Source String Research Articles

Sparse LCS Common Substring Alignment

The “Common Substring Alignment” problem is defined as follows. The input consists of a set of strings S 1, S 2…, S c , with a common substring appearing at least once in each of them, and a target string T. The goal is to compute similarity of all strings S i with T, without computing the part of the common substring over and over again. In this paper we consider the Common Substring Alignment problem for the LCS (Longest Common Subsequence) similarity metric. Our algorithm gains its efficiency by exploiting the sparsity inherent to the LCS problem. Let Y be the common substring, n be the size of the compared sequences, L y be the length of the LCS of T and Y, denoted |LCS[ T, Y]|, and L be max{|LCS[ T, S i ]|}. Our algorithm consists of an O( nL y ) time encoding stage that is executed once per common substring, and an O( L) time alignment stage that is executed once for each appearance of the common substring in each source string. The additional running time depends only on the length of the parts of the strings that are not in any common substring.

On the Common Substring Alignment Problem

The Common Substring Alignment Problem is defined as follows: Given a set of one or more strings S1,S2…Sc and a target string T, Y is a common substring of all strings Si, that is, Si=BiYFi. The goal is to compute the similarity of all strings Si with T, without computing the part of Y again and again. Using the classical dynamic programming tables, each appearance of Y in a source string would require the computation of all the values in a dynamic programming table of size O(nℓ) where ℓ is the size of Y. Here we describe an algorithm which is composed of an encoding stage and an alignment stage. During the first stage, a data structure is constructed which encodes the comparison of Y with T. Then, during the alignment stage, for each comparison of a source Si with T, the pre-compiled data structure is used to speed up the part of Y. We show how to reduce the O(nℓ) alignment work, for each appearance of the common substring Y in a source string, to O(n)-at the cost of O(nℓ) encoding work, which is executed only once.

Natural type selection in adaptive lossy compression

Consider approximate (lossy) matching of a source string /spl sim/P, with a random codebook generated from reproduction distribution Q, at a specified distortion d. Previous work determined the minimum coding rate R/sub 1/=R(P, Q, d) for this setting. We observe that for a large word length and with high probability, the matching codeword is typical with a distribution Q/sub 1/ which is different from Q. If a new random codebook is generated /spl sim/Q/sub 1/, then the source string will favor codewords which are typical with a new distribution Q/sub 2/, resulting in a minimum coding rate R/sub 2/=R(P, Q/sub 1/, d), and so on. We show that the sequences of distributions Q/sub 1/, Q/sub 2/,... and rates R/sub 1/, R/sub 2/,..., generated by this procedure, converge to an optimum reproduction distribution Q*, and the rate-distortion function R(P, d), respectively. We also derive a fixed rate-distortion slope version of this natural type selection process. In the latter case, an iteration of the process stochastically simulates an iteration of the Blahut-Arimoto (1972) algorithm for rate-distortion function computation (without recourse to prior knowledge of the underlying source distribution). To strengthen these limit statements, we also characterize the steady-state error of these procedures when iterating at a finite string length. Implications of the main results provide fresh insights into the workings of lossy variants of the Lempel-Ziv algorithm for adaptive compression.

Head-Transducer Models for Speech Translation and TheirAutomatic Acquisition from Bilingual Data

This article presents statistical language translation models, called ``dependency transduction models'', based on collections of ``head transducers''. Head transducers are middle-out finite-state transducers which translate a head word in a source string into its corresponding head in the target language, and further translate sequences of dependents of the source head into sequences of dependents of the target head. The models are intended to capture the lexical sensitivity of direct statistical translation models, while at the same time taking account of the hierarchical phrasal structure of language. Head transducers are suitable for direct recursive lexical translation, and are simple enough to be trained fully automatically. We present a method for fully automatic training of dependency transduction models for which the only input is transcribed and translated speech utterances. The method has been applied to create English–Spanish and English–Japanese translation models for speech translation applications. The dependency transduction model gives around 75% accuracy for an English–Spanish translation task (using a simple string edit-distance measure) and 70% for an English–Japanese translation task. Enhanced with target n-grams and a case-based component, English–Spanish accuracy is over 76%; for English–Japanese it is 73% for transcribed speech, and 60% for translation from recognition word lattices.

Numerical analysis of axionic string distribution

When the amount of the axion emitted from the axionic string and the axionic domain wall is evaluated, the number density of the source string should be necessarily determined. Although it is the simplest assumption that the axionic string obeys the scaling distribution similar to the cosmic gauged string, some of the parameters have been calculated only in the local string case. The evolution of the global string is solved numerically and it is shown that the number of the axionic string would be smaller compared with the local string so that the resulting axion density should be lowered.

Scaling Distribution of Axionic Strings and Estimation of Axion Density from Axionic Domain Walls

When the amount of the axion emitted from axionic strings is evaluated, the number density of the source string should be necessarily determined. Although it is the simplest assumption that the axionic strings obey the same distribution as that of the cosmic gauged strings, this assumption has not been completely justified. The damping due to the surrounding plasma is estimated by a detailed calculation of the fermion scatter from a global string and it is shown that the friction force on the axionic strings can be neglected before the axionic domain wall formation in most cases, which is one of the important necessary conditions for the scaling distribution of the strings. The energy density of the axion produced by the collapse of the axionic domain walls is also evaluated using the axionic string distribution, and it is found to be larger or at least comparable to the axion density which originates from the axionic strings.

A translation machine with automated top-down parsing

The top-down method of parsing programming languages has been widely used in commercially available compilers. Suprisingly, techniques for automating syntax-directed, top-down parsing have not been <u>widely</u> available. This paper describes a practical top-down parsing method which is a part of a general purpose <u>translation machine</u> automaton. The syntactical description capabilities provide for the convenient and efficient specification of linguistic constructs which are difficult or lengthy to specify in BNF metalanguage. The translation machine, which provides for semantic as well as syntactic operations, has been designed for implementation in hardware, microcode, or software. The translation machine can be utilized for convenient and efficient construction of a wide variety of programs that perform source string to object string transformations.

The structure of yet another ALGOL compiler

A high-speed “top down” method of syntax analysis which completely eliminates “back-up” of the source string has been implemented in a convenient macro-language. A technique of simulation at compile time of the use of a conventional run-time stack enables the generation of code for expressions which minimizes stores, fetches and stack-pointer motion at run time, while properly treating recursion and side effects of procedures. Block structure and recursion are handled without need for interpretive methods at run time. The “context problem” in the transmission to recursive procedures of parameters “called by name” is solved in a manner which permits the handling of the common cases of simple expressions and array identifiers with particular efficiency.