Google's AI Map
summary
An AMA that gives concrete tools and results: demos show Vertex neural architecture search producing a more accurate spam model than Model Maker, a dermatology-on-device demo using TensorFlow Lite that actually runs, and a forecasting feature that predicts right from a Google Sheet. Advice centers on notebooks, distributed training on TPUs in the Cloud, Vertex pipelines, AutoML export to TensorFlow/TFLite, Keras Tuner for hyperparameters, explainable AI and Know Your Data for debugging dataset and drift issues.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:00How Can the Cloud Help with Machine Learning
- 2:35What Tools Does Google Offer To Help with Ml Ops
- 9:37What Are the Tools and Frameworks To Help Me Solve Computer Vision Problems
- 10:42Automl Vision
- 17:11What Are the Best Ways To Train a Model if You'Re Not a Data Scientist
- 18:23Tensorflow Lite Model Maker
- 21:01Computer Vision
- 28:36What Are Best Practices for High Scale Machine Learning
related talks
31mBringing AI to Your Data
20mAI and Machine Learning for developers
32mMachine Learning with Google Cloud Platform
32mGenerative AI Text and Multimodal Embedding Models for Real World Use Cases
44mBuilding smarter solutions with no expertise in machine learning
01h 03mIt's All About AI
12mKeynote: What's New in AI
57mBuilding smarter solutions with Machine Learning, from magic to reality