Large vocabulary continuous speech recognition (LVCSR) systems traditionally represent words in terms of smaller subword units. Both during training and during recognition, they require a mapping table, called the dictionary, which maps words into sequences of these subword units. The performance of the LVCSR system depends critically on the definition of the subword units and the accuracy of t...

We prove that the combinatorial distance between any two reduced expressions of a given permutation of {1, ..., n} in terms of transpositions lies in O(n4), a sharp bound. Using a connection with the intersection numbers of certain curves in van Kampen diagrams, we prove that this bound is sharp, and give a practical criterion for proving that the derivations provided by the reversing algorithm...

We present a novel approach to building a subword speech recognizer for the task of phonetic keyword search. The recognizer, which uses short fixed-length phonetic units, is trained with phonetic transcripts that are segmented into all possible substrings of 1, 2 and 3 phones, using a lattice representation to accommodate the overlapping units. We compare the keyword search accuracy of the prop...

The calculation of radiological paths is the most important part in statistical positron emission tomography image reconstruction algorithms. We present a new, faster algorithm which replaces Siddon's. Further code transformations on this algorithm prove to be beneecial in a Maximum Likelihood-Expectation Maximization reconstruction algorithm and the result is perfectly suitable for an implemen...

This paper proposes a novel token-passing search architecture for supporting subword unit based speech recognition and a corresponding algorithm based on the well-known LZW text compression method to determine a vocabulary of subword units in an unsupervised manner. We compare our subword unit selection algorithm to an existing approach based on Minimum Description Length (MDL) modeling and als...

Most of the constructions of infinite words having polynomial subword complexity are quite complicated, e.g., sequences of Toeplitz, sequences defined by billiards in the cube, etc. In this paper, we describe a simple method for constructing infinite words w over a binary alphabet {a, b} with polynomial subword complexity pw. Assuming w contains an infinite number of a’s, our method is based on...

We proposed a subword-based tagging for Chinese word segmentation to improve the existing character-based tagging. The subword-based tagging was implemented using the maximum entropy (MaxEnt) and the conditional random fields (CRF) methods. We found that the proposed subword-based tagging outperformed the character-based tagging in all comparative experiments. In addition, we proposed a confide...

We analyze some specific aspects concerning conditional orderings and relations among them. To this purpose we define a suitable concept of reversed conditional ordering and prove some related results. In particular we aim to compare the univariate stochastic orderings ≤st, ≤hr, and ≤lr in terms of differences among different notions of conditional orderings. Some applications of our result to ...

In the change of the orientation one has to reverse the sequence (the operation introduced in [7]) and to change the second restriction one has to rotate the sequence (the operation introduced in [26]). The goal of the paper is to prove, mostly simple, facts about the relationship between properties and attributes of the finite sequence and its rotation (similar results about reversing had been...