This is a graduate-level seminar course where we discuss publications around the topic of Robot Learning. The goal is to familiarize students with fundamental aspects of learning techniques in Robotics, introduce general methodologies of conducting research and engineering in Robot Learning, and inspire independent and critical thinking of future directions in Robotics and AI.
Due to the limited time we have during lectures, the selected readings in the syllabus do not mean to cover all the representative and important papers in the field. Rather, the aim is to point out important and emerging directions for students to freely explore after the class. (Hope you are all excited about the future of Robotics and AI as much as I do!)
All students should be actively involved in all discussions to gain truly immersive experience in this course. In every lecture, we will discuss two selected papers. (Please refer to course syllabus for more information.) The class is capped at 24 students in total. Twelve students will present these two papers. In the next lecture, the rest of the tweleve students will present the next two papers. The process repeats...
Each paper will have 6 students present. Students will select different roles to present the paper. We will talk about more about this process in our first lecture. Here are the roles: "Introduction" (1 student), "Related Works" (1 student), "Method" (1 students), "Experiment" (1 student), "Future Directions" (1 student), "Limitations" (1 student).
Additionally, students are required to conduct a semester-long research project based on the topic they selected and the instructor's approval. Each team can include at most two students. Team with two students are expected to accomplish more than a single-student team. The topics are generally very broad as long as they fit Robot Learning. Students will present their project proposal, mid-term progress, and final results throughout the semester. By the end of the course, students will also need to submit their results as a final report under a conference publication format and code base. The instructor and TA will hold weekly office hours to help guide the progress of the project.
You are provided 48 grace hours for late submission on the final report and code. Late submissions for paper presentation slides and project presentation slides (proposal, mid-term prensentation, and final presentation) are not allowed.
If you have a more serious situation (medical or personal) that is preventing you from completing coursework in a timely manner, and is beyond what can be addressed by grace hours, do not send medical/personal information to the instructor or TA (they are prohibited from seeing this info due to privacy). Instead, graduate students should contact Duke student wellness officer, and undergraduate students should contact their Advising Dean to discuss reasonable accommodations. For longer term situations reach out to the Student Disability Access Office and have them assess your situation and inform the instructor.
You are not allowed to copy other people's code and written reports as your assignment submission. However, you are allowed to use other people's code as part of the module or baseline comparisons, but cite your sources appropriately and be clear about your own contributions. Cheating may lead to the following consequences: zero credit for a specific assignment, failing grade for an entire course, expulsion from the university. More information can be found at Academic Integrity Council at Duke University.