Announced yesterday in TensorFlow Developer Summit
- its faster
- its more flexible
- more production-ready than ever
And a list of other updates (in screenshots)
Also announced last week TensorFlow Fold - "makes it easy to implement deep-learning models that operate over data of varying size and structure."
you can watch the TensorFlow Developer Summit talks on YouTube
@sean_murphy Thoughts. Uber could be using this in many different ways. One area where this comes in handy is route optimization. For example all the drivers are busy and you request a ride, the system has to find someone who is dropping someone off close to your location and send him the request. There might be many Uber drivers in that area but calculating all the variables and making sure the right driver is selected is a challenging thing. Another area where this could be useful is calculating prices. Uber should estimate fares based on a number of variables like distance, traffic, etc.
For Airbnb the most important use case could be recommendations. ML-based recommendation can learn from your users and what they select, share and rent.
ContentKing