suggest that expanding query vectors with words that are lexically related to the original query words can ameliorate some of the problems of mismatched vocabularies. This paper examines the utility of lexical query expansion in the large, diverse TREC collection. Concepts are represented by Word

The researcher was interested to explore the possible effects of L1-L2 lexicalization mismatch on the acquisition and retention of receptive vocabulary knowledge by 90 Persian-speaking EFL learners. Lexicalization mismatch was defined in the context of this study as the lack of a lexically

Cross-Language Information Retrieval (CLIR) systems enable users to formulate queries in their native language to retrieve documents in foreign languages. Because queries and documents in CLIR do not necessarily share the same language, translation is needed before matching can take place. This translation step tends to cause a reduction in the retrieval performance of CLIR as compared to monolingual information retrieval. The prevailing CLIR approach and the focus of this study is query translation. The translation of queries is inherently difficult due to the lack of a one-to-one mapping of a lexical item and its meaning, which creates lexical ambiguity. This, and other translation problems, result in translation errors which impact CLIR performance. To understand the events occurring in cross-language retrieval query translation and the relation of these events to retrieval performance, the study explored the following research questions: 1) What kinds of translation events affect cross-language retrieval? 2) In what way does the presence of certain translation events in query translation affect retrieval performance? The study followed a two-phase multi-method approach. In phase one, a taxonomy of translation events was created through content analysis of queries and their translations in combination with an examination of the literature. In the second and final phase, a subset of the test queries was coded using the taxonomy resulting from phase one. These queries were then used in information retrieval experimentation to assess the impact of the translation events on retrieval performance.

This paper explores Tandem feature extraction used in a large-vocabulary speech recognition system. In this frame-work a multi-layer perceptron estimates phone probabilities which are treated as acoustic observations in a traditional HMM-GMM system. To determine a lower error bound, we simulated

This paper studies automatic image classification by modeling soft-assignment in the popular codebook model. The codebook model describes an image as a bag of discrete visual words selected from a vocabulary, where the frequency distributions of visual words in an image allow classification. One

to exploit the similarity of sounds across languages, and use this similarity to adapt models from one or more data rich languages for use in recognising data poor target languages. Pronunciation variation and cross language context mismatch combine to make the task more difficult then a monolingual ASR

An usual approach to address mismatching vocabulary problem is to augment the original query using dictionaries and other lexical resources and/or by looking at pseudo-relevant documents. Either way, terms are added to form a new query that will be used to score all documents in a subsequent

Abstract—Vocabulary and Word mismatches are common

ABSTRACT We describe a new algorithm for finding the hypothesis in a recognition lattice that is expected to minimize the word error rate (WER). Our approach thus overcomes the mismatch between the word-based performance metric and the standard MAP scoring paradigm that is sentence

than considering noise robustness techniques as compensating clean acoustic models for environmental noise, they may be thought of as reducing the acoustic mismatch between training and test conditions. This report examines the application of VTS model compensation or model-based Joint uncertainty