3 papers
cs.LG2026
Generalised Balanced Softmax: A Finite-Data Perspective on Logit Adjustment for Long-Tailed Recognition
Yi-Hang Zhu, Rajeev Raman, Shiqi Su +4
Models trained on long-tailed data using standard softmax tend to exhibit higher training error and a larger generalisation gap for classes with fewer training samples. We characte…
cs.CV2026
Multi-Domain Learning with Global Expert Mapping
Pourya Shamsolmoali, Masoumeh Zareapoor, Huiyu Zhou +3
Human perception generalizes well across different domains, but most vision models struggle beyond their training data. This gap motivates multi-dataset learning, where a single mo…
cs.CV2025
Action-Dynamics Modeling and Cross-Temporal Interaction for Online Action Understanding
Xinyu Yang, Zheheng Jiang, Feixiang Zhou +8
Online action understanding, encompassing action detection and anticipation, is central to numerous practical applications. However, untrimmed videos contain substantial redundant…