Frequency versus probability formats in statistical word problems

Frequency Estimates for Statistical Word Similarity Measures

Statistical measures of word similarity have application in many areas of natural language processing, such as language modeling and information retrieval. We report a comparative study of two methods for estimating word cooccurrence frequencies required by word similarity measures. Our frequency estimates are generated from a terabyte-sized corpus of Web data, and we study the impact of corpus...

Joint or Conditional Probability in Statistical Word Learning: Why decide?

Three experiments investigated the ability of human learners to concurrently extract and track both joint and conditional probabilities in statistical word learning. In each experiment, participants were briefly trained on novel word– novel object pairs and asked to learn correct mappings by the end of training. Across a series of learning conditions, we systematically manipulated conditional a...

Whole-word versus part-word phonotactic probability/ neighborhood density in word learning by children

Computational Modeling of Statistical Learning: Effects of Transitional Probability Versus Frequency and Links toWord Learning

Statistical learning mechanisms play an important role in theories of language acquisition and processing. Recurrent neural network models have provided important insights into how these mechanisms might operate. We examined whether such networks capture two key findings in human statistical learning. In Simulation 1, a simple recurrent network (SRN) performed much like human learners: it was s...

On Infinite-dimensional Hierarchical Probability Models in Statistical Inverse Problems

In this article, the solution of a statistical inverse problem M = AU + E by the Bayesian approach is studied where U is a function on the unit circle T, i.e., a periodic signal. The mapping A is a smoothing linear operator and E a Gaussian noise. The connection to the solution of a finite-dimensional computational model Mkn = AkUn+Ek is discussed. Furthermore, a novel hierarchical prior model ...