Super Data Science: ML & AI Podcast with Jon Krohn 785: Math, Quantum ML and Language Embeddings, with Dr. Luis Serrano
5 snips
May 21, 2024 
Dr. Luis Serrano discusses making Math & Quantum ML accessible, teaching AI to beginners, & the power of embeddings. Topics include easy math understanding, learner categories, importance of embeddings, semantic search, exciting AI applications, & Quantum ML.
AI Snips
Chapters
Transcript
Episode notes
Three Core Learner Types
- Learners fall into three dominant types: formulaic, visual, and builder, and effective courses address all three.
- Serrano designs content with small numeric examples, cartoons, and hands-on Python projects to cover each learning style.
Use RAG To Reduce Model Hallucinations
- Use Retrieval-Augmented Generation (RAG) to reduce hallucinations by making models search external documents before answering.
- Serrano recommends embedding and reranking corpora so LLMs answer from verified sources rather than inventing facts.
Embeddings Are Where The Rubber Meets The Road
- Embeddings are the fundamental bridge from human data to machine computation; better embeddings improve nearly every downstream ML task.
- Serrano likens embeddings to mapping words into coordinates so similar meanings cluster closely in high-dimensional space.