The Feynman Algorithm

This quarter, I’m taking more theory classes than I ever though I could handle. One is on optimization and approximation algorithms (CS261) and the other is on geometric algorithms (CS268). Both classes are heavy on problem sets, 4 in one and 3 in the other. Collaboration is encouraged in both classes which is nice. It seems like that is how these kind of problems are supposed to be solved anyway.

The unfortunate thing is that these low frequency problem sets managed to align and I find myself having two weeks to finish twelve theory problems. Even if some of these questions are on the easy side, it’s a lot of work! And I am the kind of person who likes to type up these kinds of things. On the flipside, I really get to work on problem solving. I have always wanted to learn, how to solve problems. I was convinced reading a lot of books on it was the key, but I’m less inclined to think it’s true. I’ve learned that problem solving comes with a bit of ingenuity and a whole lot of practice. I’ve managed to fake my way through it by knowing a few bag of tricks, but in college, that quickly runs dry.

Richard Feynman made an algorithm for solving problems. It is as follows:

1. Write down the problem.
2. Think about the problem.
3. Solve the problem.

I think this is less of an algorithm, and more of a dependency graph. You can’t move onto the next step until you’ve finished your current step. Still, it’s fun to look at the truth that lies in this silly thing. (Wow, that last sentence has the phrase “truth that lies”). The goal of AI should not be the Turing Test, but instead should be implementing this algorithm.

It’s hard to put into words how I approach problems. I think any attempt would boil down to Feynman’s Algorithm. I think one thing that has proved particularly useful when solving computer science problems has been to ask myself, “How are you solving this?” Consciously asking myself this somehow sheds light on things. One would think that asking yourself that question should’ve been automatic. Not sure what’s wrong with my brain though.