Ballesteros, L., Croft, W.: Resolving ambiguity for cross-language retrieval. In: Research and Development in Information Retrieval. (1998) 64--71.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and documents need to be standardized into a common representation, so that monolingual retrieval techniques can be applied. From this point of view, we classify existing CLIR methods into the following three fundamental categories. The first method translates queries into the document language [1, 7, 14]. On the other hand, the second method translates documents into the query language [13, 15] The third method projects both queries and documents into a language independent space by way of thesaurus classes [9, 18] and latent semantic indexing [2, 11] We shall call those methods, query ....

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st Annual International ACM SIGIR Conference on pages 64--71, 1998.

....as can be predicted, CLIR needs to standardize queries and documents into a common representation, so that monolingual IR techniques can be applied. From this point of view, existing CLIR can be classified into three approaches. The first approach translates queries into the document language [2, 4, 5, 16], while the second approach translates documents into the query language [13, 17] The third approach projects both queries and documents into a languageindependent representation by way of thesaurus classes [6, 18] and latent semantic indexing [3, 11] Although extensive comparative experiments ....

Lisa Ballesteros and W. Bruce Croft. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 64--71, 1998.

....methods in query translation. 6] The dictionaries used are usually bilingual machine readable dictionaries (MRDs) designed for a human reader and they are converted for CLIR purposes by removing unnecessary material. Also bi or multilingual thesauruses have been developed for CLIR purposes (see [2]) Translation in CLIR is a simpler process than what is normally meant by translation: query words are often translated separately, one by one, without taking into consideration their relations to each other. When using an MRD, each word of a query is simply replaced with all of its translation ....

....Jrvelin for his constructive and valuable comments on this research. 7. References [1] Ballesteros, L. A. 2000) Cross Language Retrieval via Transitive Translation. In Croft, W. B. ed. Advances in Information Retrieval: Recent Research from the CIIR, pp. 203 234. Kluwer Academic Publishers. [2] Gilarranz, J. Gonzalo, J. Verdejo, F. 1997) An Approach to Conceptual Text Retrieval Using the EuroWordNet Multilingual Semantic Database. In Working Notes of AAAI Spring Symposium on Cross Language Text and Speech Retrieval, Stanford, CA, pp. 51 57. 3] Gollins, T. Sanderson, M. 2001) ....

[Article contains additional citation context not shown here]

Ballesteros, L. and Croft, W.B., Resolving ambiguity for cross-language retrieval. in Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, (Melbourne, Vic., Australia, 1998), ACM Press.

....translations and InQuery s weight averaging operator (#sum) to combine the weights from each synonym set into document scores [9] Pirkola s initial experiments were performed using English queries and Finnish documents; similar results are now available for a broad array of language pairs (c.f. [1, 6]) In this paper, we present an alternative to structured queries that achieves a similar effect at indexing time. The approach to structured query formulation that Pirkola introduced raises two important issues that limit the range of scenarios to which it can be applied: The computation ....

Lisa Ballesteros and W. Bruce Croft. Resolving ambiguity for cross-language retrieval. In C.J. Van Rijsbergen W. Bruce Croft, Alistair Moffat, editor, Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 64--71. ACM Press, August 1998.

....by installation at user sites and concluding with extensive usability testing. 2.1 Cross language retrieval for new languages with few resources A considerable amount of research has been conducted on the CLIR problem. The best results approach the effectiveness of a monolingual system [2]. Much of the research assumes the availability of translation resources that, for the majority of languages, are not available. Therefore other methods are required. The core of our approach is the use of bilingual dictionaries or lists of word (phrase) pairs from two languages derived from ....

....information retrieval techniques for Baltic languages; 5. 1 Cross language retrieval As has already been stated, a considerable amount of research has been conducted on the cross language retrieval problem [11] The best results of this research approach the effectiveness of a monolingual system [2], 24] Much of this research assumes the availability of expensive language resources such as machine translation systems [23] parallel corpora [6] or large comparable corpora such as news databases covering the same time period [34] Our approach, in contrast, has focused on using bilingual ....

[Article contains additional citation context not shown here]

Ballesteros, L. and Croft, W.B. Resolving Ambiguity for Cross-Language Retrieval, Proceedings of

....documents need to be standardized into a common representation, so that monolingual retrieval techniques can be applied. From this point of view, existing CLIR methods are classified into the following three fundamental categories. The first method translates queries into the document language [1, 8, 17], and the second method translates documents into the query language [16, 18] The third method projects both queries and documents into a language independent space by way of thesaurus classes [10, 21] and latent semantic indexing [3, 14] Among those above methods, the first method (i.e. query ....

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st Annual International ACM SIGIR Conference on pages 64--71, 1998.

....are translated. Since the document translation is computationally expensive [10] most efforts focus on accurate translation of the query. There are three main approaches to CLIR: machine translation, bilingual dictionary, and parallel or comparable corpora methods. Dictionary based methods [1,2,3,4] perform query translation by looking up terms on a bilingual dictionary and generating a target language query by considering some or all of the translations. In investigating SpanishEnglish CLIR, Ballesteros and Croft [2] introduced the notion of pre translation, post translation, and combined ....

