Imputing missing values in high dimensional time-series is a difficult problem. This paper presents a strategy for training energy-based graphical models for imputation directly, bypassing difficulties probabilistic approaches would face. The training strategy is inspired by recent work on optimization-based learning (Domke, 2012) and allows complex neural models with convolutional and recurren...

We propose rectified factor networks (RFNs) as generative unsupervised models, which learn robust, very sparse, and non-linear codes with many code units. RFN learning is a variational expectation maximization (EM) algorithm with unknown prior which includes (i) rectified posterior means, (ii) normalized signals of hidden units, and (iii) dropout. Like factor analysis, RFNs explain the data var...

Latent generative models can learn higher-level underlying factors from complex data in an unsupervised manner. Such models can be used in a wide range of speech processing applications, including synthesis, transformation and classification. While there have been many advances in this field in recent years, the application of the resulting models to speech processing tasks is generally not exp...

The stable marriage problem has received considerable attention both due to its practical applications as well as its mathematical structure. While the original problem has all participants rank all members of the opposite sex in a strict order of preference, two natural variations are to allow for incomplete preference lists and ties in the preferences. Both variations are polynomially solvabl...

We propose comparing cosmological solutions in terms of their total spatial volumes V (τ ) as functions of proper time τ , assuming synchronous gauge, and with this intention evaluate the variations of V (τ ) about the Friedmann-Lemâıtre-Robertson-Walker (FLRW) solutions for dust. This can be done successfully in a simple manner without solving perturbation equations. In particular, we find tha...

We examine a way to extend routing using transfer patterns – a state of the art route planning approach for fast routing in public transportation networks – to a multi-modal scenario that allows using public transport, cars, and walking. We specifically examine how to precompute transfer patterns for multi-criteria, multi-modal route planning. When done naively with a Pareto cost model, this le...

We present a method for annotating gene products in the literature automatically with the terms of Gene Ontology (GO), which provides a dynamic but controlled vocabulary. Although GO is well-organized with such lexical relations as synonymy, ‘is-a’, and ‘partof’ relations among its terms, GO terms show quite a high degree of morphological and syntactic variations in the literature. As opposed t...

The stable marriage problem has received considerable attention both due to its practical applications as well as its mathematical structure. While the original problem has all participants rank all members of the opposite sex in a strict order of preference, two natural variations are to allow for incomplete preference lists and ties in the preferences. Both variations are polynomially solvabl...