Improvement of Image Classification: Novel Method Surpasses Softmax for Calibrated Predictions which utilizes Fuzzy Set Theory and Spatial Statistics.
Abstract: Problem of poorly calibrated predictions in neural network image classifiers will be discussed and brief overview of its solutions will be presented. A novel approach to estimation of class probabilities will be introduced which replaces traditionally used Softmax function. The approach stems from fuzzy set theory and spatial statistics.
Who: Lubomir Soukup
When: 10:00 a.m. Friday, January 12
Where: The session will occur physically at the Institute of Information Theory and Automation (UTIA). Depending on the number of listeners in room 25 or 45 (café). For directions to the institute, please refer to the following link: https://www.utia.cas.cz/contacts#way
Language: Czech (if you require English, please let us know in advance)
