Instructor: Shashank Srivastava
ssrivastava@cs.unc.edu
When: Wednesdays
Time: 2:30 pm - 5:00 pm
Where: SN 115
Office hours: Wed 1:30-2:30pm, SN235
Announcements:
In light of the Coronavirus situation, class logistics have changed (Moving forward, classes will be through online meetings on Zoom)!
Sign ups to present papers live here.
Online feedback form for paper presentations.
Course Description:
This is a graduate seminar course on grounded language understanding, which focuses on exploring the interface between human language, perception and the external world. This is related to sub-areas of Artificial Intelligence, NLP, cognitive science and robotics. We will explore exciting recent research as well as seminal papers in this domain. The central topics that we will discuss include:
Language acquisition in humans and machines
Grounded language learning from interaction and dialogue
Instruction following for navigation and game-playing robots
Language grounding to symbolic representations and databases
Language driven policy learning, planning and control
Models for compositional and hierarchical grounding of language
Natural Language Generation in pragmatic contexts
Neuro-symbolic and multi-modal methods
The course will center around reading and presenting interesting papers, participating in class discussions, and doing a research project.
Please email me if you have any questions.
Prerequisites:
This course assumes previous experience with Machine Learning and NLP. If you have taken similar classes before, feel free to take the course. If you are uncertain, you should email me.
Grading (tentative):
Paper presentations: 20%
Paper summaries (due at beginning of each class, starting Jan 15): 20%. Each week you will be expected to submit a short critical summary on papers that are to be discussed in the class.
Project presentations and reports: 45%
Initial proposal and Midterm presentation (15%)
Final presentation and report (30%)
Class participation and discussion (in class or on Piazza): 15%
There will be no exams
Class Schedule (tentative):