CSCI special topics courses
CSCI 8980 - Topics in Modern Machine Learning
This is a seminar style course consisting of lectures, paper presentation, and interactive discussion of selected advanced topics from modern machine learning. In the Fall 2021 version, we focus on foundational aspects of deep learning, and try to cover major theoretical results and computational insights that people have gathered on deep learning, and highlight open problems.
Registration prerequisites: CSCI 5525 or equivalent. Please review the course website for further registration information.
CSCI 8980 - Perception and Action Planning for Robotics
MW 09:45 AM ‑ 11:00 AM
Instructor: Volkan Isler
Traditional robotics algorithms often require explicit models for the perception/action interface where the perception algorithm is in charge of estimating the model which is passed on to the planning/control module. While this separation allows for independent progress in these two areas, the need for developing an explicit model often limits the general applicability of the resulting systems. In this seminar, we will go over the state of the art in robot algorithms which try to overcome this limitation by coupling perception and action planning.
Registration prerequisites: There are two prerequisites for the course:
- CSCI 5561 or equivalent computer vision class which covers multi-view geometry
- Familiarity with basic planning algorithms including search algorithms and basic reinforcement learning algorithms
CSCI 8980 - Programming Language Foundations in Agda
TTh 11:15 AM ‑ 12:30 PM
This is a seminar course exploring the foundation of programming languages using the proof assistant Agda, a tool that enables you to construct computer-checkable proofs. There are two main learning goals: (1) to learn several basic topics in the theory of programming languages and (2) to learn proof mechanization using Agda. We will closely follow the textbook Programming Language Foundations in Agda written by Wadler, Kokke, and Siek.
CSCI 8980 - Vision and Learning in Interactive Field Robotics
TTh 4:00 PM ‑ 5:15 PM
Instructor: Junaed Sattar
This course will discuss topics on machine perception and reasoning from a computational perspective, with a particular focus on robots in the field working with humans. Robots are intelligent computing engines equipped with sensors, can act on the world and work with and/or alongside their human partners. In this course we will consider the definitional problems for human-robot collaboration in field robotics, and look at how they are being solved in practice and by the research community. Particular focus will be on autonomous underwater co-robots, but general ideas will be applicable to any autonomous co-robots operating in the field. The emphasis is on algorithms, inference mechanisms, interaction design, and behavior strategies although we will also survey the kinds of mechanical and electronic systems that constitute a robot today.
In particular, the course will focus on robotic systems working outdoors, in unstructured environments, as human assistants. Generally, Robotics as a whole along with the integration of hardware systems with computer software and human-robot interaction is a key area of both industrial and research significance. The area also has appealing connections to both theoretical computer science and computer systems engineering.
CSCI 8980 - Trends in Edge Computing
TTh 9:45AM - 11:00AM
Instructor: Jon Weissman
This seminar will explore the new frontier of Edge Computing, the use of computing, communication, and storage resources, at the edges of the network close to data sources and end-users. Topics will span edge applications, architectures, and systems, and the Internet of Things. Students will read and present papers, write paper critiques/blogs, and do a final project in the topic area.
Registration prerequisites: CSCI 5103 (OS) or by instructor permission