This blog documents my attempts to make sense of artificial intelligence and how I can use it to help me to do the science I love doing better. I’m hoping others who also want to make sense of this technology might be curious to follow as I try to find my way through.
As a teenager in the 1980s I was fascinated by the excitement and promises of artificial intelligence. Concepts such as The Lambda Calculus and programming languages such as LISP captivated me, but at the time, I didn’t have the skills to do any more than look on from a distance. Back then the claims for what AI would achieve were lofty, with the darkest implications perhaps best imagined in the Sci-fi classic, The Terminator.
AI slowly drifted from my consciousness since I was too busy juggling studying for a degree and then a PhD in physics, with achieving rock and roll fame and fortune. Happily one of those worked out and I’ve now been teaching and researching in theoretical physics for over 25 years.
Over the past year, I’ve almost by accident started to think about artificial intelligence, or more precisely machine learning, and how I might use it to help solve some of the riddles that I’m working on. One use of AI in the physical sciences that has really taken off is the searching for a needle in a haystack problem. Advances in both hardware and software have enabled modern scientific instruments to collect vast amounts of data that are simply too huge to sift through by hand, even for the most dedicated PhD student. Teaching machines what kind of blips in the data we are looking for and then letting them loose on more data to find similar blips is what AI is really good at. This is the black box magic of AI. We don’t care how the AI finds the blips, we just want to know where they are. For many physicists, it is the black box nature that makes us suspicious and, I believe, has limited the uptake of AI beyond looking for needles.
This blog follows from an ‘aha’ moment I had recently. In this post I share my thoughts on why I started to renew my interest in AI..