Learning Compact Class Codes for Fast Inference in Large Multi Class Classification
Abstract
We describe a new approach for classification with a very large number of classes where we assume some class similarity information is available, e.g. through a hierarchical organization. The proposed method learns a compact binary code using such an existing similarity information defined on classes. Binary classifiers are then trained using this code and decoding is performed using a simple nearest neighbor rule. This strategy, related to Error Correcting Output Codes methods, is shown to perform similarly or better than the standard and efficient one-vs-all approach, with much lower inference complexity.
Cite
Text
Cissé et al. "Learning Compact Class Codes for Fast Inference in Large Multi Class Classification." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2012. doi:10.1007/978-3-642-33460-3_38Markdown
[Cissé et al. "Learning Compact Class Codes for Fast Inference in Large Multi Class Classification." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2012.](https://mlanthology.org/ecmlpkdd/2012/cisse2012ecmlpkdd-learning/) doi:10.1007/978-3-642-33460-3_38BibTeX
@inproceedings{cisse2012ecmlpkdd-learning,
title = {{Learning Compact Class Codes for Fast Inference in Large Multi Class Classification}},
author = {Cissé, Moustapha and Artières, Thierry and Gallinari, Patrick},
booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
year = {2012},
pages = {506-520},
doi = {10.1007/978-3-642-33460-3_38},
url = {https://mlanthology.org/ecmlpkdd/2012/cisse2012ecmlpkdd-learning/}
}