Embeddings & RAG in Science

Jefferson Building, Room 450, 17 Oxford Street, Cambridge, MA 02138

Instructors

Bret Nestor and Claudio Battiloro

Learning Outcomes

  • Embed data and view t-SNEs of representations
  • Embedding foundation models for image, text, and geospatial analysis
  • Explore diminishing returns w.r.t. training data size
  • Construct a basic RAG script

Use Cases

  • Pure embeddings vs generative models
  • Explore non-language Foundation Models
  • Implement RAG systems for research

Prerequisites

  • Familiarity with Python
  • Basic understanding of vector spaces