Tutorials

Tutorials are intended to provide fundamental knowledge and experience on topics related to ubiquitous and wearable computing through interactive discussions and hands-on exercises. Tutorials also encourage participation of young researchers, including students, and welcome people with different backgrounds and technical skills. We have two full-day tutorials in parallel with workshops.

This page will be updated with more details as the date approaches.

Dates

• Sept 7, 2015: T1 Application-Driven Research: Building Systems to Evaluate Your Ideas in the Wild [a few seats left]
• Sept 8, 2015: T2 Bridging the Gap: Machine Learning for UbiComp [sold out]

Tutorial Programs

T1: Application-Driven Research: Building Systems to Evaluate Your Ideas

Organizers: Khai Truong (University of Toronto)

Abstract:

As Weiser pointed out, "applications are of course the whole point of ubiquitous computing." Thus, it is not surprising that many researchers and designers have and continue to explore different ways that ubicomp can be used to help people in their daily lives. Sometimes, the development and evaluation of a novel application itself can potentially be a research contribution. However, application-driven research can also be about using the process of building an application to test an idea with users in the wild. The knowledge that can be learned through the process of building the system and/or evaluating that system with users can be research contributions as well. In this tutorial, we will discuss when, why and how to use applications as vehicles for answering research questions. We will review some examples of previous application-driven research projects and consider what each work's contribution is and how an application was used to obtain that contribution. We will also discuss the challenges and criteria to using an application-driven research approach.

Short Bio:

Khai Truong is an associate professor in Department of Computer Science at the University of Toronto. Khai received a Ph.D. degree in Computer Science and a Bachelor degree in Computer Engineering with highest honors from the Georgia Institute of Technology. He has been an active ubicomp researcher for over 15 years. His research interest lies at the intersection of human-computer interaction (HCI) and ubiquitous computing. Khai has served as the General Chair for Pervasive 2007, Program Co-Chair for Ubicomp 2010, and area editor for SIGMOBILE GetMobile (formerly MC2R). More information about him can be found at: http://www.cs.toronto.edu/~khai

T2: Bridging the Gap: Machine Learning for UbiComp

Organizers: Mayank Goel (University of Washington), Nils Hammerla (Newcastle University), Thomas Ploetz (Newcastle University) and Anind K. Dey (CMU)

Abstract:

The ever-increasing number of sensors are helping our devices become more useful and context-aware.  However, these sensors do not always generate readily usable or actionable data. In many cases, a machine learning-based inference systems converts the low-level sensor data into high-level inferences that can be directly used by a user or another intelligent system. Machine learning has shown considerable promise in a number of ubiquitous computing applications and is a very popular approach. However, it requires a certain amount of expertise or domain-knowledge. In the absence of such knowledge, the novice, or even intermediate user, sometimes, is prone to use machine learning algorithms as a black-box. This approach often leads to frustration and sometimes even flawed results. The goal of this tutorial is to provide an explanation of how to use machine learning effectively for sensing applications. We will explain the different stages of a machine learning-based approach through different case-studies and examples. We will discuss and provide solutions to some common pitfalls that a new student of sensor-based machine learning is prone to encounter. Through this, we hope that researchers and practitioners will learn how to develop a better understanding of their sensor data and use machine learning as a tool-box and not a black-box.

Short Bio:

Mayank Goel is a Ph.D. candidate in the Ubiquitous Computing (Ubicomp) Lab at the University of Washington. His research involves using mobile phone sensors for ubiquitous computing applications. In particular, he focuses on leveraging new sensing solutions for health and creating novel human-computer interaction techniques.He was awarded the Microsoft Research PhD Fellowship in 2014. He received his M.S. degree in computer science from the Georgia Institute of Technology in 2009, where he specialized in ubiquitous computing and human-computer interaction. He is currently pursuing his Ph.D. in Computer Science and Engineering at the University of Washington in Seattle and is advised by Prof. Shwetak Patel and Prof. Gaetano Borriello. More info at www.mayankgoel.com.