Partner Provider: MAG
Beacon Application 3 serves as a human training application which acts as a tool for synthetic data generation for AI training through Virtual Environments (VEs). It enables the integration of the XR output with AI frameworks as well as the run of virtual scenarios imitating real ones. A virtual scenario may be made to exactly imitate a real scenario or add small variations presenting the opportunity to fine-tune training variables as well as verify the VE training effectiveness.
Beacon App 3 enablers:
The following tools/enables are considered to be used by the Beacon Application 3:
- Tools for building emotion representations (pre-text task in SSL)
- Tools for using pre-trained emotion representations, on one of the modalities (audio, wearable devices, body tracking)
Expected impact:
- Creation of data sets in cases where the involvement of personnel with real machinery is costly or dangerous
- Training of technical personnel in specific types of machinery
- Evaluation of personnel based on specific procedures
- Validation and delivery of ML models for emulated VR based scene analysis as a replacement of machine vision based training and ML model generation for safety applications in a training context.
Duis autem vel eum iriure dolor in hendrerit in vulputate velit esse molestie consequat, vel illum dolore eu feugiat nulla facilisis at vero eros et accumsan et iusto odio dignissim qui blandit praesent luptatum zzril delenit augue duis dolore te feugait nulla facilisi.
