Google has offered an Artificial Intelligence code for free for anyone who wishes to distribute and modify. The software programming toolkit that is part of the code, SyntaxNet, is also part of the free offer. The company announced the news in a blog post.
The product manager at SyntaxNet, Dave Orr said the company was hoping that the move would encourage people to use the code instead of developing their own. He said he hoped people would not have to reinvent the wheel.
SyntaxNet is just one of the numerous artificial intelligence software systems that have been launched by the tech giants under the open source license. This is a license which gives anyone who wants access to a software and is able to change and modify it accordingly. This past week Amazon also made the Deep Scalable Sparse Tensor Network Engine obtainable on Github.
Last November, Google also released TensorFlow, one of its machine learning codes. Facebook is also part of the companies which have open sourced software for artificial intelligence. Facebook also has open sourced hardware for the industry.
SyntaxNet is used to help machines in their breakdown of sentences and the small parts of speech and also to help them decode the relationships between the words and phrases and sentence. The work can be seen as the same as that which is done in schools during grammar training, but the machines do it a faster rate.
The artificial intelligence computer program is called Parsey McParseface. It is able to do what other alternative programs can’t do by homing in on the most relevant words. Google has used the program in a few of its services. The search tool in Google and Smart Reply, a functionality that allows suggestion of quick email responses.
The program has also been used by Google in the Google Knowledge Graph, the database of information on which the search engine depends on. Apparently those who worked on the program had trouble thinking of a correct name and after someone’s suggestion, they stuck with Parsey McParseface.
Analysts believe the system is the first step in artificial intelligence. The system can’t understand overall context at the moment because of it’s inability to figure out the relationships between sentences. It is also only available in English. Longer sentences are very difficult for the program, and using it for a chatbot would command expertise and additional code. The ways that the programs use to read and work out in the various fields such as health care and biotech is also not known at the moment. If programs such as Parsey McParseface learn by analysis of news articles, reading and understanding of medical texts might be difficult.
