From Zero to Machine Learning in 45 Minutes
summary
Overview of a hands‑on path from zero to a working reinforcement agent using AWS DeepRacer: Python reward functions, action space (steer/brake/throttle), and a PPO policy/value pair trained in SageMaker with RoboMaker simulation and Lambda/CloudFormation orchestration. Presenter cites a 300,000× rise in AI compute (2012–2018), ImageNet top accuracy ~97.3%, and shows training workflows (recommend 30+ minutes, AWS gives 8–10 hours free training); trains on ~10 variable tracks, exports a model to a 1:18 car and demonstrates it driving successfully.
AI-written summary · about
chapters (22)
- 0:00Intro
- 3:08What is Machine Learning
- 5:13Why Care
- 10:17Historical Developments
- 15:29Types of Machine Learning
- 18:08Unsupervised Learning
- 19:40Clustering
- 23:18Evaluation
- 28:25Reward Function
- 35:14Models
- 35:41Tracks
- 36:01Input Parameters
- 36:33Model Actions
- 37:14Sample Code
- 38:07Reward
- 38:17Train
- 38:49Heatmap
- 39:02Rewards
- 39:23Virtual Races
- 40:12New Hardware
- 40:24Deep Neural Network
- 41:47Advanced Skills
related talks
27mGet rolling with machine learning on AWS DeepRacer
24mGet Rolling on Machine Learning
29mGet rolling with ML on AWS DeepRacer
23mGet rolling with machine learning on AWS DeepRacer
37mAWS DeepRacer:train, Evaluate, and Tune Your Reinforcement Learning Model
30mStart your ML journey with AWS DeepRacer for public sector
37mAWS DeepRacer: Train, Evaluate, and Tune
29mKickstart Your Machine Learning Career with AWS DeepRacer