Motivation

Multimodal conversational AI is being integrated into more and more applications. From simple chatbots, to voicebots, to fully interactive avatars and robots. Designing and developing flexible and robust conversational interaction with a user is a challenging task. Even more so, if we need multimodal interaction i.e. taking input through multiple modalities (e.g., text, GUI, gestures, emotions, etc.) and responding to the user not just through text/voice, but also through GUI, animations, sound effects, lights, etc.

ACE provides the capabilities to select and combine components that you need for your interactive avatar application. UMIM allows you to use any Interaction Manager that is UMIM compatible and seamlessly connect it with your interactive system. UMIM provides the glue between the IM and the interactive system. This will allow you to add, swap and remove ACE Components without the need to completely rewrite your interaction logic.

Next we look at the overall architectural pattern of UMIM.