https://github.com/disi-unibo-nlp/unown

https://github.com/disi-unibo-nlp/unown

Science Score: 13.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (7.8%) to scientific vocabulary
Last synced: 5 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: disi-unibo-nlp
  • Language: Jupyter Notebook
  • Default Branch: main
  • Size: 22.9 MB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed over 1 year ago
Metadata Files
Readme

README.md

unown

SciFact

You need Multivers in order to run inference on the generated code, as well as generating claims

You can perform

git clone https://github.com/dwadden/multivers.git

Once downloaded, you should run

bash script/getdatatrain.sh

In order to get the training/test data and corpus from SciFact necessary to run our generation

You should replace basepathmultivers in scifact/claimgen.ipynb by the actual path of your multivers

Finally, replace base_path value by your working directory

Feverous

You need the feverous code to run our experiments https://github.com/Raldir/FEVEROUS.git

You will also need the feverous feverouswikiv1.db dataset from https://fever.ai/download/feverous/feverous-wiki-pages-db.zip

claimgen.ipynb allows you to construct new examples

entityreplacement.ipynb permits to replace the negative of a generated/original set of claims

runPipeline.ipynb serves to run the pipeline of Feverous in a way that uses only the predictor part, and not the retriever. It will train and test it. The test is run on original dataset. It gives an output file of predicted labels and another one containing various metrics including th accuracy on the fact-checking pipeline

Owner

  • Name: DISI UniBo NLP
  • Login: disi-unibo-nlp
  • Kind: user
  • Location: Italy

NLU Research Group @ University of Bologna @ Department of Computer Science and Engineering (DISI)

GitHub Events

Total
  • Push event: 1
Last Year
  • Push event: 1