← Search

Omer Tsimhoni

2 accepted papers

2023

Gaze Pre-Train For Improving Disparity Estimation Networks

ICASSP 2023accepted

In the process of training Neural Networks, pre-training is an unsupervised training process that uses automatically generated labels for real end-goal task inputs. It usually precedes a supervised training stage, can improve neural network performance, and can reduce training loss. In this work, we…

Cited by 0SourceScholar
2022

From Bottom-Up To Top-Down: Characterization Of Training Process In Gaze Modeling

ICASSP 2022accepted

During training, artificial neural networks might not converge to a global minimum. Usually, using gradient descent, the training procedure cause the network to stroll in the high-dimensional weights’ space. This stroll passes adjacently to local minima and locations in the geometry of loss landscap…

Cited by 0SourceScholar