Machine Learning for Software Developers and Knitters

Explore machine learning for software developers and knitting enthusiasts in this NDC Melbourne 2021 conference talk. Discover the latest advancements in AI/ML tooling, making it accessible to professionals beyond data scientists. Learn about innovative applications in risk identification, conversational AI, autonomous driving, vaccine optimization, and art creation. Gain insights into AWS services designed to simplify the machine learning process. Follow along as the speaker demonstrates a unique project: teaching a computer to reverse engineer knitting patterns in pursuit of creating the world's first knitting replicator. Delve into the practical steps of machine learning, including data collection, labeling, model training, deployment, and building a web frontend for inference. Examine real-world training results and performance metrics, showcasing the potential of custom machine learning models in specialized domains.

Intro
Machine Learning for Software Developers
AWS Machine Learning Stack
Get training data 2. Label training data 3. Train ML model
1. Get training data
Deploy ML model
Build web frontend for inference
Training job #1 Number of training Images: 368 Validation-accuracy=0.458333
Rekognition Custom Labels Number of training images: 900 Average precision=0.977 Overall recall=0.985 F1 score=0.981
But wait, there's more...