Retico: A Framework for Robot/IVA-ready Spoken Dialogue

Retico is a framework written in Python for developing systems that are incremental; i.e., they process streaming dialogue at the word level. Retico is also an ecosystem of modules including speech recognition, language understanding, dialogue management, language generation, object detection, logging, robot bindings, ROS and ZeroMQ bindings, sensor modules like cameras and microphones (among others) and tools for creating custom modules. The theoretical basis of incremental processing upon which Retico is built is the work by who introduced the incremental unit (IU) theoretical framework, which has inspired implementations of the framework in Python and Java. The “Robot-ready Spoken Dialogue System" framework  is built on Retico with the goal of making spoken dialogue with robots easier to develop while keeping natural spoken interaction qualities that are lost with standard turn-based systems. This tutorial is aimed at helping roboticists and virtual agent researchers to add spoken dialogue capabilities to their robots. More information about Retico can be found at github.com/retico-team.