How to Achieve Prototypical Birth and Death for OOD Detection?

Ningkang Peng, Qianfeng Yu, Xiaoqian Peng, Linjing Qian, Yafei Liu, Canran Xiao, Xinyu Lu, Tingyu Lu, Zhichao Zheng, Yanhui Gu · 2026

Read How to Achieve Prototypical Birth and Death for OOD Detection? by Ningkang Peng, Qianfeng Yu, Xiaoqian Peng, Linjing Qian, Yafei Liu, Canran Xiao, Xinyu Lu, Tingyu Lu, Zhichao Zheng, Yanhui Gu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-15650.txt.

Description

Peng and co-authors let the number of prototypes change during training for OOD detection. Fixed prototype counts fit simple and complex classes equally poorly. Their method, PID, borrows birth and death ideas from cell biology. Birth adds prototypes where current ones are overloaded. Death prunes prototypes that blur class boundaries. The shifting count yields tighter, better-separated embeddings for known classes. The authors report state-of-the-art results on CIFAR-100, with strong gains on the FPR95 metric.

Catalog record

CREATORNingkang Peng, Qianfeng Yu, Xiaoqian Peng, Linjing Qian, Yafei Liu, Canran Xiao, Xinyu Lu, Tingyu Lu, Zhichao Zheng, Yanhui Gu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36170
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.15650

Browse the full library · About the Tower of Babel