Large Language Models are amazing, and with our research project, we aim to make them even more amazing! Our project will connect large language models to structured knowledge such as knowledge bases or databases. With this,
1. language models will stop hallucinating
2. language models’ knowledge can be audited and updated reliably, to spot biases and make them more interpretable
3. language models will become smaller and thus more eco-friendly and deployable
We work in the DIG team at Telecom Paris, one of the finest engineering schools in France, and part of Institut Polytechnique de Paris — ranked 38th in the world by the QS ranking. The institute is 45 min away from Paris by public transport, and located in the green of the Plateau de Saclay.
Check out our Web site to apply: https://suchanek.name/work/research/kb-lm/index.html