Get Started
Research questionHow should multi-expert models balance diversity and prediction variance for long-tailed visual recognition?In long-tailed visual recognition, rare classes provide limited data, making expert predictions more prone to overfitting and disagreement. Greater diversity among experts does not necessarily translate into better ensemble accuracy.
Computer Vision
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.VICAL: Vicinal Consistency Alignment for Long-Tailed Visual RecognitionThe source concerns multi-expert models for long-tailed image classification and reports evidence on CIFAR-LT, ImageNet-LT, and iNaturalist 2018. Its findings address logit adjustment, explicit diversity regularization, prediction variance, and consistency across experts.research paper · Sep 4, 2026
Related questions
How can deep imbalanced regression avoid underfitting scarce tail labels when uncertainty varies across instances?How can text-conditioned visual autoregressive models increase image diversity without sacrificing quality?How can one vision model remain robust across changing and unseen adversarial perturbation budgets?How should self-supervised visual learning combine objectives to prevent collapse while preserving semantic and spatial information?
Home
Topics
Search
Library