HannaAbiAkl / PSYCHIC

The official repository for the PSYCHIC model

PSYCHIC

PSYCHIC (Pre-trained SYmbolic CHecker In Context) is a model that is a fine-tuned version of distilbert-base-uncased on the DBLP-QuAD dataset.

This is the official code repository for the Scholarly-QALD challenge paper at the International Semantic Web Conference (ISWC) 2023: "PSYCHIC: A Neuro-Symbolic Framework for Knowledge Graph Question-Answering Grounding".

Data

This directory contains two data files:

• dev data: The data provided by the task organizers for the dev phase.
• final data: The data provided by the task organizers for the final phase.

Experiment

This directory contains the code for training and evaluating PSYCHIC on the dev and final data. The code also contains experiments with a generative T5 model.

Figures

This directory contains miscellaneous figures for the repository.

Results

This directory contains the official results submitted to the challenge.