We show how a new unsupervised approach to learning musical relationships can exploit Bayesian MAP induction of stochastic transduction grammars to overcome the challenges of learning complex relationships between multiple rhythmic parts that previously lay outside the scope of general computational approaches to music structure learning. A good illustrative genre is flamenco, which employs not...

Because many databases contain or can be embellished with structural information, a method for identifying interesting and repetitive substructures is an essential component to discovering knowledge in such databases. This paper describes the SUBDUE system, which uses the minimum description length (MDL) principle to discover substructures that compress the database and represent structural con...

I discuss the use of Kolmogorov complexity and Bayes’ theorem in Solomonoff’s inductive method to explicate a generM concept of simplicity. This makes it possible to understand how the search for simple, i.e., short, computational descriptions of (empirical) data yields to the discovery of patterns, and hence more probable predictions. I show how the simplicity bias of Langley’s BACON.2 and Tha...

Automatic segmentation is performed using watersheds of the gradient magnitude and compression techniques. Linear Scale-Space is used to discover the neighbourhood structure and catchment basins are locally merged with Minimum Description Length. The algorithm can form a basis for a large range of automatic segmentation algorithms based on watersheds, scale-spaces, and compression.

Minimum Description Length (MDL) shape modelling has recently established itself as a gold standard in optimal shape modelling. By end of 2004, available MDL packages are written in Matlab and/or incorrectly. The purpose of this technical report is to list the most critical problems that occurred during a C++ implementation of MDL–based 2D shape modelling.

Stochastic categorial grammars (SCGs) are introduced as a more appropriate formalism for statistical language learners to est imate than stochastic context free grammars. As a vehicle for demonstrating SCG estimation, we show, in terms of crossing rates and in coverage, that when training material is limited, SCG estimation using the Minimum Description Length Principle is preferable to SCG est...

We present a novel method for estimating the number of signals impinging on a uniform linear array using observed sensor data. Unlike other algorithms which apply Rissanen's Minimum Description Length (MDL) principle to the observed data for source enumeration, this method applies it to the prediction errors of a linear model which has been tted to an appropriate data matrix. It is a one-dimens...

We propose a measure for assessing the degree of influence of a set of edges of a Bayesian network on the overall fitness of the network, starting with probability distributions extracted from a data set. Standard fitness measures such as the Cooper-Herskowitz score or the score based on the minimum description length are computationally expensive and do not focus on local modifications of netw...

We propose an heuristic algorithm that induces decision graphs from training sets using Rissanen's minimum description length principle to control the tradeoo between accuracy in the training set and complexity of the hypothesis description.

We extend the Chow-Liu algorithm for general random variables while the previous versions only considered finite cases. In particular, this paper applies the generalization to Suzuki’s learning algorithm that generates from data forests rather than trees based on the minimum description length by balancing the fitness of the data to the forest and the simplicity of the forest. As a result, we s...