Vocabulary-informed Extreme Value Learning
Abstract
The novel unseen classes can be formulated as the extreme values of known classes. This inspired the recent works on open-set recognition \cite{Scheirer_2013_TPAMI,Scheirer_2014_TPAMIb,EVM}, which however can have no way of naming the novel unseen classes. To solve this problem, we propose the Extreme Value Learning (EVL) formulation to learn the mapping from visual feature to semantic space. To model the margin and coverage distributions of each class, the Vocabulary-informed Learning (ViL) is adopted by using vast open vocabulary in the semantic space. Essentially, by incorporating the EVL and ViL, we for the first time propose a novel semantic embedding paradigm -- Vocabulary-informed Extreme Value Learning (ViEVL), which embeds the visual features into semantic space in a probabilistic way. The learned embedding can be directly used to solve supervised learning, zero-shot and open set recognition simultaneously. Experiments on two benchmark datasets demonstrate the effectiveness of proposed frameworks.
Cite
@article{arxiv.1705.09887,
title = {Vocabulary-informed Extreme Value Learning},
author = {Yanwei Fu and HanZe Dong and Yu-feng Ma and Zhengjun Zhang and Xiangyang Xue},
journal= {arXiv preprint arXiv:1705.09887},
year = {2018}
}
Comments
we significantly change the content of this paper which makes it another paper. In order not to misleading, we decided to withdraw it