IS 698/800: Smart Home Health Analytics — Spring 2020

Times: Tuesday 7:10pm – 9:40pm

Location: Information Technology 104

Instructor: Nirmalya Roy

Instructor’s Office Location and Hours: ITE 421 Thursday 9:00 am – 12:00 pm, or by appointment

Instructor’s Email: nroy at umbc dot edu

Course Descriptions: This course will examine different machine learning methodologies, toolsets, wearable and ambient systems that are warranted to recognize, discover, and learn the human activities, behaviors, and their profound impact on human health and wellness in smart home environments. Human activity recognition is a growing field of research with its far-reaching impact on proactive health care, physical fitness, and healthy living. We will discuss different machine learning algorithms, evaluation metrics, and their application to human activity and behavior recognition. A special emphasize will be given to the techniques with respect to a variety of emerging smart environment contexts spanning across monitoring activities of daily living (ADLs), physiological signals (heart rate, galvanic skin response) and psychological behaviors (emotion, stress, depression, agitation) etc. These human signals can be sensed and measured using wearable and IoT (Internet-of-Things) technologies such as smart phones and wristwatches to keep people informed about their activity, life-style, health, mood, behavior, and surrounding. Students are required to do a group or individual project. The tools necessary to develop the project will be reviewed and made available in class, but students are expected to be comfortable with basic use of software, smart phones and smart watches and have an interest in using technology for smart home health monitoring, assessment, and intervention.

Course Objectives: The purpose of this course is to provide a comprehensive foundation to apply machine learning, data science, and statistical learning methodologies in solving problems in real life applications such as smart home healthcare. This course should enhance students’ reasoning, problem-solving and modeling abilities, particularly in dealing with data science problems. More specifically, the course has the following objectives:

  • Familiarize students with the concepts and applications of machine and statistical learning techniques to solve real life data science problems.
  • Teach students how to think and formalize smart environments research problems with real data and what computational techniques to apply.

We will not assume any background beyond high school level mathematics and familiarity with programming concepts. However, students are expected to spend time in learning the concepts in this course, many of which will be covered in detail.

Course Topics:

  • Gerontechnology
  • Supervised Learning
  • Bayesian Decision Theory
  • Parametric Methods
  • Decision Theory
  • Machine Learning Evaluation
  • Machine Learning Toolkits
  • Activity Learning
  • Smart Home Health Technologies
  • Functional and Behavioral Health Assessment
  • Qualitative and Quantitative Clinical Health Assessment Tools
  • Applications

Course Prerequisites: Consent of the instructor

Instructional Materials: Given that this is an interdisciplinary IS course cross cutting machine learning, data science, Internet of Things, smart home technologies and clinical psychology, there are no single textbook that fully cover the integrated aspects of the course material. At the first half of the course the instructor will discuss the core methodologies behind the course and in the second half the students will be reading various research articles to assimilate the emerging trend in the subject matter of smart home health analytics. A list of research articles will be made available at the course webpage. There are a few books that are relevant to this course are listed next.

Recommended Textbooks (Optional):

Course Project: Students are required to do a group project (individual or two students in each group) as part of this course work. Each team is required to propose a study hypothesis, collect their own datasets over a period of time using smartphones and smartwatches for a specific smart environments application as approved by the instructor and use of appropriate data analytics and machine learning methodology to recognize and evaluate the context of the users.

Reading, Writing, and Oral Assignments: Each student will be required to participate in the research reflection group discussion exercise in class, present individually one research paper and the group research project at the end of the semester. Each student is also required to write three critiques on three distinct research papers being presented by the fellow students in the class.

