Automated training feedback in an Artificial Intelligence Simulation
Invention 2019-085
Automated training feedback in an Artificial Intelligence Simulation
听
A virtual training program that provides automated feedback on a user鈥檚 expertise has been developed at UU直播.
听
Market Need
Simulation-based training has given people the opportunity to provide assessment and training of technical skills. Especially with computer-based simulators there is the potential to generate a vast amount of data on the user鈥檚 interaction with the simulated task. These tasks are then evaluated by a supervisor based on predefined metrics meant to determine their performance. While this lowers the training time for the supervisor, there is still no reliable option to provide automated and objective feedback on the simulated task.
听
Technology Summary
With a machine learning algorithm, this invention can first be trained with selected metrics of performance to distinguish users based on expertise. A Virtual Operative Assistant can then be built to classify a new user as expert or non-expert based on this automated and objective feedback. Originally designed as a virtual reality surgical simulation, this program will provide a unique breakdown of performance with regard to individual metrics. Since this framework is highly flexible, there can be a variety of modifications which will allow it to be used for many different types of simulations.
听
Advantages
- There are currently no reliable methods of providing automated feedback on a simulated task
- The machine learning algorithm provides a strong framework based on metrics inputted by experts
- The Virtual Operative Assistant can break down performance for individual metrics
- Very flexible system, so this can be used for a variety of technical fields
听
Patent Status
Filed US
听