Adversarial Generation of Real-time Feedback with Neural Networks for Simulation-based Training


Simulation-based training (SBT) is gaining popu- larity as a low-cost and convenient training tech- nique in a vast range of applications. However, for a SBT platform to be fully utilized as an effective training tool, it is essential that feedback on perfor- mance is provided automatically in real-time dur- ing training. It is the aim of this paper to develop an efficient and effective feedback generation method for the provision of real-time feedback in SBT. Ex- isting methods either have low effectiveness in im- proving novice skills or suffer from low efficiency, resulting in their inability to be used in real-time. In this paper, we propose a neural network based method to generate feedback using the adversarial technique. The proposed method utilizes a bounded adversarial update to minimize a L1 regularized loss via back-propagation. We empirically show that the proposed method can be used to gener- ate simple, yet effective feedback. Also, it was observed to have high effectiveness and efficiency when compared to existing methods, thus making it a promising option for real-time feedback gener- ation in SBT.

Electronic proceedings of IJCAI 2017