DataHour: Grounding Trustworthiness in LLM-Based Systems with Knowledge Graphs

If ChatGPT has become a part of your life to assist with your tasks, you're probably familiar with the downside that comes with its creative ability - hallucination. However, we can't solely blame language models, as hallucination is not a bug but a feature. But is there any approach that can allow us to trust the information more from LLM-based applications?

Yes, and the answer is integrating LLM-based systems with knowledge graphs. In this session, we will briefly walk you through the concept and strengths of knowledge graphs, and also provide a simple demo showcasing what the outcomes look like when LLM-based apps are grounded with factual information from knowledge graphs.