Virtual Help Assistant
This project leverages recent advances in Artificial Intelligence (AI) and Machine Learning (ML) technologies and aims to provide natural communication with artists, mainly by combining neural machine translation, speech recognition, speech synthesis, emotion recognition, gaze attention and a natural language toolkit.
Abstract
This paper describes an approach to create a virtual help assistant for artists working in the Visual Effects (VFX) industry. Having a virtual help assistant will have a beneficial impact on VFX production by increasing artists productivity and reducing workload on supporting parts of VFX pipeline (e.g. TDs, pipeline, IT support). This project leverages recent advances in Artificial Intelligence (AI) and Machine Learning (ML) technologies and aims to provide natural communication with artists, mainly by combining neural machine translation, speech recognition, speech synthesis, emotion recognition, gaze attention and a natural language toolkit. In order to have a natural conversation, the virtual assistant needs to react naturally to both VFX-related requests and general conversation requests, which are not related with VFX. The main challenge is to build a pipeline, connecting all the individual parts mentioned in the abstract to a lowlatency workflow. An artist should interface the whole system through an avatar, which will feel naturally and easy to communicate with.