Course description
This course explores the physical and cognitive dimensions of humanoid robots, robots that seek to mimic humans both in terms of their form and function. The course will introduce students to artificial intelligence (AI) and Machine Learning (ML) algorithms that enable robots to mimic human actions and behavior. The course will focus on the cognitive dimension of humanoid robotics which includes algorithms pertaining to scene/context understanding, human emotion recognition and understanding, humanoid behavior generation, and out-of-distribution-learning, to name a few. Enabling technologies such as meta-humans/meta-environments will be studied in order to understand the role of AI-driven physics simulations in accelerating the learning rate of humanoid robots.
The course will explore the technical, ethical, and philosophical dimensions of what it means for a robot to mimic human form and function and how these capabilities may shape future human-machine co-creation and interactions. This is a primarily project-based course where students will work in teams to explore the critical dimensions of humanoid robots.
Learning objectives
Humanoid
- Motivation for embodying the human form for robotics
- Imitating human behavioral characteristics (gait, facial expressions, etc.)
- Intelligence transcending embodiment (e.g., virtual reality for controlling virtual humanoids)
Cognition
- Learning without an explicit ground truth/class variable
- Making causal inferences/hypotheses generation and testing
- Learn how to evaluate the quality of AI models and the level of bias
Prerequisites
None, although students are expected to have completed a first-level graduate course in Machine Learning and to be able to program in Python.
Faculty
Conrad Tucker