The challenge of problem-based learning

https://people.eng.unimelb.edu.au/vkostakos/files/papers/ijim19.pdf

Multimodal Data as a means to understand the Learning Experience

Background: Mobile health (mHealth) apps provide an opportunity for easy, just-in-time access to health promotion and self-management support. However, poor user engagement with these apps remains a significant unresolved challenge. Objective: This study aimed to assess the effect of sending versus not sending a push notification containing a contextually tailored health message on proximal engagement, measured here as self-monitoring via the app. Secondary aims were to examine whether this effect varies by the number of weeks enrolled in the program or by weekday versus weekend. An exploratory aim was to describe how the effect on proximal engagement differs between weekday versus weekend by the time of day. Methods: The study analyzes the causal effects of push notifications on proximal engagement in 1255 users of a commercial workplace well-being intervention app over 89 days. The app employs a microrandomized trial (MRT) design to send push notifications. At 1 of 6 times per day (8:30 am, 12:30 pm, 5:30 pm, 6:30 pm, 7:30 pm, and 8:30 pm; selected randomly), available users were randomized with equal probability to be sent or not sent a push notification containing a tailored health message. The primary outcome of interest was whether the user self-monitored behaviors and feelings at some time during the next 24 hours via the app. A generalization of log-linear regression analysis, adapted for use with data arising from an MRT, was used to examine the effect of sending a push notification versus not sending a push notification on the probability of engagement over the next 24 hours. Results: Users were estimated to be 3.9% more likely to engage with the app in the next 24 hours when a tailored health message was sent versus when it was not sent (risk ratio 1.039; 95% CI 1.01 to 1.08; P<.05). The effect of sending the message attenuated over the course of the study, but this effect was not statistically significant (P=.84). The effect of sending the message was greater on weekends than on weekdays, but the difference between these effects was not statistically significant (P=.18). When sent a tailored health message on weekends, the users were 8.7% more likely to engage with the app (95% CI 1.01 to 1.17), whereas on weekdays, the users were 2.5% more likely to engage with the app (95% CI 0.98 to 1.07). The effect of sending a tailored health message was greatest at 12:30 pm on weekends, when the users were 11.8% more likely to engage (90% CI 1.02 to 1.13). Conclusions: Sending a push notification containing a tailored health message was associated with greater engagement in an mHealth app. Results suggested that users are more likely to engage with the app within 24 hours when push notifications are sent at mid-day on weekends.

To Prompt or Not to Prompt? A Microrandomized Trial of Time-Varying Push Notifications to Increase Proximal Engagement With a Mobile Health App.

Social psychology and neuroscience had confirmed that emotion state exerts a significant effect on human communication, perception, social behavior and decision making. With the wide availability of smartphones equipped with microphone, accelerometer, GPS, and other source of sensors, it is worthwhile to explore the possibility of automatic emotion detection via smartphone sensing. Particularly, we focus on a novel research problem that tries to detect the compound emotion (a set of multiple dimensional basic emotions) of smartphone users. We observe that users' self-reported emotional states have high correlation with their smartphone usage patterns and sensing data. Based on the observations, we exploit a feature extraction and selection algorithm to find the most significant features. We further adopt a factor graph model to tackle the correlations between features and emotion labels, and propose a machine learning algorithm for compound emotion detection based on the smartphone sensing data. The proposed mechanism is implemented as an APP called MoodExplorer in Android platform. Extensive experiments conducted on the smartphone data collected from 30 university students show that MoodExplorer can recognize users' compound emotions with 76.0% exact match on average.

MoodExplorer: Towards Compound Emotion Detection via Smartphone Sensing

Emotion, mood, and stress recognition (EMSR) has been studied in laboratory settings for decades. In particular, physiological signals are widely used to detect and classify affective states in lab conditions. However, physiological reactions to emotional stimuli have been found to differ in laboratory and natural settings. Thanks to recent technological progress (e.g., in wearables) the creation of EMSR systems for a large number of consumers during their everyday activities is increasingly possible. Therefore, datasets created in the wild are needed to insure the validity and the exploitability of EMSR models for real-life applications. In this paper, we initially present common techniques used in laboratory settings to induce emotions for the purpose of physiological dataset creation. Next, advantages and challenges of data collection in the wild are discussed. To assess the applicability of existing datasets to real-life applications, we propose a set of categories to guide and compare at a glance different methodologies used by researchers to collect such data. For this purpose, we also introduce a visual tool called Graphical Assessment of Real-life Application-Focused Emotional Dataset (GARAFED). In the last part of the paper, we apply the proposed tool to compare existing physiological datasets for EMSR in the wild and to show possible improvements and future directions of research. We wish for this paper and GARAFED to be used as guidelines for researchers and developers who aim at collecting affect-related data for real-life EMSR-based applications.

/pdf/toward-emotion-recognition-from-physiological-signals-in-the-2s4vx28hjk.pdf

Toward Emotion Recognition From Physiological Signals in the Wild: Approaching the Methodological Issues in Real-Life Data Collection.

