Academics
  /  
Courses
  /  
Descriptions
COMP_ENG 395, 495: Embedded Artificial Intelligence


VIEW ALL COURSE TIMES AND SESSIONS

Description

CE 395/495: Embedded Artificial Intelligence is an upper-level course on how artificial intelligence is developed, deployed, and used in the physical world, often on platforms constrained in form factor, battery, and computational resources. The topics covered in this course center on the theme of bringing intelligent perception, analysis, and understanding through artificial intelligence away from data centers and the cloud and onto mobile and edge devices that are often directly interacting with the physical environment or in the palms of our hands. Examples of such platforms include smartphones, wearables, drones, robots, and much more.
 
This course is research-oriented and based on papers and research findings from top-tier venues. Students will learn key principles, research trends, and the state-of-art in mobile and embedded intelligence. Additionally, students will learn and gain experience in paper reading, paper writing, public presentation, and carrying out original research through weekly paper readings, paper presentations, critiques, class discussions, and a class project on an original idea. For the class project, students will ideally generate (or make significant progress) and present publishable results by the end of the term.

INSTRUCTOR: Prof. Stephen Xia

GRADING: 

  • Project: 60%

  • Paper Presentation: 20%

  • Homework: 10%

  • Participation: 10%

COURSE TOPICS: 

  • Sensing, Perception, and Actuation 
  • TinyML
  • AI Accelerators
  • Mobile, Edge, and Distributed Computing
  • Federated Learning
  • On-Device Inference and Learning
  • Continual, Multimodal, and Transfer Learning
  • Physical Knowledge Informed AI
  • Applications and Areas of Impact (security/privacy, human-computer interaction, communications and networking, AR/VR, etc.)