Follow
Tina Behnia
Tina Behnia
Verified email at ece.ubc.ca
Title
Cited by
Cited by
Year
Imbalance trouble: Revisiting neural-collapse geometry
C Thrampoulidis, GR Kini, V Vakilian, T Behnia
Advances in Neural Information Processing Systems 35, 27225-27238, 2022
452022
On the Implicit Geometry of Cross-Entropy Parameterizations for Label-Imbalanced Data
T Behnia, GR Kini, V Vakilian, C Thrampoulidis
International Conference on Artificial Intelligence and Statistics, 10815-10838, 2023
142023
Supervised-Contrastive Loss Learns Orthogonal Frames and Batching Matters
GR Kini, V Vakilian, T Behnia, J Gill, C Thrampoulidis
arXiv preprint arXiv:2306.07960, 2023
52023
On how to avoid exacerbating spurious correlations when models are overparameterized
T Behnia, K Wang, C Thrampoulidis
2022 IEEE International Symposium on Information Theory (ISIT), 121-126, 2022
32022
Supervised Contrastive Representation Learning: Landscape Analysis with Unconstrained Features
T Behnia, C Thrampoulidis
arXiv preprint arXiv:2402.18884, 2024
2024
Symmetric Neural-Collapse Representations with Supervised Contrastive Loss: The Impact of ReLU and Batching
GR Kini, V Vakilian, T Behnia, J Gill, C Thrampoulidis
The Twelfth International Conference on Learning Representations, 2023
2023
Learning from imbalanced data: a geometric study on over-parameterized models
T Behnia
University of British Columbia, 2023
2023
The system can't perform the operation now. Try again later.
Articles 1–7