Smartphones are our ubiquitous, personal, wearable companions. Though, apart from their smartness and usefulness in our everyday lives they can cause displeasure. They allow us to be connected to a load of people and with a vast amount of apps - all of them requiring our attention. There is a growing need for a smart management to not be overwhelmed by the flood of information and notifications. A first step in that direction is to identify detectable contexts in which interruptibility is very high or low. In this paper, we present results of a survey taken by 68 persons. Within the survey, we assess how much smartphone notifications interrupt and disturb users at a specific location. The locations were selected based on the places that can be recognized by the Google Places API. This shall serve as a basis for future interruptibility research. We noticed that people are fairly interruptible while waiting, e.g. at bus stations or at parking lots. In contrast, they must not be disturbed at movie theaters, libraries or restaurants.

Preliminary investigations about interruptibility of smartphone users at specific place types

Traditionally, mood is assessed by either considering physiological data or smartphones-based self-reports. Physiological data is objective and continuous, but difficult to be collected in-field and lacks a subjective component. Smartphones provide subjective feedback and objective data, but lack physiological data. We propose to combine smartphones as a rich sensor system and smartwatches as a wearable heart rate monitor. Both serve as a platform for reporting mood states. Within an explorative user study with six subjects over four weeks, we collected smartphone data and heart rate in addition to subjective ground truth via self-reports. We assessed all three mood dimensions valence, energetic arousal and calmness, but only consider valence in the context of this paper. Analyzing the information gain, we identified the relevance of temporal features (daytime, weekday, type of day) as well as the heart rate. Decision tree classifiers trained on the first three weeks and tested on the fourth achieve recognition accuracies of up to 0.91.

A wearable system for mood assessment considering smartphone features and data from mobile ECGs

Smartphones keep their users up to date by notifying them about new messages or phone calls, among others. They offer different notification types such as ringtone, vibration and LED. However, users do not always perceive incoming notifications. Within a lab study with 36 subjects we investigated the perceptibility of these notification types depending on different smartphone positions: on the table, in the trouser pocket or in the backpack. Our results show that vibration and ringtone are perceived best at all positions. There is a statistically significant difference between these types and LED which was only perceived while the phone was on the table - if perceived at all. Users felt that vibration is most pleasant than the ringtone due to habit, lower obtrusiveness, lower disturbance, and lower distraction. Qualitative feedback reveals that the habits of the user, their (social) context, and their current task also influence the suitability of the notification type. We conclude that the best solution would be an automatic notification type selection that uses mostly vibration, though ringtone for important notifications and LED for unimportant notifications or those received while being surrounded by other people.

Investigating the perceptibility different notification types on smartphones depending on the smartphone position

Whether a person is in company is an important indicator for several research fields such as monitoring a patient’s mental health states in clinical psychology or interruptibility detection in experience sampling. Traditionally, social activity is assessed using self-report questionnaires. However, this approach is obtrusive. The best solution would be an automatic assessment. Smartphones are suitable sensing systems for this task. In this paper, we investigate relations between being in company and place types. First, we present results of an online survey taken by 68 persons. Within the survey, we assessed how likely users are to be in company at specific place types provided by the Google Places API. We identified that places such as night club, bar, movie theatre, and restaurant are primarily visited in company. Places such as post office, gym, bank, or library are visited rather alone. Some place types are undecidable and require additional context information. As a next step, we ran an in-field user study to gather enriched real-world data. We logged temporal features, user activity, place type, and self-reported company indicators as ground truth. We gathered data of 24 participants over a period of three weeks. Using information gain and \(\chi ^2\), we identified that place type and hour of day correlate with being in company with statistical significance shown by Cramer’s V. Using machine learning, we trained different classifiers to predict being in company. We achieved an accuracy of up to 91.1%. Our approach is a first step towards an automatic assessment of being in company as it is more accurate than pure guessing. We propose to enrich it with further context information such as transportation mode or a more accurate activity classifier.

Smartphone-Based Estimation of a User Being in Company or Alone Based on Place, Time, and Activity

Earable computing gains growing attention within research and becomes ubiquitous in society. However, there is an emerging need for prototyping devices as critical drivers of innovation. In our work, we reviewed the features of existing earable platforms. Based on 24 publications, we characterized the design space of earable prototyping. We used the open eSense platform (6-axis IMU, auditory I/O) to evaluate the problem-based learning usability of non-experts. We collected data from 79 undergraduate students who developed 39 projects. Our questionnaire-based results suggest that the platform creates interest in the subject matter and supports self-directed learning. The projects align with the research space, indicating ease of use, but lack contributions for more challenging topics. Additionally, many projects included games not present in current research. The average SUS score of the platform was 67.0. The majority of problems are technical issues (e.g., connecting, playing music).

/pdf/design-space-and-usability-of-earable-prototyping-4n08czzupp.pdf

Anja Exler

Papers

Preliminary investigations about interruptibility of smartphone users at specific place types

A wearable system for mood assessment considering smartphone features and data from mobile ECGs

Investigating the perceptibility different notification types on smartphones depending on the smartphone position

Smartphone-Based Estimation of a User Being in Company or Alone Based on Place, Time, and Activity

Design space and usability of earable prototyping