The most popular and comprehensive Open Source ECM platform
Let us know more
Artificial Intelligence: Step-wise Advances Enable Machine Learning to Evolve
AI that can be used to self-build more complex and more accurate AI is beginning to happen. Google’s NASNet researchers say that they have developed AI that can do just that. The computer-generated AI system is called NASNet, and it is able to identify objects in pictures more accurately than existing AI algorithms. For example, NASNet is able to spot objects like people and household objects in photos and videos.
Google describes their AutoML (Automatic Machine Learning) approach for NASNet here. It’s quite an accomplishment, but maybe it is not as groundbreaking as the headlines about it would have you believe. The NasNET approach is to use an accepted model for image recognition but to then add another higher-level of machine learning that tunes the parameters of the machine learning feedback layers used for image recognition, parameters that had previously been chosen by humans. The result is that NASNet is more accurate than previously developed AI algorithms.
NASNet researchers said that “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.” Machine Learning will most likely evolve in this way, with a step-wise replacement of tasks done by humans, gradually advancing the capabilities of what machines can do.
Do humans need to worry? Ray Kurzweil said that “my view is not that AI is going to displace us. It’s going to enhance us. It does already.”
