The new Google chips promise to accelerate the power of AI like never before # GoogleNext18



[ad_1]

Artificial vision, speech recognition and robotics in general can be improved. Everything, in sync with the IoT and the cloud.

Google has officially announced today new versions of chips developed specifically for artificial intelligence tasks. They also announced software that takes advantage of all the power of the units.

The huge amount of data generated by our devices is the raw material of artificial intelligence. Today, taking advantage of the right way remains a challenge in many cases.

Google introduced two new products today to develop and deploy smart devices. The first is an Edge TPU, an ASIC chip designed specifically for Google and designed to run TensorFlow Lite ML models.

These chips are optimized to speed up and enhance the features of the TensorFlow Lite. ;artificial intelligence. They can be used for industrial purposes, such as predictive maintenance, anomaly detection, artificial vision, robotics, voice recognition, among the innumerable possibilities of AI.

The TPUs are really small and focus their efforts on improving energy consumption and performance. The Edge TPU chip in the following image is referenced by a standard American penny. UU:

  1 "title =" 1 "src =" https://media.metrolatam.com/2018/07/25/image1-3c7325e4109294318ccb8e359251f6ab.jpg "/>


<p> The next announcement was Cloud IoT Edge, a software that extends the capacity of Google Cloud AI. It allows you to create and train machine learning models in the cloud, then run these models with more freedom. These are the software programs that allow IoT applications to be "smarter and safer," they said. </p>
<p>  The software allows you to run Google Cloud templates on Edge TPUs. A supplement that aims to bring the AI ​​to new levels, since it can process and badyze images, videos, gestures, acoustics and movement <strong> locally in peripherals </strong> instead of sending raw data to the cloud and waiting for an answer. </p>
<p>  As mentioned by the company, Cloud IoT Edge can run on Android Things or on Linux-based devices. From Google, they have also created a development kit that they will launch in October. You can get it by registering here </p>
</p></div>
</pre>
</pre>
[ad_2]
<br /><a href=Source link