Microsoft has made the updated version of its Cognitive Toolkit, the same set of tools that helped its speech recognition system reach a level at par with humans, available for public. The company announced on Tuesday that it is open-sourcing its machine learning tools.
Microsoft Cognitive Toolkit is essentially a system of deep learning tools that has helped in advancements of fields such as image and speech recognition and search relevance on CPUs and Nvidia GPUs, the company said in its blog post. It is also used in company's digital assistant Cortana as well as Skype Translator.
The company said that the new version of the toolkit is not only faster than its previous versions but also faster than its competition's toolkits. Microsoft has provided detailed information about the toolkit on its website.
"One key reason to use Microsoft Cognitive Toolkit is its ability to scale efficiently across multiple GPUs and multiple machines on massive data sets," company's a partner engineering manager Chris Basoglu was quoted as saying. The Cognitive Toolkit has been made available at GitHub.
Last week, the researchers at Microsoft claimed that its speech recognition system has reached 'human parity'. This effectively means that the speech recognition system can recognise words from a conversation as well as humans would do. In order to improve speech recognition to this level, the researchers made use of neural language models, which puts similar words together for better recognition.