no code implementations • 27 Feb 2022 • Jedrzej Kozerawski, Mayank Sharan, Rose Yu
We present two moment-based tailedness measurement concepts to improve performance on the difficult tail examples: Pareto Loss and Kurtosis Loss.
no code implementations • 14 Sep 2021 • Jedrzej Kozerawski, Matthew Turk
Real-world classification tasks are frequently required to work in an open-set setting.
1 code implementation • ACCV 2020 • Jedrzej Kozerawski, Victor Fragoso, Nikolaos Karianakis, Gaurav Mittal, Matthew Turk, Mei Chen
Unfortunately, this imbalance enables a visual recognition system to perform well on head classes but poorly on tail classes.
Ranked #53 on Long-tail Learning on ImageNet-LT
no code implementations • CVPR 2018 • Jedrzej Kozerawski, Matthew Turk
This work addresses the novel problem of one-shot one-class classification.