Friday, 26 August 2016

Facebook in Open Source AI Code Contest with Google

Facebook and Google are two of the top web giants even a kids know of. Recently there seem to be competition between the two as Facebook struggles to stand shoulder to shoulder with Google in making open source Artificial Intelligence (AI) code available freely regularly and as early as possible. Doing this will allow innovative ideas that will that will improve the code from experts. According to Facebook, the aim is to fast track progress in the field of machine learning by motivating research. Both companies have invested in AI and this has gone a long way in encouraging AI programmers.


According to Quartz, yesterday, Facebook made an announcement that three of its AI software tools concerned with image recognition. The social network giant says that the technology could allow Facebook users search for photos based on what they depict, without relying on the tags others had assigned to the images. Facebook also claims it could be used to identify the nutritional value of food just by taking a picture of it. 

Programmers working in the area of image recognition will find it easy to integrate this powerful tools.On the software platform, pictures can easily be dissected into necessary component parts and labelled. 

The stage is set for a contest between Google and Facebook. It is believed that a good way to measure success on either side is through citations of academic papers published alongside research tilted towards this field. Talking about citations on research papers published by the companies in the past, Google may take the lead. For instance Facebook's "Memory Networks" published has 134 citations on Google Scholar. Google and Magic Leap's "Going Deeper with Convolutions" also published the same year has more than 1,400 citations. This has nothing to do with the fact Google owns the Scholar Platform.

Another good thing about making the code open source is that a new way of scoring programmers in terms of popularity and impact among peers is introduced. Just like what is observed in GitHub one can start to see how many people are interested and experimenting with a research organization's code. To have a popular repository of repository of code means that there are other contributors that find use in the work, and further the work by locating bugs and expanding functionality.

A citation might prove influence, but seeing someone else using your code shows objective worth. Francois Chollet, a Google engineer and the author of open source deep learning library Keras said "Researchers care a lot about the codebase itsef and the developer community around it, because they are the users who will need to dig into source code and modify it for experimental purposes". 


Unlike Google and Facebook, Apple is not so generous with code. The company only publishes code necessary for building software in their own domain. Machine learning language are rarely made public.

0 comments:

Post a Comment

Add a comment here

Advert