Machine learning by an imp-op-steur
summary
A hands-on proof of concept building an LSTM-style RNN in pure Go using the gorgo math library, trained on text (Shakespeare and Les Tontons Flingueurs) and on tic-tac-toe graph paths. Shows writing and exporting learned weights (*.bin), running the learned policy across multiple runtimes including the browser via WebAssembly, and contrasts supervised training with reinforcement learning while discussing export formats and portability.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 3:56Linear regression and sales
- 7:27Perceptrons and simple nets
- 10:43Graph learning with ML
- 13:35Recurrent nets and memory
- 17:50Shakespeare text generation
- 20:42Reinforcement learning and tic-tac-toe
- 24:15Serverless deployment demo
- 28:02Future of ops and ML
- 33:05ML engineering and limits
- 37:19Model sharing and standards
- 40:33Conclusion
- 44:31Q&A
related talks
31mMachine Learning for Gophers
37mA Funny Thing Happened On The Way To Reimplementing AlphaGo
18mMachine Learning and Go
33mspaGO: a work-in-progress yet ready-to-use library for ML/NLP
53mTensorflow, deep learning and modern RNN architectures, without a PhD
33mLearn Neural Networks With Go - Not Math!
29mThe Gopher Orchestra: Distributed Machine Learning with Go
29mLearn Neural Networks With Go, Not Math