As we have already commented on some occasion, we could say that, without neglecting other branches of business, the truth is that in Microsoft seem to have bet practically their entire future on issues of Artificial Intelligence, one of the most promising lines of business. You have proof of what I say in how other great technology companies, such as Amazon, Google and even Apple, are also betting on this business model.
To try to be the spearhead in this new market, Microsoft has made the decision to make available to all interested parties a set of open source utilities with which to enhance the Machine Learning with natural speech recognition. This means that anyone can integrate the North American company's voice recognition system into their application or device, a platform that, as a reminder, stands out for having a conversation error rate of only 5,9%.
Microsoft makes its Machine Learning platform with natural speech recognition available to all interested parties.
If you are interested in the proposal, tell you that this voice recognition system has been baptized by the company as a cognitive tool kit and is already available, in beta and licensed by MIT, in the GitHub repository. Among its most interesting features, it should be noted that it offers the possibility of building neural networks or developing machine learning systems taking advantage of the use of CPUs and GPUs with sufficient scalability to become the standard for this technology.
As has been published by Microsoft itself, in addition to continuing to develop its Machine Learning platform with natural voice recognition, this effort to release the project has a lot to do with the idea they have in the company of democratize artificial intelligence in order to achieve that all these products can finally have great value in the real world.