Gloss Tagged, using BERT embeddings, with the vocabulary compression through the hypernymy/hyponymy relationships applied, as described in [our article](https://arxiv.org/abs/1905.05677).