Page 38 - SMILESENG

Intl. Summer School on Search- and Machine Learning-based Software Engineering
To guide the achievement of the aforementioned objective, we have defined the following research questions (RQ):
RQ1. What is the current state of research in the field of feature integration in mobile software ecosystems?
RQ2. What is the current state of research in the field of software- based dialogue systems?
RQ3. How mobile app and feature integration can be better supported by actively integrating users through dialogue-based feedback collection techniques?
III. INITIAL RESULTS
A. RQ1: Mobile app feature integration
We conducted a state-of-the-art review of gray and white literature in the field of mobile app feature integration. The results of this narrative literature review demonstrate that there are very few methodological and technical references to feature and app integration in the context of mobile software ecosystems which go beyond service and data integration (i.e., focusing on functionalities). Some proposals like mashup environments offer easy-to-use integration components, tools, and services which are generally focused on service and data integration. And while some approaches like MashReDroid [8] define feature integration strategies, users are required to be actively involved only by explicitly defining their own, specific integrations following a record and replay strategy.
B. RQ2: Software-based dialogue systems
We conducted a systematic literature review of secondary studies in the field of dialogue-based software systems [9]. The results of this review reinforce software-based dialogue systems as an emerging trend in recent scientific literature. The latest innovations in the field are focused on contextu- alized, personalized dialogue experiences not only through more advanced natural language understanding strategies (e.g., transformer models) but also through the integration of con- versational agents as embedded subcomponents into large, complex software systems. Consequently, contextualization and personalization are perceived as key features to achieve higher user adherence and user satisfaction.
C. RQ3: Feature integration through dialogue-based feedback
Using the conclusions from RQ1 and RQ2 as a proxy, we refined three scientific objectives based on the general objec- tive (as presented in Section II). We additionally identified three scientific objectives based on the research objective:
• Design and develop a mobile app repository for the man- agement, transformation and delivery of mobile applica- tions with semi-automatic feature integration capabilities.
• Integrate metadata and natural language data collection, data modelling and data storage techniques in order to build a natural language understanding data-set for a domain-specific sub-set of mobile applications.
• Design and develop a mobile-based conversational agent to facilitate personalized, contextualized mobile feature
integrations using dialogue-based feedback collection
techniques.
During the first iteration of the research process, we have
focused on analyzing and adopting the required technologies to build our solution, as well as to develop proof-of-concept (PoC) data-sets, artifacts, tools, and processes for each of the software components composing the solution. Specifically:
• Data collection service based on the automatic explo- ration of APIs, app stores and web scrapping of mobile apps search engines and catalogs for the collection of nat- ural language related data (e.g., app descriptions, reviews, official websites...).
• Repository of mobile app metadata and natural language data fields using a knowledge graph data model strategy.
• Keyword extraction process for the automated recognition of mobile app functionalities using natural language data sources, based on syntactic and semantic natural language
processing techniques.
• Conversational agent based on an adaptive knowledge
base for discussions based on a personalized mobile app catalog, including support for a proof-of-concept integration of mobile features.
• Proof-of-concept integration between two features of two open-source native Android applications.
• Draft model for structuring, documenting and maintaining personalized feature integrations for a specific user in a specific domain.
ACKNOWLEDGMENT
With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovacio´n under project / funding scheme PID2020- 117191RB-I00 / AEI/10.13039/501100011033.
REFERENCES
[1] K. Bahia and A. Delaporte, “The state of mobile internet connectivity report 2020 - mobile for development,” 2021. [Online]. Available: https://www.gsma.com/r/somic/
[2] E. M. Grua, I. Malavolta, and P. Lago, “Self-adaptation in mobile apps: a systematic literature study,” in 2019 IEEE/ACM 14th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS), 2019.
[3] Q. Motger, X. Franch, and J. Marco, “Integrating adaptive mechanisms into mobile applications exploiting user feedback,” in Research Chal- lenges in Information Science, Cham, 2021.
[4] K. Jasberg and S. Sizov, “Human uncertainty in explicit user feedback and its impact on the comparative evaluations of accurate prediction and personalisation,” Behaviour & Information Technology, 2020.
[5] Nivethan and S. Sankar, “Sentiment analysis and deep learning based chatbot for user feedback,” in Intelligent Communication Technologies and Virtual Mobile Networks. Springer International Publishing, 2019.
[6] C. Liu, B. Zhang, and G. Peng, “A systematic review of information quality of artificial intelligence based conversational agents in healthcare,” in Distributed, Ambient and Pervasive Interactions, 2021.
[7] R.J.Wieringa,DesignScienceMethodologyforInformationSystemsand Software Engineering. Springer Berlin Heidelberg, 2014.
[8] J. Zheng, L. Shen, X. Peng, H. Zeng, and W. Zhao, “MashReDroid: enabling end-user creation of Android mashups based on record and replay,” Science China Information Sciences, vol. 63, no. 10, 2020.
[9] Q. Motger, X. Franch, and J. Marco, “Software-based dialogue systems: Survey, taxonomy and challenges,” ACM Comput. Surv., 2022.
26