Contrastive ECOC: Learning Output Codes for Adversarial Defense
This work addresses the need for more effective and adaptive encoding mechanisms in multiclass classification, particularly for adversarial defense, though it appears incremental as it builds on existing ECOC methods.
The paper tackles the problem of multiclass classification by proposing automated codebook learning models based on contrastive learning to replace manual or random designs in Error Correcting Output Codes (ECOC), resulting in superior robustness to adversarial attacks across four datasets compared to two baselines.
Although one-hot encoding is commonly used for multiclass classification, it is not always the most effective encoding mechanism. Error Correcting Output Codes (ECOC) address multiclass classification by mapping each class to a unique codeword used as a label. Traditional ECOC methods rely on manually designed or randomly generated codebooks, which are labor-intensive and may yield suboptimal, dataset-agnostic results. This paper introduces three models for automated codebook learning based on contrastive learning, allowing codebooks to be learned directly and adaptively from data. Across four datasets, our proposed models demonstrate superior robustness to adversarial attacks compared to two baselines. The source is available at https://github.com/YuChou20/Automated-Codebook-Learning-with-Error-Correcting-Output-Code-Technique.