this article, please send e-mail to: tpami@computer.org, and reference IEEECS Log Number 108453

Ordered, labeled trees are trees in which each node has a label and the left-to-right order of its children (if it has any) is fixed. Such trees have many applications in vision, pattern recognition, molecular biology, programming compilation and natural language processing. Many of the applications involve comparing trees or retrieving/extracting information from a repository of trees. Examples include classification of unknown patterns, analysis of newly sequenced RNA structures, semantic taxonomy for dictionary definitions, generation of interpreters for nonprocedural programming languages, and automatic error recovery and correction for programming languages. Previous systems use exact matching (or generalized regular expression matching) for tree comparison. This paper presents a system, called Approximate-Tree-By-Example (ATBE), which allows inexact matching of trees. The ATBE system interacts with the user through a simple, but powerful query language; graphical devices a...

Ordered labeled trees are trees in which the sibling order matters. This paper presents algorithms for three problems having to do with approximate matching for such trees with variable-length don't cares (VLDC's). In strings, a VLDC symbol in the pattern may substitute for zero or more symbols in the data string. For example, if "comer" is the pattern, then the "" would substitute for the substring "put" when matching the data string "computer". Approximate VLDC matching in strings means that after the best possible substitution, the pattern still need not be the same as the data string for a match to be allowed. For example, "comer" matches "counter" within distance 1 (representing the cost of removing the "m" from "comer" and having the "" substitute for "unt"). We generalize approximate VLDC string matching to three algorithms for approximate VLDC matching on trees. The time complexity of our algorithms is O(jP j \Theta jDj \Theta min(depth(P ); leaves(P )) \Theta min(de...

. This paper describes and discusses some theoretical and practical problems arising from developing a system to combine the structured but incomplete information from machine readable dictionaries (MRDs) with the unstructured but more complete information available in corpora for the creation of a bilingual lexical data base, presenting a methodology to integrate information from both sources into a single lexical data structure. The bicord system (BIlingual CORpus-enhanced Dictionaries) involves linking entries in Collins English-French and FrenchEnglish bilingual dictionary with a large English-French and French-English bilingual corpus. We have concentrated on the class of action verbs of movement, building on earlier work on lexical correspondences specific to this verb class between languages (Klavans and Tzoukermann, 1989), (Klavans and Tzoukermann, 1990a), (Klavans and Tzoukermann, 1990b). 1 We first examine the way prototypical verbs of movement are translated in the Collin...

We present a new (continuous) quadratic programming approach for graph- and tree-isomorphism problems which is based on an equivalent maximum clique formulation. The approach is centered around a fundamental result proved by Motzkin and Straus in the mid-1960s, and recently expanded in various ways, which allows us to formulate the maximum clique problem in terms of a standard quadratic program. The attractive feature of this formulation is that a clear one-to-one correspondence exists between the solutions of the quadratic programs and those in the original, combinatorial problems. To approximately solve the program we use the so-called "replicator" equations, a class of straightforward continuous- and discrete-time dynamical systems developed in various branches of theoretical biology. We show how, despite their inherent inability to escape from local solutions, they nevertheless provide experimental results which are competitive with those obtained using more sophisticated mean-fiel...

this paper we present a novel method to represent changes and to compare trees that avoids these and other problems. The intuitive idea is to apply edit operations "in parallel" as opposed to in sequence. That is, we apply a set of edit operations, called a transformation, to a tree by first disassembling the given tree into "chunks," then operating on each chunk independently, and finally reassembling the resulting chunks to get the final tree. (In Section 2 we describe our model in detail.) Our model is free from the the unintuitive artifacts resulting from the interdependencies between edit operations in the linear edit script model. Even more importantly, searching for a minimum-cost parallel transformation is simpler that searching for a minimum-cost edit script (when moves and copies are allowed). This simplicity is because the essential information in a transformation, including its cost, can be compactly represented in a signature. Thus, we can search for a minimum-cost signature and then map it back to the corresponding transformation. In this paper we show how signatures are constructed, and how they map to transformations. The mapping between signatures and transformation is independent of the cost model used, making our methods for detecting changes useful in diverse application domains. The idea of working with signatures is widely used in the literature of differencing algorithms, in various forms (such as "traces" or matchings) [WF74, Mye86, ZS89, Yan91, CGM97]. However, the introduction of move and copy operations makes it hard to recover a script from a signature, and this makes it difficult to detect changes using signatures. To illustrate some of these difficulties, Figure 1(b) shows the "traditional" signature of the edit script in Figure 1(a). The t...

IBM is engaged in advanced research and devel-opment projects on various aspects of machine translation, between several language pairs. The activities reported on hero are all parts of a rather large-scale, international effort, fol-lowing Michael McCord’s LMT approach. The paper focuses on seven selected topics: recent enhancements made in the Slot Grammar for-malism and the specific analysis components; specification of a semantic type hierarchy and its use for verb sense disambiguation; incorpo-ration of statistical techniques in the transla-tion process; anaphora resolution; linkage of target morphology modules; methods for the construction of large MT lexicons; and interac-tive disambiguation. 1