Addressing imbalanced or long-tailed data is a major challenge in visual recognition tasks due to disparities between training and testing distributions and issues with data noise. We propose the Wrapped Cauchy Distributed Angular Softmax (WCDAS), a novel softmax function that incorporates data-wise Gaussianbased kernels into the angular correlation between feature representations and classifier weights, effectively mitigating noise and sparse sampling concerns. The class-wise distribution of angular representation becomes a sum of these kernels. Our theoretical analysis reveals that the wrapped Cauchy distribution excels the Gaussian distribution in approximating mixed distributions.Additionally, WCDAS uses trainable concentration parameters to dynamically adjust the compactness and margin of each class. Empirical results confirm label-aware behavior in these parameters and demonstrate WCDAS's superiority over other state-of-the-art softmaxbased methods in handling long-tailed visual recognition across multiple benchmark datasets. The code is public available.

No abstract

No abstract

No abstract

On account of growing demands for personalization, the need for a so-called few-shot TTS system that clones speakers with only a few data is emerging. To address this issue, we propose Attentron, a few-shot TTS model that clones voices of speakers unseen during training. It introduces two special encoders, each serving different purposes. A fine-grained encoder extracts variable-length style information via an attention mechanism, and a coarse-grained encoder greatly stabilizes the speech synthesis, circumventing unintelligible gibberish even for synthesizing speech of unseen speakers. In addition, the model can scale out to an arbitrary number of reference audios to improve the quality of the synthesized speech. According to our experiments, including a human evaluation, the proposed model significantly outperforms state-of-the-art models when generating speech for unseen speakers in terms of speaker similarity and quality.

No abstract

Abstract. In this paper, we propose a fast and accurate approximation to the information potential of Information Theoretic Learning (ITL) using the Fast Gauss Transform (FGT). We exemplify here the case of the Minimum Error Entropy criterion to train adaptive systems. The FGT reduces the complexity of the estimation from O(N 2 ) to O(pkN) where p is the order of the Hermite approximation and k the number of clusters utilized in FGT. Further, we show that FGT converges to the actual entropy value rapidly with increasing order p unlike the Stochastic Information Gradient, the present O(pN) approximation to reduce the computational complexity in ITL. We test the performance of these FGT methods on System Identification with encouraging results.

No abstract