Page 37 - SMILESENG

Intl. Summer School on Search- and Machine Learning-based Software Engineering
Transforming Mobile Software Ecosystems with Semi-Automatic Feature Integration through Dialogue-Based Feedback
Quim Motger
PhD Student Department of Service and Information System Engineering Universitat Polite`cnica de Catalunya Barcelona, Catalonia, Spain [email protected]
Xavier Franch Co-supervisor Department of Service and Information System Engineering Universitat Polite`cnica de Catalunya Barcelona, Catalonia, Spain [email protected]
Jordi Marco Co-supervisor Department of Computer Science Universitat Polite`cnica de Catalunya Barcelona, Catalonia, Spain [email protected]
Abstract—Mobile applications have become a daily-use com- modity worldwide. Users need to manage a wide variety of com- plex use cases using multiple, isolated applications in combination to achieve higher, more complex goals. As each user’s goals are unique, defining these integrations becomes a challenging task. To address these challenges, users can play a proactive role by providing valuable feedback for runtime integration. Hence, the use of conversational agents to assist users by collecting natural language feedback can be considered as a key research trend towards this end. Following a design science methodology, we aim at exploring how the integration of mobile application features can be better supported by actively integrating users through dialogue-based feedback collection techniques. To this end, we define the following scientific objectives: (1) transforming mobile applications with automatic feature integration mechanisms; (2) generating natural language data-sets for training conversational agents in the context of mobile software ecosystems; and (3) designing explicit feedback collection techniques for feature inte- gration. We envisage that our research will contribute to explore the potential of users’ natural language feedback for personalized software experiences in the context of mobile software ecosystems.
I. INTRODUCTION AND MOTIVATION
With the adoption of smartphone devices as ubiquitous tools for both personal and professional use cases [1], mobile soft- ware ecosystems [2] have become complex, heterogeneous and constantly-evolving environments, with particular challenges from the user experience perspective. As users’ goals and needs are unique, managing and fully exploiting the potential of the applications’ portfolio is a challenging task, which includes the ability of such systems to adapt hardware and software components to match the users’ needs [3]. Despite the intrinsic interconnectivity between mobile apps and their features, there is a lack of focused research in the field of cross-app feature integration, and these strategies vaguely offer any customization to extend mobile software ecosystems in order to deliver personalized feature integration paradigms. To achieve a highly personalized, customizable user experience in terms of feature integration, actively involving the user
is fundamental to learn about their unique goals and needs. To this end, effective analysis of explicit user feedback in adaptive software systems is a fundamental strategy in adaptive software systems [4], to which the use of natural language interfaces or conversational agents is emerging as a leading research trend in a wide variety of domains [5], [6].
Based on this context, in this paper, we summarize the objectives, research questions and initial results of the the- sis titled “Transforming Mobile Software Ecosystems with Semi-Automatic Feature Integration through Dialogue-Based Feedback”. This research is intended to explore how the integration of a mobile-based conversational agent can be used to provide personalized user experiences in terms of cross- app feature integration among the applications’ portfolio of mobile users. We expect that our research will contribute to lay the groundwork for future research both in adaptive mobile software ecosystems and the adoption of personalized conversational agents.
II. RESEARCH METHOD
The research method is based on an adaptation of the Design Science methodology for Information Systems and Software Engineering as defined by Wieringa [7]. We designed a research plan based on three research process iterations, in- cluding the following activities: (i) definition of objectives; (ii) design and development; (iii) demonstration; (iv) evaluation and verification; and (v) dissemination.
Following the Goal Question Metric (GQM) template, we state the general objective of this project as follows:
25
Analyze feature integration supported by dialogue-based feedback for the purpose of enhancing and personalizing the user experience with respect to cross-app feature integration
from the point of view of users
in the context of mobile software ecosystems.