For many years, a combination of principal component analysis (PCA) and independent component analysis (ICA) has been used as a blind source separation (BSS) technique to separate hidden sources of natural data. However, it is unclear why these linear methods work well because most real-world data involve nonlinear mixtures of sources. We show that a cascade of PCA and ICA can solve this nonlinear BSS problem accurately as the variety of input signals increases. Specifically, we present two theorems that guarantee asymptotically zero-error BSS when sources are mixed by a feedforward network with two processing layers. Our first theorem analytically quantifies the performance of an optimal linear encoder that reconstructs independent sources. Zero-error is asymptotically reached when the number of sources is large and the numbers of inputs and nonlinear bases are large relative to the number of sources. The next question involves finding an optimal linear encoder without observing the underlying sources. Our second theorem guarantees that PCA can reliably extract all the subspace represented by the optimal linear encoder, so that a subsequent application of ICA can separate all sources. Thereby, for almost all nonlinear generative processes with sufficient variety, the cascade of PCA and ICA performs asymptotically zero-error BSS in an unsupervised manner. We analytically and numerically validate the theorems. These results highlight the utility of linear BSS techniques for accurately recovering nonlinearly mixed sources when observations are sufficiently diverse. We also discuss a possible biological BSS implementation.

Consider a movie studio aiming to produce a set of new movies for summer release: What types of movies it should produce? Who would the movies appeal to? How many movies should it make? Similar issues are encountered by a variety of organizations, e.g., mobile-phone manufacturers and online magazines, who have to create new (non-existent) items to satisfy groups of users with different preferences. In this paper, we present a joint problem formalization of these interrelated issues, and propose generative methods that address these questions simultaneously. Specifically, we leverage the latent space obtained by training a deep generative model---the Variational Autoencoder (VAE)---via a loss function that incorporates both rating performance and item reconstruction terms. We then apply a greedy search algorithm that utilizes this learned latent space to jointly obtain K plausible new items, and user groups that would find the items appealing. An evaluation of our methods on a synthetic dataset indicates that our approach is able to generate novel items similar to highly-desirable unobserved items. As case studies on real-world data, we applied our method on the MART abstract art and Movielens Tag Genome dataset, which resulted in promising results: small and diverse sets of novel items.

Action recognition has attracted increasing attention from RGB input in computer vision partially due to potential applications on somatic simulation and statistics of sport such as virtual tennis game and tennis techniques and tactics analysis by video. Recently, deep learning based methods have achieved promising performance for action recognition. In this paper, we propose weighted Long Short-Term Memory adopted with convolutional neural network representations for three dimensional tennis shots recognition. First, the local two-dimensional convolutional neural network spatial representations are extracted from each video frame individually using a pre-trained Inception network. Then, a weighted Long Short-Term Memory decoder is introduced to take the output state at time t and the historical embedding feature at time t-1 to generate feature vector using a score weighting scheme. Finally, we use the adopted CNN and weighted LSTM to map the original visual features into a vector space to generate the spatial-temporal semantical description of visual sequences and classify the action video content. Experiments on the benchmark demonstrate that our method using only simple raw RGB video can achieve better performance than the state-of-the-art baselines for tennis shot recognition.

We investigated the impact of noisy linguistic features on the performance of a Japanese speech synthesis system based on neural network that uses WaveNet vocoder. We compared an ideal system that uses manually corrected linguistic features including phoneme and prosodic information in training and test sets against a few other systems that use corrupted linguistic features. Both subjective and objective results demonstrate that corrupted linguistic features, especially those in the test set, affected the ideal system's performance significantly in a statistical sense due to a mismatched condition between the training and test sets. Interestingly, while an utterance-level Turing test showed that listeners had a difficult time differentiating synthetic speech from natural speech, it further indicated that adding noise to the linguistic features in the training set can partially reduce the effect of the mismatch, regularize the model, and help the system perform better when linguistic features of the test set are noisy.

