Xiang Ding

Nihao: A Predictive Smartphone Application Launcher

Increasingly large number of the applications installed on smartphones tends to harm the application lookup efficiency. In this paper, we introduce Nihao, a personalized intelligent app launcher system, which could help the users to find apps quickly. Nihao predicts which app the user will likely open next based on a Bayesian Network model leveraging the contextual information such as the time of day, the day of week, the user’s location and the last used app with the hypothesis that the users’ app usage pattern is context dependent. Through the field study with seven users over six weeks, we first validate the above hypothesis by comparing the prediction accuracy of Nihao with other predictors. We found that the larger UI change did not necessarily yield longer app lookup time as the app lookup time highly depended on the app icon position on screen, which suggested the prediction accuracy was the most important factor in designing such a system. At the end of the study, we conducted a user survey to evaluate Nihao qualitatively. The survey results show that five out of seven users were quite satisfied with the prediction of Nihao and thought it could help to save both app lookup and management time by ranking the app icons automatically while Nihao did not help the other two users much since they used their phones primarily for calling and texting (not for apps).

A Pilot Study of an Inspection Framework for Automated Usability Guideline Reviews of Mobile Health Applications

Usability is of significant importance for any interactive software. In the mobile domain, applications face more challenges to deliver good experiences to end users due to the characteristics and usage of mobile devices in ubiquitous computing contexts. The situation may be exacerbated for mobile health applications since the target population or domain may impose even stricter usability requirements. Heuristic Evaluation (HE) or guideline review has proven itself to be an effective approach among many usability evaluation methods. Organizing heuristic evaluation by usability professionals, however, can be costly and time consuming, particularly for frequent prototype updates generated by fast iterations. Manual inspection by human experts also suffers from scalability issues as mobile applications often need to run on a diverse set of hardware platforms. To help find potential usability problems at an early stage and reduce the workload of human usability experts, we propose an inspection framework to conduct automated guideline reviews of mobile health applications. The inspection framework is based on the Health Information Management Systems Society (HIMSS) usability guidelines for mHealth applications. First, we translate the high level descriptions of usability guidelines into operationalized metrics that can be measured by software. Second, we demonstrate the translation is meaningful by providing detailed analysis of suggested metrics and real-world case studies. We hope this framework can be used to enforce a minimum bar for the usability of mobile health applications and further adapted when new products in the field are developed.

Changing health behaviors through social and physical context awareness

The single greatest opportunity to improve health and reduce premature death lies in personal behavior. While technology-based behavior intervention has been around for many years, the emerging smartphone and wearable sensing technology brings great promise to push health behavior change further by inferring and predicting real-time behavior occurrence and its context. In this paper, we envision how social and physical context awareness could sustain behavior change motivation and assist health habit formation. We describe our preliminary work that supports this vision and outline the research challenges to be addressed.

Monitoring Sleep and Detecting Irregular Nights through Unconstrained Smartphone Sensing

Sleep is essential for a persons health and well-being. Recent advances of wearable devices and smartphone sensing have led to the proliferation of at-home sleep monitoring solutions for the consumer market. In this paper, we study how to monitor basic sleep behavior and how to detect irregular sleep nights, through unconstrained smartphone sensing, which can serve as an important indicator for both mental and physical health if the sleep problems persist. We first propose a supervised learning approach to predict bedtime and sleep duration with a light-weight context sensing schedule to minimize battery consumption. The proposed solution is validated through an extensive user study, and the prediction accuracy of bedtime and sleep duration significantly outperformed the state-of-art solution. In addition, we propose an unsupervised approach to detect irregular sleep nights by profiling and detecting contextual variations. The experiment results show that the proposed solution is effective in detecting irregular sleep nights. To the best of our knowledge, this is the first work that uses unconstrained smartphone sensing to detect sleep pattern changes with the benefits of reduced training efforts and improved robustness against behavior diversity.

Beyond Smartphone Overuse: Identifying Addictive Mobile Apps

Current research on smartphone addiction has mainly focused on addiction at the device level. This motivated us to explore more specifically on app addiction. We investigate smartphone usage for college students using surveys, logged data, and interviews. Specifically, we adapted existing smartphone addiction assessment instruments to measure app addiction. The analysis of our data shows that social and communication apps are the top 2 most addictive categories among participants. Female and male participants show no significant difference in terms of smartphone addiction. However, female participants tend to report that they are addicted to more apps. The psychological factors associated with app addiction are different between app categories. For example, compared to communication apps, participants report that it is easier to withdraw from social apps, but more difficult to control time spent on them. Correlation analysis between app usage features and app addictiveness scores reveals that compulsive open times, usage duration, and regularity of usage are good indicators of app addiction, though response time to notifications has limited predictive power.

A New Method for Automated GUI Modeling of Mobile Applications

It is often necessary to construct GUI models for automated testing of event-driven GUI applications, so test cases can be generated by traversing the GUI models systematically. It is, however, difficult to apply traditional modeling techniques directly for mobile platforms as common static models cannot reflect application behaviors under different contexts. To address these challenges, we propose a novel approach for automated GUI modeling of mobile applications and introduce our unique definition of GUI state equivalence, which can reduce state space and facilitate model merging. The proposed modeling method can already discover subtle implementation issues. Real-world case studies show that the proposed approach is effective for adaptive GUI modeling on the Android platform.

Automatic mobile photo tagging using context

The market of smartphones has been exploding, and taking pictures is a basic, maybe one of the most important functions of a smartphone. In this paper we address the problem of managing a large amount of mobile photos by automatically tagging the photos, so they can be easily browsed or searched later. Unlike other content-based photo tagging approaches, this papers main contribution is to explore an alternative opportunity of automatic photo tagging using contextual information. Both clustering and similarity-based approaches were studied for photo tagging using context such as date, time, location, environment noise, and human faces. The results show that there are intrinsic connections between contextual information and photo tags, and similarity-based approach outperforms clustering-based tagging significantly.

Unsupervised Detection of Abnormal Moments for Usability Testing of Mobile Apps

In this paper, we explore the idea of mining unexpected user interactions with mobile apps, as a way to detect usability problems through an unsupervised learning approach. We consider an interaction (or UI event) during task execution as the detection target, and derive novel features from the perspectives of transition probabilities, dwell time distribution, and context-related statistics. We propose an unsupervised learning approach to differentiate between normal and abnormal interactions through clustering on these features. Experimental results validate that the proposed method is effective in detecting different types of abnormal interactions and identifying usability issues.