16:198:549 - Artificial Intelligence for Visual Computing
- Course Number: 16:198:549
- Course Type: Graduate
- Semester 1: Spring
- Credits: 3
This course will introduce students to recent trends in Artificial Intelligence for applications to visual computing. Visual computing is a sub-field of Computer Science which broadly encompasses areas including but not limited to: Computer Graphics, Computer Vision, Virtual and Augmented Reality, and Interactive Techniques. As part of this course, students will therefore be exposed to concepts in Artificial Intelligence which directly apply to one or more of these topic areas.
- Category: B (M.S.), B (Ph.D.)
- Prerequisite Information:
Proficiency in one or more of the following areas:
- Artificial Intelligence
- Computer Graphics
- Computer Vision
- Machine Learning
Sample Topics include but not limited to:
Week 1: Motivation and Introduction
The students will be introduced to the exciting new developments in Visual Computing and how Artificial Intelligence stands to disrupt these application domains. Students will be briefly introduced to different topics including Character Animation, Digital Storytelling, Computer-Aided Design, and Virtual and Augmented Reality.
Week 2 -3: Advancements in Animating Autonomous Virtual Humans
The students will be introduced to recent advancements in Artificial Intelligence (Motion Planning, Machine Learning) for animating humanoid virtual characters. Beyond character animation, we will also describe methods for synthetic visual and auditory perception, goal-directed collision avoidance, navigation, and multi-actor behavior authoring. This series of lectures will conclude with an open question of what’s next? How can these enabling technologies be deployed in the application domains described in Week 1 ?
Week 4: Models of Personality, Mood, Emotion, Behavior in Autonomous Virtual Humans
The students will be introduced to well-established models of personality (e.g., the OCEAN personality model), mood, and emotion, and behavior, and how they may be integrated into established pipelines for animating virtual humans.
Week 5 - 6: Deep Reinforcement Learning for Humanoid Agents
This lecture will describe recent trends in deep reinforcement learning (specifically Deep Hierarchical Reinforcement Learning and Multi-Agent Reinforcement Learning) for learning optimal controllers for humanoid agents.
Week 7 - 8: Recent Trends in Crowd Simulation
We will provide an overview of the most recent trends in simulating human crowd behavior. Topics will include social force models, reciprocal velocity obstacles, and hybrid methods.
Week 9 – 10 : Machine Learning for Crowd Modeling:
We will describe recent advancements in learning how to predict long-term crowd activity patterns from observations of real crowd behavior.
Week 11 - 12: Cognitive Modeling for Human Wayfinding
In contrast with expert-based models and machine learning techniques, we will also investigate how cognitive principles can be utilized in the modeling of autonomous agents – focusing on human wayfinding as an application domain.
Week 13: AI for Digital Storytelling
We will review recent advancements in machine learning for multi-modal story comprehension, and synthesizing animated stories. We will also describe how passive animated content can be enhanced to create immersive, interactive stories.
Week 14: AI for Computer-Aided Design
We will introduce concepts in Artificial Intelligence for computer-aided design applications, focusing on architectural design as a case study.
Week 15: Virtual Reality for Serious Games
We will describe recent advancements in Virtual Reality (esp. shared online virtual reality) and how these platforms can be used for serious games and the study of collective human behavior.
- Expected Work: Reading research papers and final project.
- Exams: None
- Learning Goals:
- Develop an understanding of recent AI trends for Visual Computing applications
- Gain both conceptual and practical exposure to different Visual Computing application domains listed above