

LLMs and Graphs Synergy
59 snips Feb 10, 2025
Garima Agrawal, a senior researcher and AI consultant, dives into the dynamic synergy between large language models and knowledge graphs. She highlights how knowledge graphs can reduce AI hallucinations and enhance accuracy by integrating domain expertise. They discuss real-world applications, from smarter customer support systems to AI-driven decision-making pipelines. Garima emphasizes the importance of human oversight in AI integration and advocates for a strategic, rather than fearful, approach to adopting these technologies.
AI Snips
Chapters
Transcript
Episode notes
Cross-Industry Innovation with Knowledge Graphs
- A company used knowledge graphs to find similar processes across industries.
- They found a more efficient way to wax wood by magnetizing it, inspired by the auto industry.
LLMs and Hallucinations
- LLMs frequently hallucinate, which is not ideal.
- Augmenting LLMs with knowledge graphs may reduce hallucinations.
Knowledge Graph History
- Knowledge graphs were popular around 2019-2020 but then declined.
- Building question-answering systems on static knowledge graphs is cumbersome due to the modeling and updating required.