Google is developing TensorFlow, the machine learning system which powers several products, such as Google Photos search, Google app speech recognition and Inbox Smart Reply.
According to search engine giant, the technology is now smarter than the first generation and the company will continue building and creating neural nets up to five times than before, and endlessly improve its products, reported TechCrunch.
Years ago, a user could not talk to a Google app through a noisy sidewalk, or read a sign in German using Google Translate or quickly find photos of a favorite dog in Google Photos. The apps were not smart enough but in a short time, they have gotten much smarter. Machine learning contributes a lot to that.
The Google Brain Team engineers and researchers originally developed TensorFlow in the Machine Intelligence research organization to conduct deep research on neural networks as well as machine learning, but the technology also applies to other domains.
"TensorFlow is what we use every day in the Google Brain team, and while it's still very early days and there are a ton of rough edges to be ironed out, I'm excited about the opportunity to build a community of researchers, developers and infrastructure providers around it," said Vincent Vanhoucke, Brain Team Tech Lead and Manager.
The goal of machine learning is to create a technology that works like the human brain.
CEO Sundar Pichai who made the announcement via the Google official blog pointed out that the best systems of today find it difficult to do what a four-year-old child can do, like knowing a dinosaur after seeing samples. He explained that open-sourcing the technology will speed up machine learning research to improve the technology and benefit the whole community.
Google believes machine learning will be useful in research so complex data would make sense, such as, crunching astronomy data and protein folding. The tech can identify what is in the pictures and videos; it can read and understand speech and written text. It can do more. For example, it powers Inbox's Smart Reply, a way to create automatic responses to emails. It reads the email content, then suggests phrases for the user to use as a reply.
Google is not the only big tech firm that builds machine learning systems. Other companies do the same and name it IBM Watson, Azure Machine Learning and Amazon Machine Learning.
