Live Session
Hall 405
22 Sep
8:30
SGT
Friday Posters
Late Breaking Results
Power Loss Function in Neural Networks for Predicting Click-Through Rate
Ergun Biçici (Huawei R&D Center Turkey).
View Paper PDFView Poster
Abstract
Loss functions guide machine learning models towards concentrating on the error most important to improve upon. We introduce power loss functions for neural networks and apply them on imbalanced click-through rate datasets. Power loss functions decrease the loss for confident predictions and increase the loss for error-prone predictions. They improve both AUC and F1 and produce better calibrated results. We obtain improvements in the results on four different classifiers and on two different datasets. We obtain significant improvements in AUC that reach $0.44\%$ for DeepFM on the Avazu dataset.
Join the Conversation
Head to Slido and select the paper's assigned session to join the live discussion.
Conference Agenda
View Full Agenda →No items found.