CIS 640: Advanced Topics in Software Systems:
Data-Driven IoT/Edge Computing
Spring 2020
- Instructors:
- Class: Tuesday, Thursday 10:30am-11:45 am, via Zoom (formerly Towne 315)
Prerequisite
Working knowledge in data analytics and programming; permission by instructor
Description
This course is to explore selected topics in data-driven IoT/Edge Computing. We are currently witnessing a technological paradigm shift, in which the IoT systems are increasingly deployed in society. This course is to study emerging paradigms in IoT/Edge Computing and to learn how to develop data-driven applications that can harness the power of the IoT/Edge computing. For application domains, the course will target connected medical devices, smart home for aging, and connected automotive systems. Topics to be covered include data processing and learning for IoT computing, the Internet of medical things, connected vehicles, anomaly detection. The course will require a significant term project in connected health or connected automotive domains.
Learning goals for students:
- Understand, discuss, and apply state-of-the-art techniques in data analysis and learning for IoT.
- Develop data-intensive applications in medical, automotive, or other IoT domains.
Student responsibilities
Students are required to complete readings, participate in class discussions, and present on technical topics in class. Students are required complete a term project with significant design, implementation, written report, and demo/presentation.
Presentation. Each student will be required to give at least one presentation in class. Students can propose to present on a topic of interest, or be assigned one by the instructors. For each presentation, a student sends 2 reading questions and 1 discussion question/topic. The timeline for presentations (deadlines are at midnight EST):
- 2+ weeks before presentation: agree on the presentation topic and the reading material for the rest of the class.
- 1 week before presentation: send the instructors 2 reading questions (based on the reading material) and 1 discussion question (based on the reading material and the presentation).
- 3 days before presentation: send the instructors a slide deck for feedback.
- Presentation: 50-60 minutes long, with a heavy focus on the background and educating the audience about the technical material. The remaining time is left for a discussion.
Reading questions. The day before each class (deadline: midnight EST), the students are required to provide short answers to 2 reading questions (via Canvas). The questions will be sent at least 2 days prior to the deadline.
Class participation. Students are expected to engage with the presented material by asking and offering insights. Each student presentation is followed by a free-form discussion.
Term project. Projects can be done individually or in a team of 2 students, with the latter having an larger scope. Students should find a significant application of the learned IoT-related techniques in an analysis and/or development project. The expected term project timeline:
- By February 21: the initial idea and team composition communicated & discussed with instructors.
- By March 16: a proposal draft with the initial exploratory/setup tasks done.
- By late March: if/as requested by instructors, meet with them and discuss the project.
- By April 28 or 30: demo/presentation completed in class and a draft of project report sent.
- By May 9 11:59 pm: final edits to the report completed.
Tentative Schedule
- Week 1: Introduction to IoT: challenges and opportunities
- Week 2: Medical IoMT and decision-support applications for healthcare; technical background
- Week 3: Time series analysis and tools
- Week 4: Connected automotive domain; motifs detection
- Week 5: Confidence evaluation and calibration
- Weeks 6-13: student presentations TBD
- Week 14: Project presentation/demo
See the actual schedule here.
Reading Materials
- Outlier Analysis (2nd edition) by Charu C. Aggarwal, Springer, 2016 (link).
- Time Series Analysis and Its Applications with R Examples by Robert H. Shumway, David S. Stoffer. 4th Edition, Springer 2016. (link).
- A collection of papers on IoT, edge computing, anomaly detection, confidence calibration, data generation, connected health, connected vehicles, and IoT middleware/resource management.
Grading
Grading will be based on the quality of deliverables and class participation:
- 50%: Term project execution and report. Focus: interesting problem and technical depth.
- 15%: Term project demo/presentation. Focus: cohesive/correct technical presentation and providing valuable insights.
- 15%: Paper presentation (including prep of reading questions). Focus: audience-friendly presentation and educational value to the class.
- 10%: Answering reading questions. Focus: comprehension of reading material.
- 10%: In-class participation (questions and discussion). Focus: interesting and engaging discussion.