DialogOS is a (fairly classical) state-based dialog management environment with built-in scripting and – as a major feature for an open-source system – fully-working graphical user interface that enables even school kids to get their dialogs up and running in minutes. Yet, it scales to large and complex systems that interact with databases, LEGO- or ROS-based robots, and the built-in plugin-mechanism allows developers to quickly integrate it into their system workflows.
I am one of the main developers of DialogOS and we use it in research as the dialog engine for the Smooth robot, previously on the server-end of our Android-based SARA agent, as well in teaching university students, in our frequent outreach programs such as the university's Girls' Day and the "Schnupperstudium".
For details, see our Interspeech 2018 poster and/or the DialogOS webpage. DialogOS is developed on Github.
Related Publications:
- Alexander Koller, Timo Baumann and Arne Köhn (2018).
"DialogOS: Simple and extensible dialog modeling".
in: Proceedings of Interspeech,. Hyderabad, India.
URN, pdf, poster, webpage, code, bibtex - Vivian Tsai, Timo Baumann, Florian Pecune and Justine Cassell (2018).
"Faster Responses are Better Responses: Introducing Incrementality into Sociable Virtual Personal Assistants".
in: International Workshop on Spoken Dialog Systems 2018. Singapore.
pdf, bibtex - Timo Baumann, Maike Paetzel, Philipp Schlesinger and Wolfgang Menzel (2013).
"Using Affordances to Shape the Interaction in a Hybrid Spoken Dialogue System".
in: Proceedings of Elektronische Sprachsignalverarbeitung (ESSV 2013). Bielefeld, Germany. TUDpress, pages 12-19.
URN, pdf, slides, video, bibtex - Maike Paetzel, Philipp Schlesinger, Mircea Pricop, Radu Comaneci, Timo Baumann and Wolfgang Menzel (2013).
"Inkrementelle und zustandsbasierte Verarbeitung in einem hybriden Sprachdialogsystem".
in: Postersession der Sektion Computerlinguistik. Potsdam, Germany.
abstract, poster, bibtex
Funding
This project was previously founded in the Yahoo!/Oath/CMU InMind project and currently is funded in the SMOOTH project for seamless human-robot interaction.