....over transnational dictionary based approaches. They later investigated the effect ofphrasal translation in improving the effectiveness [3] A co occurrence method was used to resolve the ambiguity. An approach to reduce ambiguity of phrasal and term translation was eventually developed [4]. In corpus based methods [12, 14] queries are translated on the basis of multilingual terms extracted from parallel or comparable document collections. In parallel corpora, the pair or set of documents are identical but in different languages. A comparable corpus contains similar documents in ....

Ballesteros, L., and Croft, B.,"Resolving Ambiguity for Cross-Language Retrieval", SIGIR, 1998, Pages 64-71.

....narrower the window size, the less effect of unrelated terms. However, term pairs that do not co occur will increase even if the corpus is relatively large. Furthermore, mutual information has an undesired characteristic, which is the assignment of unexpectedly high values to rarely occurred terms [ 11 ]. Although our method[8, 12] also uses mutual information. In order to avoid the possibility of selecting only inappropriate translations, we take all translation candidates that exceed a certain threshold value. In addition, in order to avoid the undesired effect of rarely occurred terms, we ....

....of average precisions compared with NODIS, the proposed disambiguadon method improved 1.0 point for NTCIR and 0.8 point for NEAR, but decreased 0.7 point for AND. In general, the effectiveness of using all translation candidates in a dictionary is about a half of the one using monolingual retrieval[11], but our result of NODIS has greatly surpassed it and has achieved 92 of manual translation (MAN) It is probably because: 1) queries consist mosdy of technical terms and their ambiguities were relatively low, 2) query structuring using Boolean operator was much more effective than expected. The ....

Ballesteros, L. and Croft, W. B. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st Annual International ACM S1G1R Conference on Research and Development in Information Retrieval (S1G1R'98), 1998, pp. 64-71.

....recent findings have shown the positive influence of concept based query structuring. For instance, strong query structures improve retrieval performance when queries are expanded [15, 16] The positive effect of strong query structures seems to hold also for translation dictionary based CLIR [17, 18, 19]. However, in these studies the relevance assessments were dichotomous. In the present study, we investigate the effects of query structures and expansion on retrieving documents at different relevance levels. Earlier statistical text analyses in the context of experimental IR have focused on the ....

Ballesteros, L. and Croft, W.B. Resolving ambiguity for cross-language retrieval. In: Croft, W.B., Moffat, A., van Rijsbergen, C.J., Wilkinson, R. and Zobel, J. eds. Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. New York, NY: ACM, 1998, 64--71.

....to lists, requires prior specific permission and or a fee. SIGIR 01, September 9 12, 2001, New Orleans, Louisiana, USA. Copyright 2001 ACM 1 58113 331 6 01 0009. 5.00. extracted from parallel sentences in French and English, the performance of CLIR is improved. Ballesteros and Croft [3] performed phrase translation using information on phrase and word usage contained in the Collins machine readable dictionary. They demonstrated that translations of multi word concepts as phrases are more precise. However, a critical problem remains: if a phrase is not stored in a lexicon, how ....

....sense disambiguation. The impact of disambiguation for CLIR is debatable. Xu and Weischedel [19] estimated an upper bound on CLIR performance. They concluded that even if the translation ambiguity were solved correctly, only limited improvement can be obtained. In contrast, Ballesteros and Croft [3] showed that using co occurrence statistics from corpora can help with reducing translation ambiguity. Other studies including [1, 10, 12] also used similar approaches to select the best translation(s) In this paper, by extending the work in [1, 3] we present a disambiguation method, which can ....

[Article contains additional citation context not shown here]

Ballesteros, L., and Croft, W. B. (1998). Resolving ambiguity for cross-language retrieval. In Proceedings of

....space model retrieval system. Unweighted summing of frequencies is implemented in the Inquery system as the synonym operator . Grouping possible translations of a source language term by the Inquery synonym operator has shown to be a successful approach to cross language information retrieval [1, 10]. The model does not require the translation probabilities # i (j) to sum up to one for each i, since they are conditioned on the target language query term and not on the source language query term. Interestingly, for the final ranking it does not matter what the actual sum of the translation ....

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st ACM SIGIR Conference Research and Development in Information Retrieval (SIGIR'98), pages 64--71, 1998.

....corpus, consisting of 243,000 documents, was used. For Spanish the database ISM ALL of 208 MB was used. Tests are performed using 21 queries from the TREC crosslingual IR set with provided relevance judgements. Using these corpora allowed for comparison with results obtained by Lisa Ballesteros [Ball98]. Step 1 A: Expansion Optionally, a source language expansion step can be performed before the translation process. Implemented as a call to the INQUERY function get modified query with one of the given query words at a time, frequently occurring words, i.e. expected to be relevant to the query, ....