Week

Date

Topic

Handout/ Assignment

Due

Notes

1

1/28

  • Course Introduction
  • Review of Probability & Statistics


2

2/4

  • Machine Learning Introduction
  • Weka Tutorial


Introduction ML (Textbook IML: Chapter 1, Section 1.1 and 1.2)

Weka Tutorial

Example Datasets

3

2/11

  • Supervised Learning
  • Research Logistics & Reflection


HW1

Class Cancelled

4

2/18

  • Supervised Learning
  • Research Logistics & Reflection



Supervised (Textbook IML: Chapter 2, Sections 2.1, 2.3, 2.4, 2.5, 2.7, and 2.8)

Research Logistics

Research Reflection

Research Reflection Sign Up Sheet

5

2/25

  • Bayesian Decision Theory
  • Research Reflection Discussion


Bayesian (Textbook IML: Chapter 3, Sections 3.1, 3.2 and 3.6)

6

3/3

  • Parametric Methods
  • Research Reflection Discussion


7

3/10

  • Parametric Methods (Contd..)
  • Research Reflection Discussion




8

3/17

Spring Break




9

3/24

  • Decision Trees


10

3/31

  • Decision Trees (Contd..)
  • Evaluation
  • Research Paper Presentation [Zahid Hasan]


11

4/7

  • Evaluation
  • Research Paper Presentation [Emon Dey]


12

4/14




14

4/28

Research Paper Presentation

Exam Review



15

5/5

Exam




Research Paper Presentation Schedule:

(See above)

Research Papers:

In this course we will be discussing various papers recently published in ACM/IEEE MobiSys, SenSys, BuildSys, PerCom 2019. Please check the following papers.

Smartphone Sensing

VitaMon: Measuring Heart Rate Variability Using Smartphone Front Camera; Sinh Huynh (Singapore Management University); Rajesh Balan (Singapore Management University); JeongGil Ko (Yonsei University); Youngki Lee (Seoul National University); ACM SenSys 2019 Zahid Hasan

SenseHAR: A Robust Virtual Activity Sensor for Smartphones and Wearables; Jeya Vikranth Jeyakumar (University of California, Los Angeles); Lianghzen Lai (ARM); Naveen Suda (ARM); Mani Srivastava (University of California, Los Angeles); ACM SenSys 2019 Jason Jozwiak

ALT: Towards Automating Driver License Testing using Smartphones; Akshay Nambi (Microsoft Research); Ishit Mehta (Microsoft Research); Anurag Ghosh (Microsoft Research); Vijay Lingam (Microsoft Research); Venkat Padmanabhan (Microsoft Research); ACM SenSys 2019 Sabrina Mamtaz Nourin

HyperSight: Boosting Distant 3D Vision on a Single Dual-camera Smartphone; Zifan Liu (Shanghai Jiao Tong University); Hongzi Zhu (Shanghai Jiao Tong University); Junchi Chen (Shanghai Jiao Tong University); Shan Chang (Donghua University); Lili Qiu (University of Texas at Austin); ACM SenSys 2019

Deep Learning

MetaSense: Few-Shot Adaptation to Untrained Conditions in Deep Mobile Sensing; Taesik Gong (KAIST); Yeonsu Kim (KAIST); Jinwoo Shin (KAIST); Sung-Ju Lee (KAIST); ACM SenSys 2019

Moving Target Defense for Embedded Deep Visual Sensing against Adversarial Examples; Qun Song (Nanyang Technological University); Zhenyu Yan (Nanyang Technological University); Rui Tan (Nanyang Technological University); ACM SenSys 2019

Neuro.ZERO: A Zero-Energy Neural Network Accelerator for Embedded Sensing and Inference Systems;Seulki Lee (University of North Carolina at Chapel Hill); Shahriar Nirjon (University of North Carolina at Chapel Hill); ACM SenSys 2019

DeepAPP: A Deep Reinforcement Learning Framework for Mobile Application Usage Prediction; Zhihao Shen (Xi'an Jiaotong University); Kang Yang (Xi'an Jiaotong University); Wan Du (University of California, Merced); Xi Zhao (Xi'an Jiaotong University); Jianhua Zou (Xi'an Jiaotong University); ACM SenSys 2019 Millie Zeng

Infrastructure Sensing

WISDOM: Watering Intelligently at Scale with Distributed Optimization and Modeling; Daniel A. Winkler (University of California, Merced); Alberto E. Cerpa (University of California, Merced); ACM SenSys 2019 Hemlata Kohin

