A pretraining algorithm, which is a layer-bylayer greedy learning algorithm, for a deep Boltzmann machine (DBM) is presented in this paper. By considering the deep belief net type of pretraining for the DBM, which is a simplified version of the original pretraining of the DBM, two interesting theoretical facts about pretraining can be obtained. (1) By applying two different types of approximati...

This extended abstract describes the pattern discovery submission to MIREX 2014 of an algorithm that uses music segmentation (or music structure analysis) techniques and a refined greedy method in order to identify the repetitive musical patterns of a given music piece, either represented symbolically or with an actual audio file. We obtain a harmonic representation of the input and compute the...

The Walksat local search algorithm has previously been extended to handle quantification over variables. This greatly reduces model sizes, but in order to guide greedy moves the algorithm still maintains a set of violated clauses. For very large problems, or at the start of a search, this can cause memory problems. We design a new local search algorithm that does not maintain this set and is th...

The Noah's Ark Problem (NAP) is a comprehensive cost-effectiveness methodology for biodiversity conservation that was introduced by Weitzman (1998) and utilizes the phylogenetic tree containing the taxa of interest to assess biodiversity. Given a set of taxa, each of which has a particular survival probability that can be increased at some cost, the NAP seeks to allocate limited funds to conser...

We present an optimal, combinatorial 1−1/e approximation algorithm for monotone submodular optimization over a matroid constraint. Compared to the continuous greedy algorithm (Calinescu, Chekuri, Pál and Vondrák, 2008), our algorithm is extremely simple and requires no rounding. It consists of the greedy algorithm followed by local search. Both phases are run not on the actual objective functio...

In live broadcasting, the break lengths available for commercials may not always be fixed and known ex ante (e.g., strategic and injury time-outs are of variable duration in live sport transmissions). Because advertising represents a significant share of the broadcasters’ revenue, broadcasters actively manage that revenue by jointly optimizing their advertising sales and scheduling policies. We...

Comparing protein interaction networks can reveal interesting patterns of interactions for a specific function or process in distantly related species. In this paper we present GASOLINE, a Cytoscape app for multiple local alignments of PPI (protein-protein interaction) networks. The app is based on the homonymous greedy and stochastic algorithm. GASOLINE starts with the identification of sets o...

We have recently shown that classical planning problems can be characterized in terms of a width measure that is bounded and small for most planning benchmark domains when goals are restricted to single atoms. Two simple algorithms have been devised for exploiting this structure: Iterated Width (IW) for achieving atomic goals, that runs in time exponential in the problem width by performing a s...

Recent work by Rauhut and Ward developed a notion of weighted sparsity and a corresponding notion of Restricted Isometry Property for the space of weighted sparse signals. Using these notions, we pose a best weighted sparse approximation problem, i.e. we seek structured sparse solutions to underdetermined systems of linear equations. Many computationally efficient greedy algorithms have been de...

Greedy Restrictive Boltzmann Machines yield an fairly low 0.72% error rate on the famous MNIST database of handwritten digits. All that was required to achieve this result was a high number of hidden layers consisting of many neurons, and a graphics card to greatly speed up the rate of learning. Keywords—ANN (Artificial Neural Networks), RBM (Restrictive Boltzmann Machine), MNIST handwritten da...