Any-Class Presence Likelihood for Robust Multi-Label Classification with Abundant Negative Data

Overview

Improves multi-label classification when datasets contain many examples in which none of the target classes is present.

Research context

This work contributes to Computer Vision. The publication record above is the authoritative source for its current status; accepted papers will be updated when final bibliographic metadata becomes public.

@inproceedings{AnyClassPresenceLikelihood2026,
  title = {Any-Class Presence Likelihood for Robust Multi-Label Classification with Abundant Negative Data},
  author = {D. Tissera, O. Awadallah, M. U. Danish, A. Sadhu,, K. Grolinger},
  year = {2026},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}
}