Can a Video Really Teach Artificial Intelligent Common Sense?

Technology has come a long way over the last 20 years, particularly when it comes to software. Just five years ago, a massive leap was made in the accuracy of image interpreting software. The technology behind it is responsible for the even more recent boom in artificial intelligence (AI) and is what’s known as artificial neural networks.


This advancement in technology opens the door further for new applications that use facial recognition. Neural networks are now being used for machine vision and could lead to the development of software that has common sense. Machine vision is all about recognizing images and give data about that object. It can include specific details such as a particular species of plant, or a certain breed of dog and can even distinguish between different events such as a wedding or a birthday party.

But, there’s still a long way to go before this technology is anywhere near perfected.  One way in which this technology lacks currently is when it comes to common sense. These machines may be able to detect various objects when they appear in the situations in which they first learned about them, but anything out of the ordinary and they don’t have a clue.  A better way of going about this issue is to use video to allow the AI to learn through real observations opposed to looking at static images. This is the way in which babies learn in the first few months of their life and is one of the most effective forms of learning.


Moving forward it would be great to see learning systems in place that can also predict the future by viewing a few frames first. This is the first step in developing an unsupervised learning system, in which many interesting things could happen, not just for machine vision, bur for AI in general.  It’s also a sign of good things to come, so watch this space.


More News to Read

Comments

comments