Microsoft has just open-sourced it’s artificial intelligence framework, CNTK. This is the same AI that powers the company’s own Cortana personal assistant, as well as the new Skype Translator applications.
Released back in April, the framework is used to allow systems to recognize human speech, as demonstrated with both Cortana and Skype, but also photos and videos. It does this by simulating the function and structure of the human brain. It was initially closed off and available only for non-commercial use, but Microsoft has just made its brain available to anyone, bringing new meaning to the phrase “collective intelligence”.
Microsoft wants to make its technology available for other companies and academics to take advantage of. This falls in line with what other companies have been doing as of late with their own AI technology, such as Facebook and Baidu. Google’s open-source TensorFlow framework was also released last year, powering the company’s search engine and other applications for Android.
Microsoft’s CNTK has one upped Google’s efforts to open-source AI framework in terms of scale, as it is able to utilize the power of multiple servers simultaneously. According to Microsoft’s chief speech scientist, Xuedong Huang (pictured), Google’s release is more limited, and thus less practical for real-world AI applications. “We want this to be useful not just for academics but for commercial artificial intelligence and deep learning companies.”
Another advantage of CNTK is that it is one of the few AI frameworks compatible with Windows. It is also more efficient than other AI tools. The downside is that it currently only supports C++ and its own proprietary language, which can make using it difficult. Fortunately Microsoft plans to add support for Python and C#, the former of which is more commonly used among AI research.
Source: Wired