Caesar: Cross-camera Complex Activity Recognition; Xiaochen Liu (University of Southern California); Pradipta Ghosh (University of Southern California); Oytun Ulutan (University of California, Santa Barbara); B.S. Manjunath (University of California, Santa Barbara); Kevin Chan (ARL); Ramesh Govindan (University of Southern California); ACM SenSys 2019

SmrtFridge: IoT-based, User Interaction-Driven Food Item & Quantity Sensing; Amit Sharma (Singapore Management University); Archan Misra (Singapore Management University); Vengateswaran Subramaniam (Singapore Management University); Youngki Lee (Seoul National University); ACM SenSys 2019 Prachi Mardhekar

WideSee: Towards Wide-Area Contactless Wireless Sensing; Lili Chen (Northwest University); Jie Xiong (UMass Amherst); Xiaojiang Chen (Northwest University); Sunghoon Ivan Lee (UMass Amherst); Kai Chen (Northwest University); Dianhe Han (Northwest University); Dingyi Fang (Northwest University); Zhanyong Tang (Northwest University); Zheng Wang (Lancaster University); ACM SenSys 2019

A Closer Look at Quality-Aware Runtime Assessment of Sensing Models in Multi-Device Environments; Chulhong Min (Nokia Bell Labs); Alessandro Montanari (Nokia Bell Labs); Akhil Mathur (Nokia Bell Labs and University College London); Fahim Kawsar (Nokia Bell Labs and TU Delft); ACM SenSys 2019

Smart Cities

Thermal Comfort

Sample Research Project Reports:

David Welsh and Nirmalya Roy. Smartphone-based Mobile Gunshot Detection, In Proceedings of the 13th Workshop on Context and Activity Modeling and Recognition (CoMoRea’17), co-located with PerCom, March 2017. [pdf]

H M Sajjad Hossain, Md Abdullah Al Hafiz Khan, and Nirmalya Roy. SoccerMate: A Personal Soccer Attribute Profiler using Wearables, in Proceedings of the 1st IEEE PerCom International Workshop on Behavioral Implications of Contextual Analytics (BICA), co-located with PerCom, March 2017. [pdf]

Abu Zaher Md Faridee, Sreenivasan Ramasamy Ramamurthy, H M Sajjad Hossain, and Nirmalya Roy. HappyFeet: Recognizing and Assessing Dance on the Floor, in Proceedings of the ACM 19th International Workshop on Mobile Computing Systems and Applications (HotMobile), Feb. 2018 [pdf]

Varun Mandalapu, Lavanya Elluri and Nirmalya Roy. Developing Machine Learning based Predictive Models for Smart Policing, in Proceedings of the 1st IEEE International Students Workshop on Smart Computing (SmartStudents), co-located with SmartComp, pp. 1-6, Washington D.C., June 2019

Research Projects:

Team No

Team Members

Topic/Title

Devices/ Datasets

1

Jason Jozwiak

Monitoring the Correctness of Rehabilitation Exercises for Physical Therapy

Actigraph, Android smartphone

2

Zahid Hasan

Contact-less Physiological Health monitoring

RGB and Thermal Camera

3

Prachi Mardhekar

Signature of Walking: Can it be used for Identifying Criminal Activities?

UCI user identification from walking activity datasets

4

Emon Dey

Detecting Mental Disorders with Substance and Non-Substance Users During COVID-19

Emotiv Epoch EEG headset

5

Hemlata Kohin

Monitoring Air Quality

Air Quality Dataset from UCI repository

6

Pretom Roy Ovi

Heart Disease Prediction

VA medical center & Cleveland clinic

7

Masud Ahmed

Detection of Crack in Concretes using Drones

Concrete crack datasets

8

Sabrina Mamtaz Nourin

Which Health App to Use? Comparative Study between Different Health Apps

Health Apps

9

Millie Zeng

Impact of Stay-at-home Quarantine Order on Mental Health

Activity datasets

10

Saydeh Karabatis

Impact of Administrative Actions on COVID-19 Cases

COVID-19 cases in MD