Type-Driven Automated Learning with Lale
summary
Lale is an open-source Python library that enables semi-automated machine learning by letting users search over pipeline compositions with a few lines of code. It integrates with scikit-learn and existing AutoML backends, returns runnable pipeline objects plus human-readable pipeline reports, and supports freezing components so only chosen operators are tuned; demos show end-to-end reproducible workflows for classification and regression.
AI-written summary · about
chapters (8)
- 0:00Lale introduction
- 3:57Automation in ML pipelines
- 10:30Operator selection lifecycle
- 17:36Type-driven automation
- 24:47Use case demos
- 31:48Customizing pipelines
- 36:10Interoperability challenges
- 39:07Summary & Q&A
related talks
25mScaling ML pipelines with KALE — the Kubeflow Automated Pipeline Engine
41mFairness for Scikit-Learn Pipelines with Lale
23mAutomation: Build A Training Pipeline
35mMachine Learning Pipeline using Luigi and Scikit Learn
47mBuilding Python based AI Systems with LLMs
35mA Generative AutoML for Tabular Data
17mCrafting Superior ML Pipelines with SapientML
49mML Pipelines