CS 598tar
Sensing in Social
Spaces
Syllabus
Instructor
Tarek Abdelzaher, 4126
Office Hours: Thursday, 2-3pm, 4126
Lecture
Times
Wednesdays and Fridays 12:30-1:45pm, 0216
Description
This is a paper-reading course with a research project component about applications of social/urban/mobile sensing, which are gaining momentum because of their applicability to urban areas and other populated spaces. The proliferation of smart phones and social networks gives rise to large amounts of information becoming available in real-time that can be thought of as "sensing" the physical world. According to the United Nations, presently 54% of the world population live in cities. This percentage will increase to 66% by 2050. Arguably, the most versatile "sensor" in urban areas is the human observer. Collectively, human observers post over 500 million tweets and over 80 million Instagram photos per day, making social media an interesting new "sensor network" for obtaining insights on a variety of events. This course investigates unfolding research challenges and directions in distributed social sensing, overviews the broader landscape of its urban applications, including sustainability, green computing, IoT, and urban cyber-physical systems, discusses common misconceptions, presents the underlying theoretical foundations, and sheds light on related recent technologies and publications. The course includes an experimental project on a social sensing testbed. Projects will involve crowd-sensing across the board from mobile/smartphone-based applications in urban/social spaces to exploitation of Twitter and Instagram, where humans act as pervasive "networked sensors" of current events. Links to CPS with humans-in-the-loop and IoT will be made.
Textbook
The course is
designed to be self-contained. The primary source of information will be the
papers listed in the reading list. Students interested
in a deeper understanding of the analytical foundations of social sensing can
find them in:
Social Sensing: Building Reliable Systems on
Unreliable Data, 1st Edition
By: Dong Wang, Tarek Abdelzaher, Lance
Kaplan
Elsevier
Buy from Elsevier here,
or from Amazon (cheaper) here.
Grading
10% of the grade will be assigned on individuals' class participation, and discussion of research papers.
10% will be assigned on group summaries of selected papers/topics covered (project group members share the same group summary). Roughly one summary will be due per week (but not every week). More credit will be given to groups with creative and original opinions, and on their ability to defend their correctness. Assigned summaries (one per group) are to be submitted to the instructor by e-mail by 9pm of the day before the assigned paper is covered in class. (Please include the words "598 SUMMARY" in the subject of the e-mail, and include the group members and title of the critiqued paper in the body. A discussion of the research paper/topic will ensue in class.
5% of the grade will be assigned on an in-class mid-semester presentation.
5% of the grade will be assigned on an in-class final project presentation.
15% of the grade will be assigned for an open-book take-home midterm.
15% will be assigned for an open-book take-home final.
40% of the grade will be determined by a substantial group course project. This grade includes a project paper, and a project demonstration (to the instructor). The project will implement some innovative social sensing application, service, protocol, or computing environment. Students will work in groups of 3 on the project. Access may be provided to a testbed depending on project topic.