Concave losses for robust dictionary learning
This work addresses robustness in dictionary learning for applications dealing with noisy or outlier-prone data, representing an incremental improvement over existing methods.
The authors tackled the problem of dictionary learning being sensitive to outliers by proposing a robust framework based on concave losses, which improved outlier detection and generated better dictionaries, outperforming state-of-the-art methods like K-SVD and LC-KSVD in experiments.
Traditional dictionary learning methods are based on quadratic convex loss function and thus are sensitive to outliers. In this paper, we propose a generic framework for robust dictionary learning based on concave losses. We provide results on composition of concave functions, notably regarding super-gradient computations, that are key for developing generic dictionary learning algorithms applicable to smooth and non-smooth losses. In order to improve identification of outliers, we introduce an initialization heuristic based on undercomplete dictionary learning. Experimental results using synthetic and real data demonstrate that our method is able to better detect outliers, is capable of generating better dictionaries, outperforming state-of-the-art methods such as K-SVD and LC-KSVD.