Growing Clojure via InferenceQL Synthetic Data Generators
summary
InferenceQL is a Clojure/ClojureScript platform that learns probabilistic programs from tabular data and exposes a SQL-like generate expression that returns an infinite synthetic table; models trained on its synthetic data show lower error than a neural‑net approach and a banking baseline. It models hundreds to thousands of rows and 10 to several thousand columns across datasets like oncology trials, whole‑genome RNAseq, the 2021 Stack Overflow survey and US Census, supports tracing via let‑traced traces, editable programs to remove gender–salary coupling, and a GPT‑3 integration for natural‑language queries.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:28Overview
- 4:15InferenceQL
- 8:25Generating Synthetic Data
- 13:19Reducing Algorithmic Bias
- 14:58Growth
related talks
38mProbabilistic programming and meta-programming in Clojure
29mDeep Learning Needs Clojure
35mScalable structure learning and inference for domain-specific probabilistic prog
41mA Clojure fusion of symbolic and data driven AI
19mHow to use generative AI to deliver self-service data products
47mApproximate Lifted Inference with Probabilistic Databases
18mAccelerate generative AI development with integrated data platforms
01h 00mThe Synthetic Data Advantage: How It Helps You to Get Both AI Privacy & Explainability