The collected papers and books about artificial intelligence and robotics are the most dominant resource available to investigate the subject in detail. The assumption is, that every subject was described already in the Gutenberg galaxy, the only question is where exactly and how well a certain text was written.
Especially for newbies its important to identify possible entry point to start the journey to AI. Such an entry point are “np hard” and “Np complete” problems which are discussed since the 1990s in common books about algorithms. The idea is that apart from vanilla problems like sorting an array and searching in a string, some advanced challenges are available. These advanced challenges have no known algorithm, or the algorithm is working with so called heuristics.
The good news is, that these np hard problems are a good starting point to investigate the subject of AI. Every AI problem problem is also an np hard problem. That means, AI is trying to solve only the advanced problems not researched in classical computer science, e.g. path planning, motion planning, decision making in games or human to machine communication.
There are some attempts available to solve even np hard problems, e.g. guided search in the state space or evaluation function. Especially evaluation functions are a widely discussed subject within the AI community because they are needed for computer chess. There are hundreds or even thousands of books available which are dedicated only to evaluation functions, so we can say that this subject might be a good starting point to become familiar with AI.
No comments:
Post a Comment