Only as good as the data
Practical AI
00:00
Data Strategies in Machine Learning
This chapter explores the critical aspects of creating effective test and evaluation sets in machine learning, emphasizing the significance of random sampling and benchmark data. It discusses the interplay between training data and existing benchmarks, showcasing examples such as machine translation and question answering tasks. Additionally, the chapter introduces retrieval augmented generation (RAG) and its impact on data quality, illustrating how users can enrich generative models with their own data while maintaining their core functionality.
Transcript
Play full episode
