Question Generation

from Knowledge Bases

Unleashing the Power of Information

🌍 Multilingual Question Generation with Flan-T5

Unleash the power of our fine-tuned Flan-T5 model! Seamlessly generate complex questions in English, German, and French. Immerse yourself in a world where language barriers are a thing of the past!

📊 Advanced KBQA Training via Graph Linearization

Our cutting-edge approach to graph linearization enables you to train your KBQA model with ease. Transform your sub-graphs into a textual form and unlock the full potential of your data!

🎓 Dynamic Question-Answer Pair Generation from any KGs

Our novel method for generating question–answer pairs from any given KG is a game-changer! Use graph isomorphism to sample a sub-graph and watch your knowledge transfer dreams come true!

INFO

This project is part of the Master Thesis "Multilingual Question Generation from Knowledge Graphs" at the University of Zurich (UZH), Department of Informatics (IFI). The final version will be published soon. The project received a grade of 5.75 🏅, which is near the top of the grading scale in Switzerland, where 6.0 is the highest possible grade.

Abstract

Coming soon!

The abstract will be published soon. Stay tuned!

Keywords

Knowledge Base Question GenerationKnowledge Base Question AnsweringKnowledge GraphGraph IsomorphismKnowledge TransferGraph LinearizationSequence-to-SequenceLarge Language ModelNatural Language ProcessingQuestion GenerationMultilingual Question GenerationGrade 5.75 🏅