Many researches demonstrated that the DNA methylation, which occurs in the context of a CpG, has strong correlation with diseases, including cancer. There is a strong interest in analyzing the DNA methylation data to find how to distinguish different subtypes of the tumor. However, the conventional statistical methods are not suitable for analyzing the highly dimensional DNA methylation data with bounded support. In order to explicitly capture the properties of the data, we design a deep neural network, which composes of several stacked binary restricted Boltzmann machines, to learn the low dimensional deep features of the DNA methylation data. Experiments show these features perform best in breast cancer DNA methylation data cluster analysis, comparing with some state-of-the-art methods.

In supervised machine learning for author name disambiguation, negative training data are often dominantly larger than positive training data. This paper examines how the ratios of negative to positive training data can affect the performance of machine learning algorithms to disambiguate author names in bibliographic records. On multiple labeled datasets, three classifiers - Logistic Regression, Na\"ive Bayes, and Random Forest - are trained through representative features such as coauthor names, and title words extracted from the same training data but with various positive-negative training data ratios. Results show that increasing negative training data can improve disambiguation performance but with a few percent of performance gains and sometimes degrade it. Logistic Regression and Na\"ive Bayes learn optimal disambiguation models even with a base ratio (1:1) of positive and negative training data. Also, the performance improvement by Random Forest tends to quickly saturate roughly after 1:10 ~ 1:15. These findings imply that contrary to the common practice using all training data, name disambiguation algorithms can be trained using part of negative training data without degrading much disambiguation performance while increasing computational efficiency. This study calls for more attention from author name disambiguation scholars to methods for machine learning from imbalanced data.

A novel text-independent speaker identification (SI) method is proposed. This method uses the Mel-frequency Cepstral coefficients (MFCCs) and the dynamic information among adjacent frames as feature sets to capture speaker's characteristics. In order to utilize dynamic information, we design super-MFCCs features by cascading three neighboring MFCCs frames together. The probability density function (PDF) of these super-MFCCs features is estimated by the recently proposed histogram transform~(HT) method, which generates more training data by random transforms to realize the histogram PDF estimation and recedes the commonly occurred discontinuity problem in multivariate histograms computing. Compared to the conventional PDF estimation methods, such as Gaussian mixture models, the HT model shows promising improvement in the SI performance.

The first step to realize automatic experimental data analysis for fusion plasma experiments is fitting noisy data of temperature and density spatial profiles, which are obtained routinely. However, it has been difficult to construct algorithms that fit all the data without over- and under-fitting. In this paper, we show that this difficulty originates from the lack of knowledge of the probability distribution that the measurement data follow. We demonstrate the use of a machine learning technique to estimate the data distribution and to construct an optimal generative model. We show that the fitting algorithm based on the generative modeling outperforms classical heuristic methods in terms of the stability as well as the accuracy.

Bayesian inference plays an important role in advancing machine learning, but faces computational challenges when applied to complex models such as deep neural networks. Variational inference circumvents these challenges by formulating Bayesian inference as an optimization problem and solving it using gradient-based optimization. In this paper, we argue in favor of natural-gradient approaches which, unlike their gradient-based counterparts, can improve convergence by exploiting the information geometry of the solutions. We show how to derive fast yet simple natural-gradient updates by using a duality associated with exponential-family distributions. An attractive feature of these methods is that, by using natural-gradients, they are able to extract accurate local approximations for individual model components. We summarize recent results for Bayesian deep learning showing the superiority of natural-gradient approaches over their gradient counterparts.

This paper presents a useful method to achieve classification in satellite imagery. The approach is based on pixel level study employing various features such as correlation, homogeneity, energy and contrast. In this study gray-scale images are used for training the classification model. For supervised classification, two classification techniques are employed namely the Support Vector Machine (SVM) and the Naive Bayes. With textural features used for gray-scale images, Naive Bayes performs better with an overall accuracy of 76% compared to 68% achieved by SVM. The computational time is evaluated while performing the experiment with two different window sizes i.e., 50x50 and 70x70. The required computational time on a single image is found to be 27 seconds for a window size of 70x70 and 45 seconds for a window size of 50x50.