DBpedia eXplainable Chatbot (DBpedia XChat)
- Mentors
- Ram G Athreya, Ricardo Usbeck, Aleksandr Perevalov, Andreas Both
- Organization
- DBpedia
- Technologies
- python, javascript, rdf, docker, rest api, sparql, natural language processing, shell scripting, GitHub Actions
- Topics
- machine learning, natural language processing, semantic web, natural language understanding, linked data, knowledge graphs, Explainable AI, Named Entity Recognition, question answering
DBpedia unifies the amount of information on the web in Wikimedia projects and provides access in the form of an Open Knowledge Graph. Started in 2021, a DBpedia GSoC project led to the development of a Dialogflow-based chatbot that enables users to access the DBpedia Knowledge Graph (KG) using Natural Language (NL). This QA system provides accessibility but is still quite static. The goal of this GSoC project is to make the QA chatbot more explainable and accessible to increase credibility, accountability, and trust, which helps the user obtain explanations on how the result was obtained, what resources were used, intermediate processing steps and components’ behaviour. This project would achieve the goal by 1) Work with the current codebase to refine and refactor, followed by deployment and a CI/CD pipeline using Docker and GitHub Actions. 2) Introduce new features to the chatbot by adding scenarios that adds explainability on Qanary pipeline and its components. 3) Evaluation of the DBpedia chatbot by running A/B tests to measure user satisfaction 4) Machine Learning integrations might be used to create recommendations for improved QA pipeline configurations.