So, what’s TensorFlow?

TensorFlow is an open access software library that gives computers the ability to learn without any specific classical programming. It is the part of fifty research and production teams at google and used in various google products, such as Gmail, Google Photos, Speech recognition and Search. It works as an artificial intelligence and helps google design very powerful products. It helps machines to actually see and understand what is in front of them; for example, an image in a short video clip, different languages and processes.

[embedyt] http://www.youtube.com/watch?v=oZikw5k_2FM[/embedyt]

Tooling

TensorFlow was originated in 2011 during the complex research on machine intelligence for the purpose of machine learning and deep neural network research. After development of the software, Google’s Brain team researchers and engineers found that the system is general enough to understand; and can have diverse applications in wide variety of other domains as well. It can run on multiple CPUs, GPUs, as well as on mobile operating systems like Android and iOS.

Senior research scientist Greg Corrado at google said,

“Machine learning is the secret sauce for the products of tomorrow. It no longer make sense to have separate tools for researchers at machine learning and people who are developing real products”.

There should really be one set of tools that researchers can use to try out their crazy ideas, and if those ideas work they can move it directly into products without having to rewrite code.

Problem solving for existing problems

On the research side the goal is to bring new understanding to existing problems, bring state of the art solutions for the existing problems and understand new problems that were not considered before. On the engineering side the goal is to take these insights from the research and make it part of the product features which was never thought before.

Collaboration enabler

Another part of the TensorFlow is to allow collaboration and communication between researchers. It allows the researcher on one location to develop an idea, explore it and just send the code that someone else can use on the other side of the world. TensorFlow is making people to easily use the devices around them and having TensorFlow as an open source tool really helping developers to do a lot more than they were doing before.

Speeding up AI evolution

Google has the best machine learning infrastructure of the world and it is a great thing that they are sharing that with the world.  Google believes that it can speed up the AI evolution; and via open source, outsiders can help google bring innovation and improvements. In recent years, other companies, including Facebook, Microsoft and Twitter also made huge progress in the field of Artificial Intelligence and have free software like TensorFlow; but google is still five to seven years ahead from the rest of the world. Typically, google didn’t share its design and research paper with the world to keep an advantage. However, with TensorFlow, the company changed its track. But google is still keeping an advantage because google is not yet opensourcing the version of TensorFlow that helps in running it across massive range of machines.

[embedyt] http://www.youtube.com/watch?v=bYeBL92v99Y[/embedyt]


If you would like to contribute on Infoginx, feel free to reach out at info@infoginx.com.