Machine Intelligence at Google Scale
summary
Overview of available Google Cloud ML tools and concrete outcomes: ready-made APIs like Cloud Vision API, Speech, Natural Language, Translate (now neural MT) and Video Intelligence return labels, bounding boxes, sentiment and OCR; demo chains speech→NL→custom search and Vision OCR; TensorFlow and Tensor Processing Units power models and Cloud Machine Learning Engine runs distributed training to speed custom model training 10x–40x using GPUs/TPUs and many servers, or export pretrained models (Inception v3) for local retraining.
AI-written summary · about
chapters (21)
- 0:00Start
- 3:44How did we escape the Al winter?
- 6:21Google Photos
- 8:26Machine Learning is everywhere at Google
- 11:17A shift from mobile-first to Al-first apps & services...
- 17:37Two ways Google can help you benefit from ML
- 18:23Face detection
- 19:01Landmark detection
- 20:02Vision API-new features
- 20:36Web annotations
- 28:29Extract entities
- 29:23Analyze sentiment
- 29:43Analyze syntax
- 32:00How the demo works
- 36:59airbnb-connecting guests through translation
- 38:21Calling the translation API
- 40:53Neural machine translation improvements
- 44:58Video API Response: Label detection
- 45:25TensorFlow-Google's 2nd gen. OSS deep learning library
- 47:42Cloud Machine Learning Engine
- 49:25Try them all in your browser!
related talks
42mMachine Learning APIs, from Training [...]
47mBuilding smarter apps with Machine Learning, from magic to reality
44mBuilding smarter solutions with no expertise in machine learning
57mBuilding smarter solutions with Machine Learning, from magic to reality
42mUse Machine Learning in your code, without being a ML expert
54mBuilding smarter apps with Machine Learning, from magic to reality
07mMachine learning at Google
36mCloud AI Services: What they are and how to use them