Social media platforms struggle to protect users from harmful content through content moderation. These platforms have recently leveraged machine learning models to cope with the vast amount of user-generated content daily. Since moderation policies vary depending on countries and types of products, it is common to train and deploy the models per policy. However, this approach is highly inefficient, especially when the policies change, requiring dataset re-labeling and model re-training on the shifted data distribution. To alleviate this cost inefficiency, social media platforms often employ third-party content moderation services that provide prediction scores of multiple subtasks, such as predicting the existence of underage personnel, rude gestures, or weapons, instead of directly providing final moderation decisions. However, making a reliable automated moderation decision from the prediction scores of the multiple subtasks for a specific target policy has not been widely explored yet. In this study, we formulate real-world scenarios of content moderation and introduce a simple yet effective threshold optimization method that searches the optimal thresholds of the multiple subtasks to make a reliable moderation decision in a cost-effective way. Extensive experiments demonstrate that our approach shows better performance in content moderation compared to existing threshold optimization methods and heuristics.

When there is a mismatch between the target identity and the driver identity, face reenactment suffers severe degradation in the quality of the result, especially in a few-shot setting. The identity preservation problem, where the model loses the detailed information of the target leading to a defective output, is the most common failure mode. The problem has several potential sources such as the identity of the driver leaking due to the identity mismatch, or dealing with unseen large poses. To overcome such problems, we introduce components that address the mentioned problem: image attention block, target feature alignment, and landmark transformer. Through attending and warping the relevant features, the proposed architecture, called MarioNETte, produces high-quality reenactments of unseen identities in a few-shot setting. In addition, the landmark transformer dramatically alleviates the identity preservation problem by isolating the expression geometry through landmark disentanglement. Comprehensive experiments are performed to verify that the proposed framework can generate highly realistic faces, outperforming all other baselines, even under a significant mismatch of facial characteristics between the target and the driver.

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.

Keyword spotting (KWS) plays a critical role in enabling speech-based user interactions on smart devices.Recent developments in the field of deep learning have led to wide adoption of convolutional neural networks (CNNs) in KWS systems due to their exceptional accuracy and robustness.The main challenge faced by KWS systems is the trade-off between high accuracy and low latency.Unfortunately, there has been little quantitative analysis of the actual latency of KWS models on mobile devices.This is especially concerning since conventional convolution-based KWS approaches are known to require a large number of operations to attain an adequate level of performance.In this paper, we propose a temporal convolution for real-time KWS on mobile devices.Unlike most of the 2D convolution-based KWS approaches that require a deep architecture to fully capture both low-and high-frequency domains, we exploit temporal convolutions with a compact ResNet architecture.In Google Speech Command Dataset, we achieve more than 385x speedup on Google Pixel 1 and surpass the accuracy compared to the state-of-the-art model.In addition, we release the implementation of the proposed and the baseline models including an end-to-end pipeline for training models and evaluating them on mobile devices.

We suggest a use of genetic programming for transformation from a vector space to an understandable graph representation, which is part of a project to inspect the latent space in matrix factorization. Given a relation matrix, we can apply standard techniques such as non-negative matrix factorization to extract low dimensional latent space in vector representation. While the vector representation of the latent space is useful, it is not intuitive and hard to interpret. The transformation with the help of genetic programming allows us to better understand the underlying latent structure. Applying the method in the context of a stock market, we show that it is possible to recover the tree representation of technical patterns from a relation matrix. Leveraging the properties of the vector representations, we are able to find patterns that correspond to cluster centers of technical patterns. We further investigate the geometry of the latent space.

We tackle the problem of knowledge discovery in time series data using genetic programming and GPGPUs. Using genetic programming, various precursor patterns that have certain attractive qualities are evolved to predict the events of interest. Unfortunately, evolving a set of diverse patterns typically takes huge execution time, sometimes longer than one month for this case. In this paper, we address this problem by proposing a parallel GP framework using GPGPUs, particularly in the context of big financial data. By maximally exploiting the structure of the nVidia GPGPU platform on stock market time series data, we were able see more than 250-fold reduction in the running time.