Following my previous post from Day 133, this post is a summary of another talk from the Law Society's conference held last year. You can watch a recording of the conference on YouTube.
Dale Lane - Developer, IBM Watson
Dale began by giving an example of one of the early showcases of Watson, where it beat some of the best players of an American TV quiz show, after learning from past episodes of the show. He explained how the same principles of machine learning can be applied to more important applications such as in the diagnosis of cancer.
Dale highlighted the fact that machine learning algorithms are now being designed which are more generalised, as opposed to older ones which were designed for very use cases. He also spoke of the ease of access to these methods, such as Google's open-source TensorFlow library, to the point where their use is simply a hobby for some people - even children!
Dale concluded with three points:
AI systems are getting smarter
Advances in both hardware and software have made things viable that did not seem possible a few years ago
The use of machine learning libraries is getting easier