On AI, from your AI
Monday, February 6th, 2006 by Umar SyedI’m Umar Syed, another one of your preceptors. My main reasearch interest is artificial intelligence (AI). I thought I’d briefly tell you what I like about AI, especially the CS approach to it.
AI is studied in many other disciplines, such as psychology, philosophy and linguistics, and in those fields, its study often revolves around deep questions such as ”Can computers think?”. Unfortunately, this line of inquiry can get bogged down in sematic questions about what the terms “computer” and “think” really mean. Computer scientists neatly sidestep these issues by making all the definitions strictly operational. In other words, we describe a task that represents an example of a computer thinking, e.g. identifying an email as spam or non-spam, and then prove that computers can think simply by exhibiting a computer program that (usually) accomplishes the task.
While this approach may seem too prosaic, it often leads to interesting theoretical questions, such as about the intrinsic “learnability” of various tasks. And these questions can be investigated in a very concrete way by computer scientists.
I hope you enjoy the course, and I’m looking forward to this semester. Feel free to ask me for help on anything.