We present a novel unsupervised artificial neural network for the extraction of common features in multiple data sources. This algorithm, which we name Exploratory Correlation Analysis (ECA), is a multi-stream extension of a neural implementation of Exploratory Projection Pursuit (EPP) and has a close relationship with Canonical Correlation Analysis (CCA). Whereas EPP identifies ”interesting” s...

We tightly analyze the sample complexity of CCA, provide a learning algorithm that achieves optimal statistical performance in time linear in the required number of samples (up to log factors), as well as a streaming algorithm with similar guarantees.

Canonical correlation analysis (CCA) is a classical representation learning technique for finding correlated variables in multi-view data. Several nonlinear extensions of the original linear CCA have been proposed, including kernel and deep neural network methods. These approaches seek maximally correlated projections among families of functions, which the user specifies (by choosing a kernel o...

Given a bivariate distribution, the set of canonical correlations and functions is in general finite or countable. By using an inner product between two functions via an extension of the covariance, we find all the canonical correlations and functions for the so-called Cuadras-Augé copula and prove the continuous dimensionality of this distribution.

The asymptotic behaviour of an estimator of multiinformation is investigated. It is shown that it qualitatively depends on the value of certain numerical characteristic. If this characteristic is non-zero then the estimator is asymptotically normally distributed. In the opposite case the asymptotic distribution of the estimator is the distribution of a weighted sum of squares of independent nor...

Analysis and design of multielement antenna systems in mobile fading channels require a model for the space–time cross correlation among the links of the underlying multiple-input multiple-output (MIMO) channel. In this paper, we propose a general space–time cross-correlation function for mobile frequency nonselective Rice fading MIMO channels, in which various parameters of interest such as th...

Spatial correlation modeling comprises both spatial autocorrelation and spatial cross-correlation processes. The spatial autocorrelation theory has been well-developed. It is necessary to advance the method of spatial cross-correlation analysis to supplement the autocorrelation analysis. This paper presents a set of models and analytical procedures for spatial cross-correlation analysis. By ana...

State-of-the-art pronoun interpretation systems rely predominantly on morphosyntactic contextual features. While the use of deep knowledge and inference to improve these models would appear technically infeasible, previous work has suggested that predicate-argument statistics mined from naturally-occurring data could provide a useful approximation to such knowledge. We test this idea in several...