Aarhus University Seal / Aarhus Universitets segl

How Implementing Physical Actuation in a Smart Speaker can reveal its Underlying Activities

PhD defence, Tuesday, 20 April 2021. Mirzel Avdic.

2021.04.13 | Mette Nielsen

Date Tue 20 Apr
Time 16:00 18:00
Location

Mirzel Avdic

Digital assistants like Amazon’s Alexa have come a long way in the past few years, and are becoming increasingly present in homes through stand-alone devices like voice-enabled smart speakers (e.g. Amazon Echo). Smart speakers are also used as primary actors in home automation, especially acting as the interface for controlling smart appliances in homes due to their hands-free speech interface. Unfortunately, smart speakers lack transparency about their underlying activities and clarity about how they work. While enthusiasts are considered to be the primary users of such technology, passive users (e.g. guests of smart speaker owners) may be less familiar with such technology, which increases the challenges faced with using smart speakers as primary interfaces to control smart appliances in homes. Moreover, smart speakers rely on smartphones to provide users with more information about how the smart speakers work, however not everyone has configured their smartphones to access such information.

I have investigated alternative approaches to making smart speakers transparent and clear about what they do and how one can control smart appliances. More specifically, I investigated the use of physical actuation: drawing on familiar non-verbal human behaviour (e.g. gesticulation) to make smart speakers’ functions and underlying activities transparent to users.

  

The PhD study was completed at the Computer Science Department, Faculty of Natural Sciences, Aarhus University.

This summary was prepared by the PhD student.

Time: Tuesday, 20 April 2021 at 16:00.
Place: The PhD defence will be held online. To receive a link to the event, please send an e-mail to Henriette Gammelgaard Farup; henriettefarup@cs.au.dk.
Title of PhD thesis: Physical Actuation as an Alternative Approach to the Intelligibility of Smart Speakers
Contact information: Mirzel Avdic, e-mail: miavd18@cs.au.dk, tel.: +45 22230952
Members of the assessment committee:
Assistant Professor Stuart Reeves, School of Computer Science, University of Nottingham, United Kingdom
Professor Dag Svanæs, Computer Science, Norwegian University of Science and Technology, Norway Associate Professor Claudio Orlandi (chairman), Department of Computer Science, Aarhus University, Denmark
Main supervisor:
Principal Research Scientist Jo Vermeulen, Autodesk Research, Canada
Co-supervisor:
Professor Susanne Bødker, Department of Computer Science, Aarhus University, Denmark

Language: The PhD dissertation will be defended in English

The defence is public.
Due to the coronavirus situation, the PhD thesis will not be available for reading at the offices of the PhD administration. Instead, interested parties may send an e-mail to gradschool.nat@au.dk to receive a digital copy of the PhD thesis. Please note, it may in certain cases be necessary to make an individual arrangement to read the PhD thesis.

PhD defence
16882 / i43