CS232 Artificial Intelligence

Gesture recognition in the Google Android phone.

For as long as scientists have been building computers, they have also been trying to make them intelligent. But what does it mean to build intelligent machines? Should they be like the ruthlessly efficient HAL in the "2001: A Space Odyssey" movie, or like the love-needing child robot in Steven Spielberg's "Artificial Intelligence"? Or has Google already become the intelligent machine, by aggregating for years our collective intelligence, so that now it can answer our questions, translate in 53 natural languages, transcribe our phone conversations, recognize gestures, or give us turn-by-turn voice instructions while driving.

Voice recognition in the Google Android phone.

Autonomous vehicles in the DARPA urban challenge.

Personalized recommendations by the online store Netflix.

In this course, we will discuss what makes an agent (human or machine) intelligent and primarily focus on its ability to learn. We will study computational models of human learning (Bayesian learning and handling uncertainty) and the ways they can inform us to build machine learning systems. Finally, we will try to understand and build several machine learning algorithms that have been applied successfully to diverse applications such as: machine translation, speech recognition, information retrieval on the web, autonomous vehicles, or personalized recommendation.

The Python programming language will be used to implement algorithms in the course. No previous knowledge of Python is necessary. Prerequisites: CS230 or by permission of instructor.

The class will meet 3 times a week. More information will be available soon.