CHI 2009 Workshop

Developing Shared Home Behavior Datasets to Advance HCI and Ubiquitous Computing Research



More information on this workshop is available from the main web site at http://web.mit.edu/datasets

Below you will find links to all of the accepted position papers for the workshop. Authors confirmed as attending have been highlighted. A representative for each paper will be asked to select two of the followup questions listed below and present a brief 5-minute (total for both questions) response during the workshop. Please use the PowerPoint template available for download at http://web.mit.edu/datasets/downloads/WorkshopSlides-YourNameHere.ppt when preparing your slides.

Prior to the workshop, it is recommended that all participants familiarize themselves with the positions of the other participants. This will help promote more targeted discussions among researchers with complementary interests.

Netcarity Multimodal Data Collection
Alessandro Cappelletti, Bruno Lepri, Nadia Mana, Fabio Pianesi, Massimo Zancanaro
University of Trento, Italy
  1. How would low-bandwidth sensing (e.g. passive infrared motion detection, object movement sensors, RFID) complement the methods used in the NETCARITY project?
  2. How could the data on target behaviors in NETCARITY be used to improve segmentation of activities in recordings of ongoing natural behavior?
  3. How might recordings from the high density microphone arrays used in this project provide value to other researchers? Would this justify the cost?

Collecting and Disseminating Smart Home Sensor Data in the CASAS Project
D.J. Cook, M. Schmitter-Edgecombe, Aaron Crandall, Chad Sanders, Brian Thomas
Washington State University
  1. What are some of the advantages and disadvantages of using a database model for storing sensor data?
  2. How could the CASAS dataset be scaled to support high-bandwidth datastreams such as audio and video recordings?
  3. What lessons did you learn from the erroneous activity datasets that could be used to improve analysis of sensor data from natural home environments?

Gathering Datasets for Activity Identification
Lorcan Coyle, Juan Ye, Susan McKeever, Stephen Knox, Matthew Stabeler, Simon Dobson, Paddy Nixon
University College Dublin, Ireland
  1. In what ways could datasets from home and workplaces could be used to improve each other? Would the sensing needs differ?
  2. In the CASL dataset, how might overlapping data from the UbiSense locator, pressure mats, and Bluetooth spotters be used to good advantage?
  3. Describe the concept of "bootstrapping" datasets for new users and discuss how this might be done efficiently for large, long-terms datasets.

Detailed Human Data Acquisition of Kitchen Activities: the CMU-Multimodal Activity Database (CMU-MMAC)
Fernando de la Torre, Jessica Hodgins, Javier Montano, Sergio Valcarcel
Carnegie Mellon University
  1. How might body motion capture be most practically implemented in a natural home environment?
  2. Given the choice between high resolution (1024x768, 30fps) or high frame rate (640x480, 60fps) video, which do you think would be more beneficial to the greatest number of researchers?
  3. Explain how different types of microphones pickup (cardiod or omnidirectional) affect the audio recordings. What would you consider an optimal layout for "generic" home activity?

HCI is Different: We Need Something Else
James Fogarty
University of Washington
  1. Describe one way researchers might solicit and distill community input before undertaking a data collection project?
  2. Do you believe the existence of shared datasets might provoke interesting/useful research questions that might not otherwise be asked? Describe why/why not?
  3. Discuss why you believe it is or is not possible to collect general-purpose shared datasets on home behavior.

Experience-based requirements on data sets for user modeling in the home
(1) Johan Hjelm, Toshikane Oda, (2) Sitorius Timothy Lawrence, Hua Si, Hiroyuki Morikawa, Shunsuke Saurwatari
(1) Ericsson Research, (2) University of Tokyo
  1. How reliably do you think the context inference model of Synapse could be applied to data collected from other sensor infrastructures?
  2. How might user intention modeling be implemented to facilitate the annotation of activities in shared datasets?

Open Home: Approaches to Constructing Sharable Datasets within Smart Homes
(1) Xin Hong, Chris Nugent, Dewar Finlay, Liming Chen, Richard Davies, Haiying Wang, Mark Donnelly, Huiru Zheng, Maurice Mulvenna, (2) Josef Hallberg, Kåre Synnes
(1) University of Ulster, UK, (2) Luleå University of Technology, Sweden
  1. What are the advantages and disadvantages of using XML vs SQL for collecting and sharing home sensor data?
  2. What is your vision for how Semantic Smart Home architecture could help other researchers studying home behavior?
  3. How do you believe open standards for smart home datasets could be best developed and promoted within the research community?

Sensing Room and Its Resident Behavior Mining
Taketoshi Mori, Masamichi Shimosaka, Akinori Fujii, Kana Oshima, Ryo Urushibata, Tomomasa Sato, Hajime Kubo, Hiroshi Noguchi
University of Tokyo
  1. Describe your schema for annotating behaviors in Sensing Room. What were strengths and weaknesses of the annotation procedure?
  2. What challenges would you anticipate for installing magnetic motion capture in natural homes? What alternative strategies for capturing bodily motion would you consider?
  3. How might you use data collected by other researchers (with different sensor types) to advance your work on behavior modeling?

Homebase: Developing a Corpus of Domestic Network Usage
(1) Tom Rodden, Paul Tennent, Andrew Crabtree, (2) Matthew Chalmers, (3) Beki Grinter, Keith Edwards
(1) Nottingham University, (2) Glasgow University, (3) Georgia Institute of Technology
  1. What ethnographic methods would you encourage other researchers to employ when collecting sensor datasets in homes?
  2. What sensor data types do you feel might be problematic for use in DRS. Why or why not?
  3. How might the homework sensor network plane be extended to data type and acqusition/storage strategies used by other researchers

A sensing and annotation system for recording datasets in multiple homes
T.L.M. van Kasteren, B.J.A. Kröse
University of Amserdam
  1. How might you change the realtime voice-activated annotation procedure to reduce the burden on the user?
  2. Describe how the activities to be annotated were selected. How and why would you change this list in future data collection?
  3. Discuss the feasibility of using a Bluetooth headset to automatically record ongoing audio in the home environment

Simulating Events to Generate Synthetic Data for Pervasive Spaces
(1) Andres Mendez-Vazquez, Abdelsalam (Sumi) Helal, (2) Diane J. Cook
(1) University of Florida, (2) Washington State University
  • Presented by Aaron Crandall

WARD: A Wearable Action Recognition Database
(1) Allen Y. Yang, Ruzena Bajcsy, (2) Philip Kuryloski
(1) University of California, Berkeley, (2) Cornell University
  1. Discuss the pros and cons for wearable vs. environmental sensing to measure human action. How are they complementary?
  2. Describe a few of the practical outcomes of sharing a public dataset (e.g. publications, inventions, etc.)
  3. Based on feedback from WARD users, how would you improve future datasets (e.g. data types and formats, tutorials, analysis tools, etc.)