Poster: Toward Context-aware Proactive Conversation for Smart
Speakers
Soowon Kang
sw.kang@kaist.ac.kr
KAIST
South Korea
Heepyung Kim
heepkim@kaist.ac.kr
KAIST
South Korea
Youngtae Noh
ytnoh@inha.ac.kr
Inha University
South Korea
Uichin Lee
∗
uclee@kaist.ac.kr
KAIST
South Korea
ABSTRACT
Despite the increasing popularity of novel conversational services
for smart speakers, the current ones have supported limited proac-
tive interactions. In this work, we focus to understand opportune
moments for proactive conversational interactions in domestic con-
texts. To achieve our goal, we built a speech-based experience sam-
pling device and performed a 1-week feld study with 40 students
living in university dormitories. From 3,572 in-situ experiences,
we identifed eleven categories to fnd contextual features related
to opportune moments. We showed that the key determinants for
interruptibility are relevant to personal contextual factors, user
mobility, and social presence. Through considering the aforemen-
tioned factors, we envision that a smart speaker can intelligently
manage the timing of conversations.
CCS CONCEPTS
· Human-centered computing → Ubiquitous and mobile com-
puting; User interface management systems.
KEYWORDS
Smart Speakers, Conversational Interaction, Interruptibility
ACM Reference Format:
Soowon Kang, Heepyung Kim, Youngtae Noh, and Uichin Lee. 2021. Poster:
Toward Context-aware Proactive Conversation for Smart Speakers. In Ad-
junct Proceedings of the 2021 ACM International Joint Conference on Per-
vasive and Ubiquitous Computing and Proceedings of the 2021 ACM Inter-
national Symposium on Wearable Computers (UbiComp-ISWC ’21 Adjunct),
September 21–26, 2021, Virtual, USA. ACM, New York, NY, USA, 3 pages.
https://doi.org/10.1145/3460418.3479306
1 INTRODUCTION
Recently, top shares in the smart speaker market (e.g., Amazon
Echo [9], and Google Home [11]) started to pay attention to the
∗
This is the corresponding author.
UbiComp-ISWC ’21 Adjunct, September 21–26, 2021, Virtual, USA
© 2021 Copyright held by the owner/author(s).
This is the author’s version of the work. It is posted here for your personal use.
Not for redistribution. The defnitive Version of Record was published in Adjunct
Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous
Computing and Proceedings of the 2021 ACM International Symposium on Wearable
Computers (UbiComp-ISWC ’21 Adjunct), September 21–26, 2021, Virtual, USA, https:
//doi.org/10.1145/3460418.3479306.
proactive services, while the current smart speakers have generally
supported reactive services via voice commands. Although proac-
tive services allow the users to obtain useful information, prior
studies also informed that timing and relevance of the services are
crucial to the user experience [1, 3]. To the best of our knowledge,
the interruptibility studies of smart speakers in home contexts lacks
depth in spite of the popularity of proactive services in smart speak-
ers. In this study, we explained the opportune moments for the
proactive services of smart speakers in diverse activities at home.
To collect in-situ experiences in home environment, we imple-
mented a smart speaker that supports recording user contexts and
responses via a voice-based experience sampling method (ESM).
Our smart speaker asked users to answer the question, łIS now a
good time to talk?ž in yes or no, with detailed contextual informa-
tion. The 3,572 ESM responses were collected from 20 pairs (n=40)
living in dormitories. We examined the responses by conducting the
exit interview to understand the core factors that afects interrupt-
ibility related to home activities. Thus, we identifed the following
factors relevant to the user interruptibility of conversational inter-
actions: (i) personal contextual factors (e.g., engagement, urgency,
psychological/physical states, auditory/verbal channel availability),
(ii) movement-related factors (e.g., entrance/departure behaviors, ac-
tivity switching), and (iii) social presence (e.g., roommate’s current
activity). Our fndings emphasize that the proactive conversation
management is required to the proactive conversational services
for smart speakers.
2 BACKGROUND
Traditional smart speakers have mainly provided reactive services.
The recent commercial smart speakers, including Amazon and
Google, started to look into several proactive services. Amazon
Echo supported friendly reminders (e.g., informing to lock the door)
and ambient sensing to warn the user of safety issues (e.g., win-
dow break) [9]. Google Home can notify the upcoming events that
depend on the user’s Google Calendar [11].
These prior research on smart speakers supported limited chances
to proactively interact with users. One important factor of intelli-
gent services is detecting the appropriate time to open the conver-
sation session. Researchers have been studied the interruptibility
in wide contexts (e.g., smartphone [8], in-vehicle [7]). However,
home environment is more challenging since it is more dynamic
compared to other environments due to multi-user, mobility factors,
38