Google supercomputer creates its own ‘AI child’

Google’s Brain team builds its own AI child, that has an increased level of recognition than any other AI in existence.

  • Published: December 5, 2017 11:26 AM IST

These days Artificial Intelligence has become a part of any and everything. Google, in May this year announced the creation of AutoML, an artificial intelligence (AI) that’s capable of generating its own AIs. Now, Google’s AutoML artificial intelligence (AI) system has created its own “AI child”.

In simple terms, this Google-made computer-system, which is known as NASNe, is able to identify people and objects separately. In technical terms they are known as ImageNet image classification and COCO for object detection.

The new system was able to identify objects with 82.7 percent accuracy. According to the report, the accuracy level is 1.2 percent higher than any other system. It’s also listed as being 4 percent more efficient than man-made devices. However, the Google Brain team acknowledge that this is just the beginning of machine learning.

Google states that it created its child AI using a technique called reinforcement learning. Further, Google says that this works similar to machine learning, except it’s entirely automated where AutoML acts as the neural network for its task-driven AI child. It’s also worth noting that Google has open-sourced NASNet for inference on image classification and detection in TensorFlow repositories.

“We hope that the larger machine learning community will be able to build on these models to address multitudes of computer vision problems we have not yet imagined,” Google says in its blog post.

Coming back to AI, Gartner recently predicted that AI will be virtually present in almost every new software product and service by 2020.

  • Published Date: December 5, 2017 11:26 AM IST