Google open-sources AI engine of its core services

Google yesterday took a big step toward building artificial intelligence by open-sourcing part of TensorFlow, the ''deep learning'' engine that powered the company's speech recognition, its Photos app, its translation services, and other services.

TensorFlow, a library of algorithms allows Google to train computer systems, called ''neural networks,'' to think in ways similar to the manner in which humans do.

The neural networks perform complex mathematical operations on arrays of data, called tensors, identifying patterns and relationships worked through the available information.

Google's Photos app is powered by a neural network that allows the app to learn the relationship between an object's name and its appearance so it could identify a cat within a photo, on the basis of similar pictures it had seen before, while another neural network allows the Google Translate app to learn the use of words in conversation, so it could provide more accurate renderings.

With the progress, Google had achieved in the field of deep learning over the past few years, it thinks the technology could be improved by letting others help out.

''We hope [open-sourcing TensorFlow] will let the machine learning community - everyone from academic researchers, to engineers, to hobbyists- exchange ideas much more quickly, through working code rather than just research papers,'' Google CEO Sundar Pichai wrote in a blog entry.

''And that, in turn, will accelerate research on machine learning, in the end making technology work better for everyone."

Companies ranging from Google to Facebook Inc. to Microsoft Corp are seeking AI play by staffing up research labs, publishing academic research papers, giving presentations at conferences and even guest lecturing at universities.

They are looking to attract top talent from academia to work for them, while encouraging the wider community to work on new AI technologies.

Other companies which had been which had been more secretive in the past, are seeking to be more open, Amazon and Apple to name two.

With the release of TensorFlow, Google aimed to make the software it built to develop and run its own AI systems a part of the standard toolset used by researchers, according to Jason Freidenfelds, a spokesman for Mountain View, California-based Google.

He added Google had shared TensorFlow ahead of its release with a few people. Christopher Manning, a professor of Linguistics and Computer Science at Stanford University, had with two student associates, been writing AI programs using TensorFlow.