this paper, we focus on the translation modeling problem. Before we turn to this problem, however, we should address an issue that may be a concern to some readers: Why do we estimate Pr(e) and Pr(fle) rather than estimate Pr(elf ) directly? We are really interested in this latter probability. Wouldn't we reduce our problems from three to two by this direct approach? If we can estimate Pr(fle) adequately, why can't we just turn the whole process around to estimate Pr(elf)? To understand this, imagine that we divide French and English strings into those that are well-formed and those that are ill-formed. This is not a precise notion. We have in mind that strings like Il va la bibliothque, or I live in a house, or even Colorless green ideas sleep furiously are well-formed, but that strings like lava I1 bibliothque or a I in live house are not. When we translate a French string into English, we can think of ourselves as springing from a well-formed French string into the sea of well-formed English strings with the hope of landing on a good one. It is important, therefore, that our model for Pr(elf ) concentrate its probability as much as possible on wellformed English strings. But it is not important that our model for Pr(fle ) concentrate its probability on well-formed French strings. If we were to reduce the probability of all well-formed French strings by the same factor, spreading the probability thus 265 liberated over ill-formed French strings, there would be no effect on our translations: the argument that maximizes some function f(x) also maximizes cf(x) for any positive constant c. As we shall see below, our translation models are prodigal, spraying probability all over the place, most of it on ill-formed French strings. In fact, as we discuss in Section 4.5, two...

The concept of maximum entropy can be traced back along multiple threads to Biblical times. Only recently, however, have computers become powerful enough to permit the widescale application of this concept to real world problems in statistical estimation and pattern recognition. In this paper we describe a method for statistical modeling based on maximum entropy. We present a maximum-likelihood approach for automatically constructing maximum entropy models and describe how to implement this approach efficiently, using as examples several problems in natural language processing.

We describe and evaluate experimentally a method for clustering words according to their dis- tribution in particular syntactic contexts. Words are represented by the relative frequency distributions of contexts in which they appear, and relative entropy between those distributions is used as the similarity measure for clustering. Clusters are represented by average context distributions derived from the given words according to their probabilities of cluster membership. In many cases, the clusters can be thought of as encoding coarse sense distinctions. Deterministic annealing is used to find lowest distortion sets of clusters: as the an- nealing parameter increases, existing clusters become unstable and subdivide, yielding a hierarchi- cal "soft" clustering of the data. Clusters are used as the basis for class models of word coocurrence, and the models evaluated with respect to held-out test data.

We present a technique for constructing random fields from a set of training samples. The learning paradigm builds increasingly complex fields by allowing potential functions, or features, that are supported by increasingly large subgraphs. Each feature has a weight that is trained by minimizing the Kullback-Leibler divergence between the model and the empirical distribution of the training data. A greedy algorithm determines how features are incrementally added to the field and an iterative scaling algorithm is used to estimate the optimal values of the weights. The random field models and techniques introduced in this paper differ from those common to much of the computer vision literature in that the underlying random fields are non-Markovian and have a large number of parameters that must be estimated. Relations to other learning approaches, including decision trees, are given. As a demonstration of the method, we describe its application to the problem of automatic word classifica...

SRILM is a collection of C++ libraries, executable programs, and helper scripts designed to allow both production of and experimentation with statistical language models for speech recognition and other applications. SRILM is freely available for noncommercial purposes. The toolkit supports creation and evaluation of a variety of language model types based on N-gram statistics, as well as several related tasks, such as statistical tagging and manipulation of N-best lists and word lattices. This paper summarizes the functionality of the toolkit and discusses its design and implementation, highlighting ease of rapid prototyping, reusability, and combinability of tools. 1.

This paper presents a statistical model which trains from a corpus annotated with Part-OfSpeech tags and assigns them to previously unseen text with state-of-the-art accuracy(96.6%). The model can be classified as a Maximum Entropy model and simultaneously uses many contextual "features" to predict the POS tag. Furthermore, this paper demonstrates the use of specialized features to model difficult tagging decisions, discusses the corpus consistency problems discovered during the implementation of these features, and proposes a training strategy that mitigates these problems.

We describe a parsing system based upon a language model for English that is, in turn, based upon assigning probabilities to possible parses for a sentence. This model is used in a parsing system by finding the parse for the sentence with the highest probability. This system outperforms previous schemes. As this is the third in a series of parsers by different authors that are similar enough to invite detailed comparisons but different enough to give rise to different levels of performance, we also report on some experiments designed to identify what aspects of these systems best explain their relative performance. Introduction We present a statistical parser that induces its grammar and probabilities from a hand-parsed corpus (a tree-bank). Parsers induced from corpora are of interest both as simply exercises in machine learning and also because they are often the best parsers obtainable by any method. That is, if one desires a parser that produces trees in the tree-bank ...

This article presents a measure of semantic similarityinanis-a taxonomy based on the notion of shared information content. Experimental evaluation against a benchmark set of human similarity judgments demonstrates that the measure performs better than the traditional edge-counting approach. The article presents algorithms that take advantage of taxonomic similarity in resolving syntactic and semantic ambiguity, along with experimental results demonstrating their e#ectiveness. 1. Introduction Evaluating semantic relatedness using network representations is a problem with a long history in arti#cial intelligence and psychology, dating back to the spreading activation approach of Quillian #1968# and Collins and Loftus #1975#. Semantic similarity represents a special case of semantic relatedness: for example, cars and gasoline would seem to be more closely related than, say, cars and bicycles, but the latter pair are certainly more similar. Rada et al. #Rada, Mili, Bicknell, & Blett...

Trigrams'n'Tags (TnT) is an efficient statistical part-of-speech tagger. Contrary to claims found elsewhere in the literature, we argue that a tagger based on Markov models performs at least as well as other current approaches, including the Maximum Entropy framework. A recent comparison has even shown that TnT performs significantly better for the tested corpora. We describe the basic model of TnT, the techniques used for smoothing and for handling unknown words. Furthermore, we present evaluations on two corpora.

Syntactic natural language parsers have shown themselves to be inadequate for processing highly-ambiguous large-vocabulary text, as is evidenced by their poor per- formance on domains like the Wall Street Journal, and by the movement away from parsing-based approaches to textprocessing in general. In this paper, I describe SPATTER, a statistical parser based on decision-tree learning techniques which constructs a complete parse for every sentence and achieves accuracy rates far better than any published result. This work is based on the following premises: (1) grammars are too complex and detailed to develop manually for most interesting domains; (2) parsing models must rely heavily on lexical and contextual information to analyze sentences accurately; and (3) existing n-gram modeling techniques are inadequate for parsing models. In experiments comparing SPATTER with IBM's computer manuals parser, SPATTER significantly outperforms the grammar-based parser. Evaluating SPATTER against the Penn Treebank Wall Street Journal corpus using the PARSEVAL measures, SPATTER achieves 86% precision, 86% recall, and 1.3 crossing brackets per sentence for sentences of 40 words or less, and 91% precision, 90% recall, and 0.5 crossing brackets for sentences between 10 and 20 words in length.