Postgres and the Artificial Intelligence Landscape
summary
Shows an actual PL/Perl demo that builds tensors and neurons inside Postgres, trains on 10,000 training rows, and tests with additional runs, achieving roughly ~72% accuracy on the toy task of predicting whether a number contains a zero in a given digit position. Covers running ML in-server with PL/Python and MADlib, GPU-accelerated workloads, and concrete production reasons—data locality, transactionality, indexing and realtime fraud/voice/image inference.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 2:17What is artificial intelligence?
- 6:28Machine learning and deep learning
- 12:31Demo using Tensors with Postgres
- 17:08Hardware/software efficiency
- 19:05Tasks AI is particularly good for
- 22:16Why use a database?
related talks
01h 00mPostgreSQL and Artificial intelligence
45mPostgres and the Artificial Intelligence
50mFrom Transactions to Intelligence: Postgres for Open AI & Analyti
35mPostgres in Artificial Intelligence
37mRound Table Discussion
53mModern PostgreSQL for AI...
27mPutting AI in Production
32mSimplify End-To-End MLOps with PostgresML