Hacker News new | ask | show | jobs
by psyklic 4466 days ago
AI is a very broad field. Much of the academic "classical" AI requires some advanced math to understand in depth.

A fun place to start is video game AI because it is visual. For example, path-finding algorithms, the ghost AI in pacman, chess/checkers, bots which play video games.

Another fun visual subfield is computer vision. If you have a webcam, you could get OpenCV and play around with recognizing faces and motion detection.

If you are really into math, I always think that attempting an automated theorem prover would be fun. It is an up and coming science (in fact there was a HN post today about it), and I bet there are a lot of ad hoc approaches you could take if you choose a specific area of math.

If you want something very practical, nowadays search/recommendations/etc. are important on the web. There are some high-rated books on Collective Intelligence that are more practical-minded than the academic AI classics. And mining the web for a dataset could be a fun project.

As for the best language, it depends on which subfield you want to pursue. Scheme/Lisp are probably the most strongly associated with AI. However, in practice they are rarely used. I would say it does not matter, as long as you are motivated :)

1 comments

Hey, thank you! I will look for Recommendations Algorithms and I will try implement the A* Algorithm.

You have some recommendations in math books?