(MY) Connected Car Project: the Next Big Thing

Yes, solutions of the future are built as mashup. Assembling building blocks.

Some months ago I purchased the Google Voice Kit and I assembled it.
I have also used it for some demo at work. It is a very nice and cost effective way to show how you can build custom solutions using Google Assistant Technology, Speech-to-Text (STT) and Text-to-Speech (TTS).

Now, the idea is: why not to run the OBDII python code on the Google Voice AIY kit?
It should be possible and easy: we have an RPI 3, and Python SDK from Google.

Therefore, this is MY Next Big Thing: my OBDII Python Code running on this:

If you want more details, see older posts and have a look at the MY Connected Car Project

Last but not the least: my idea of playing with "cars" started from a suggestion, and work we did together,  from one great collegues and friend: Fabrizio Marini. Thanks guy!

Real Time Objects Detection

Wow: from time to time you find something that take your breath away.
It is simply astonishing what you can do today with powerful Deep Learning algorithms and the power of GPUs.
One of the area where the progress is "so fast" is Computer Vision. It is not a surprise, for example Computer Vision is very important for the future of the self-driving car.

This project YOLO (You Only Look at Once) is terrific. Real Time object detection, so fast that you can identify objects real-time on a video.

See the demo !

YOLO

Deep Learning and Keras

Ok, let's talk about one wonderful subject: Deep Learning.
Every time you go to Google Translate you use Deep Learning. Every time you use technology, like the Chatbot I was talking about in my previous post, you use Deep Learning.
The promises of AI are coming back. Now almost reality. Exciting.
 
In one of my previous posts I talked about one of the most promising framework for Deep Learning in Python: TensorFlow.

TensorFlow is powerful, but it is a little bit low-level. Good if you want to enter in details and control, efficiently, all the aspects.

A really interesting "Frond-End" framework, born to simplify usage of Deep Models is:

Keras

Keras normally use TensorFlow as backend, but it can be configured to use Theano.
With Keras it is easier to design, compile and run a model. It is easy to switch from CPU to GPU.
 In the following gist, with code taken from Keras public documentation, you see how easy is to define, train, test a Neural Netwok Model.
In the documentation you find also full fledged examples, with data (for example MNIST).
One word of caution: if you try more complicated models, with many layers, the training can take too long on CPU. I'll return back.