Put Your Machine Learning on Autopilot
summary
Shows a concrete AutoML pipeline that takes the Amazon Customer Reviews CSV (130M records available, demo uses camera category), loads a pre-trained bird model from Hugging Face, and runs Amazon SageMaker Autopilot to auto-detect a multi-class problem, preprocess with TF‑IDF, train and tune up to the chosen maximum candidates (demo uses 3–5). A 20‑minute run raises accuracy from ~27% to ~53%, emits two Jupyter notebooks documenting preprocessing and candidate code, and deploys the chosen candidate as an endpoint.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:56Overview
- 1:44Machine Learning Workflow
- 4:03Data Preparation
- 6:13Prototyping
- 7:10Demo
- 15:09Automation
- 15:36Machine Learning
- 16:03AutoML
- 17:23Transparency Control
- 18:29AWS ML Stack
- 19:58Autopilot
- 20:46Whitepaper
- 21:38Steps
- 22:03Demonstration
- 29:52Outro
related talks
31mUse AutoML to create high-quality models with just a few clicks
32mBuild quality ML models easily & quickly with Amazon SageMaker Autopilot
29mAutomatically Build Models With Amazon Sagemaker
30mBring AutoML to all users across your business
31mUse AutoML to create high-quality models with just a few clicks
19mDrive transformation through machine learning
51mProfundizando en Aprendizaje Automático para Desarrolladores
45mAutomate your ML workflows with pipelines