Member of our project team, Renato Soic, participated in the conference KES-AMSTA 2020 which was held as a virtual event due to COVID-19 pandemic. KES-AMSTA is an international scientific conference for research in the field of agent and multi-agent systems. HR-SYNTH team presented its paper titled “Natural Language Agents in a Smart Environment”, which describes an agent-based model for spoken notifications in a smart environment.
Here is a short abstract of the paper:
There have been significant achievements in the area of human-computer interaction using natural language (i.e. speech), applied in various domains. As a result, users can intuitively interact with various smart devices and systems. Unfortunately, the full extent of capabilities is almost exclusively related to English language. This is expected, as most advanced solutions and achievements in the field come from tech giants originating from English-speaking countries. English is also considered the global lingua franca, so this is definitely the most viable language from the economic perspective. However, in case of minority languages, such as Croatian, there are many challenges ahead in order to achieve comparable results. In this paper, we present a model which intends to enable spoken notifications in Croatian language. The presented model defines cognitive and linguistic processes which enable generation and reproduction of understandable, human-friendly notifications in smart environments.