Agile experiments in Machine Learning with F#
summary
Covers using F# on .NET to run agile experiments in machine learning and forecasting models, focusing on practical techniques for applying math to business problems. Names F# and .NET as the technologies and emphasizes test-driven development (TDD) and iterative/experimental workflows for building ML solutions.
AI-written summary · about
chapters (9)
- 0:00Intro
- 1:03Why this talk? - Machine learning competition as a team Team work requires process
- 12:00Observation Single feature • Never complete, no binary test - Many experiments • Possibly in parallel • No correct model - any model could work. If it performs better, it is better.
- 21:40Recap Traditional software: incrementally build solutions by completing discrete features, • Machine Learning: Create experiments, hoping to improve a predictor Traditional process likely inadequate
- 32:04Strategy Use types to represent what we are doing Automate everything that doesn't change data loading, algorithm leaming, evaluation Make what changes often (and is valuable) easy to change: creation of features
- 38:13Food for thought Use types for modelling Model the process, not the entity Cross-validation replaces tests
- 40:43Recap -F# Types to model Domain with common language across scripts - Separate code elements by role, to enable focusing on high value activity, the creation of features
- 41:33Reproducible research Anyone must be able to re-compute everything from scratch Model is meaningless without the data Don't tamper with the source data Script everything
- 49:33Open questions Better way to version features? • Experiment is not an entity? • Is pre-processing a feature? Something missing in overal versioning computation...
related talks
58mAgile Experiments in Machine Learning with F#
01h 02mMachine Learning with F#, Redux
27mN Ways to Do Machine Learning on .NET
30mUsing Machine Learning and Generative AI for Agile Transformations
43mAgile Strategies for High-Impact Data Science & Machine Learning Projects
57mMachine Learning for .Net Developers
37mForecasting using machine learning methods
57mMachine Learning for .NET Developers with ML.NET