AI is making my life very difficult these days. And no, this isn’t a rant about AI taking jobs; I’m talking about the Artifical Intelligence class I’m taking in school this semester.
The class itself is very interesting, and the approach our professor takes in teaching is quite challenging, something that until now I’ve felt was largely lacking in other classes. On day 1 of the class we were given an assignment due in two weeks to come up with a topic of interest in AI research and learn enough about it to write a formal research proposal. The proposal had to include a description of some problem within an area of AI and the state of research into that issue, and a plan for improving upon existing research with a novel approach to solving the problem. (I’ll talk more about that project another time.)
No sooner had we finished that than our professor was giving us our first homework assignment: create an A*-based search algorithm capable of solving a sliding box puzzle. If you’re familiar with A* search, you know it’s an approach generally used to solve problems in which each state is the spatial position of some element within a coordinate system. This problem required instead that we think about the states of the system as a whole, which for someone brand-new to A* and other algorithms in the AI arena was no easy task, and indeed I mulled over this problem for several days.
The solution ended up being inspired by my wife, who was patiently listening as I explained the basics behind the problem as a way of talking through how I might solve it. While she is very smart and highly educated, computer science is not her forte or her passion. As it turns out, this was quite helpful for my brainstorming, as she asked several simple but significant questions that helped me to finally see a path to the solution. A more CS-literate person might have easily glossed over or at least approached those questions from a more technical point of view; her layman’s perspective led me to reason through answers in a way that broke through the obstacles I was encountering.
That perspective is valuable in any domain, because too often experts find themselves trapped in mental ruts that are formed by the knowledge and practices typical of their profession. Those ruts are necessary, a sort of mental muscle memory that reduces low-level thinking to nearly a reflex and frees up brainpower to focus on more abstract and esoteric problems. And while using colleagues as a sounding board is certainly helpful and often crucial in problem solving, they are usually bound by those same ruts. An outsider’s perspective, while perhaps naive, can provide unique insights into the way that things work, giving one a sort of “periscope” to look outside the structure of one’s domain and see a problem from a new or unexpected angle.
So I’ll continue to entertain my wife with nerdy problems, for as long as she’ll let me.