Postgres and the Artificial Intelligence
summary
Demo shows running simple neural nets inside Postgres using plpython and a hand-rolled plperl example, training on ~100–1,000 synthetic rows and reporting poor accuracy (~10–30%) until data volume grows. Argument claims major benefits: avoid exporting CSVs, run models next to production data, do continuous training and DB-focused predictions (eg transaction wraparound, fraud, resource forecasting). Notes support for GPU-based computation and concrete examples of tensors, random splits, weights, and test/train pipelines that actually run inside the database.
AI-written summary · about
chapters (8)
- 0:00Start
- 1:11What is Artificial Intelligence?
- 2:41Types of Artificial Intelligence
- 15:37Deep Learning
- 25:53Generate Tensor
- 34:58Final Table Expression
- 36:34Why Use Database?
- 39:10Advantages of doing Machine Learning in a Database?
related talks
01h 00mPostgreSQL and Artificial intelligence
25mPostgres and the Artificial Intelligence Landscape
35mPostgres in Artificial Intelligence
50mFrom Transactions to Intelligence: Postgres for Open AI & Analyti
27mPutting AI in Production
37mRound Table Discussion
53mModern PostgreSQL for AI...
34mAnalytics and AI with PostgreSQL - Mission Possible!