Generating Conversational Interfaces from Web sites

Ficha técnica

Resumen

Conversational User Interfaces (CUI), both in voice interaction and chatbot way, have become an important and popular way of interacting with current applications. These kinds of interfaces are nowadays spreading around the Web for multiple purposes, although the vast majority of Web sites do not provide them. There is still a huge gap between what Web applications allow users to do via Graphical User Interfaces (GUI), and what CUI offers. Also, there are unanswered questions about CUI design and its usability impacts on web browsing. The proposed approach transfers the responsibility of creating the conversational interaction to the users of the Web browser, instead of remaining an implementation decision taken by the Web site owners. An End-User Development (EUD) environment was designed to allow end users to define their own conversational interfaces, based on annotations of UI elements from Web Sites. In the case of the Chatbots, plugged into the target Web sites, and in the case of Voice Interfaces, creating applications for smart speakers devices based on third-party Web sites content. The underlying ideas and the reference architecture could be used to create tools that allow fast prototyping of CUI, which could be important to conduct exploratory studies about the impact of this kind of interface used in Web applications.

Publicaciones

Ripa, Gonzalo, Torre, Manuel, Urbieta, Matias, Rossi, Gustavo, Fernandez, Alejandro, Tacuri, Alex and Firmenich, Sergio, "Generating voice user interfaces from web sites", Behaviour & Information Technology, pp. 1-24, 2023. https://doi.org/10.1080/0144929X.2023.2272192 

Proyectos

Extensibilidad de navegadores Web basada en P2P: apropiación, producción y consumo de servicios y contenidos por parte de usuarios finales en el marco de la aumentación Web. Financiado por PICT - FONCYT. Dirigido por Sergio Firmenich. Duración: 1/3/2022-28/2/2025

Técnicas y herramientas de ingeniería de software para el desarrollo ágil de soluciones colaborativas basadas en crowdsourcing y colaboración, modelos e inteligencia artificial.. Financiado por UNLP. Dirigido por Matías Urbieta. CoDirigido por Leandro Antonelli. Duración: 1/1/2023-31/12/2024

Técnicas y herramientas para ingeniería de software web adaptable y ágil basada en modelos con soporte semántico y de crowdsourcing. Financiado por Universidad Nacional de la Plata (Acreditado en el Programa de Inventivos). Dirigido por Antonelli, Ruben Leandro. CoDirigido por Urbieta, Mario Matias. Duración: 1/1/2020-31/12/2022

Las extensiones web P2P y su potencial para perforar la burbuja de filtrado. Financiado por Facultad de Informática, UNLP. Dirigido por Alejandro Fernández. CoDirigido por Sergio Firmenich y Diego Torres. Duración: 1/3/2020-28/2/2021