....a dependence tree for the Spanish terms and searching for Mutual Information only among those terms, which were directly connected in the tree for the source language Spanish. 1. Complete Destination Language Dependence Tree In a first step the EM measure is computed for all pairs of terms, cf. [Ball98] for more detail: 0 , max ) b n a n n b n a n b a n b a em = The components are as following: n is the number of documents in the corpus, n(a) the document frequency of term a, n(a,b) the co occurrence of terms a and b. The co occurrence statistics ....

[Article contains additional citation context not shown here]

Lisa Ballesteros and Bruce Croft. "Resolving ambiguity for cross-language Retrieval"

....translations of query terms can provide a form of query expansion that can improve performance. It has been shown that simpler and less resource costly techniques can work at least as effectively and that, for query translation, dictionary based techniques can outperform commercial MT systems [7]. Multilingual Thesauri: Early experiments showed that multilingual thesauri can give acceptable results for cross language retrieval and there are now a number of thesaurus based systems available commercially. A multilingual thesaurus for indexing and searching with a controlled vocabulary ....

Ballestreros, L., Croft, W.B.: Resolving Ambiguity for Cross-language Retrieval. In Proceedings of the 20 th International ACM SIGIR Conference on Research and Development in Information Retrieval, Philadelphia, PA, pp 84--91, 1997.

.... information retrieval (CLIR) techniques that are based on term by term translation depends on the coverage and accuracy of the available translation lexicon(s) Two types of translation lexicons are commonly used, one based on translation knowledge extracted from bilingual dictionaries [1] and the other based on translation knowledge extracted from bilingual corpora [8] Dictionaries provide reliable evidence, but often lack translation preference information. Corpora, by contrast, are often a better source for translations of slang or newly coined terms, but the statistical ....

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In W. B. Croft, A. Moffat, and C. V. Rijsbergen, editors, Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 64--71. ACM Press, Aug. 1998.

....thought of as estimating the weights for query language terms (as if the documents had been written in the query language) and then performing retrieval using those weights. Remarkably, the best known alternative to balanced translation was also simultaneously reported, in this case at SIGIR 98 [1, 8]. Lacking a better title for the technique, we refer to it simply as Pirkola s method, since Pirkola wrote more extensively on the issue. In Pirkola called the technique a structured query, but balanced translation also produces queries with structure. Term translation term C1 ....

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In C. V. R. W. Bruce Croft, Alistair Moffat, editor, Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 64--71. ACM Press, Aug. 1998.

....for Boolean queries was introduced earlier by Harman [4] for on line stemming. Harman did not present her algorithm as an extension of Boolean searching, but instead called it grouping . A somewhat similar approach for cross language information retrieval was adopted by Ballesteros and Croft [1] by using a synonym operator on term translations with more than one target term equivalent. The operator treats occurrences of all words within it as occurrences of a single pseudo term whose document frequency is the sum of df s for each word in the operator. If translation probabilities are ....

L. Ballesteros and W.B. Croft. Resolving ambiguity for cross-language retrieval. In W.B. Croft, A. Moat, C.J. van Rijsbergen, R. Wilkinson, and J. Zobel, editors, Proceedings of the 21st ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR'98), pages 64-71, 1998.

....words have less than 3 translation (English) words. We thought these words might be suitable as entries of the bilingual dictionary. Remaining 9 of words should be disambiguated. Using parallel corpora for disambiguation of translation equivalent is major approach among past CLIR researches [1][3] We took similar approach. We again utilized the term similarity calculation facility of MEISTER. The disambiguation process is: Calculate the similarity between a Japanese word and its translation (English word) based on the parallel corpus. We used the same corpus described in 3.1. ....

Ballesteros, L. and Croft, W.B. Resolving Ambiguity for Cross-Language Retrieval, Proc. of SIGIR98 (1998), 6471.

....if a specific term has an unique substitution this term is retained even though it not exists in the context of the target query. Note that in this process all the terms appearing in the target context are retained we do not select only the best translation as it is done in some other works [1]. 4 Experiment and Results 4.1 Multilingual experiment Two runs using English topics and retrieving documents from the pool of documents in all four languages (German, French, Italian and English) were submitted. The queries were 3 translated using the downloaded dictionaries. No ....

L. Ballesteros, W. Croft. Resolving Ambiguity for Cross-Language Retrieval. in Proceedings of the 21st ACM SIGIR'98, pages, 64-71.

....from Spanish to French via English degraded performance by 91 when compared to bilingual translation direct from Spanish to French. Ballesteros attempted to reduce the ambiguity introduced by transitive translation using query structuring and expansion techniques developed in her earlier work [2]. With their application, Ballesteros obtained an average precision for transitive translation at 67 of the monolingual performance in the target language, which compared favourably but was still below, the 79 monolingual performance obtained from a direct translation. This paper presents an ....

Ballesteros, L. and Croft, W.B., Resolving ambiguity for cross-language retrieval. in Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, (Melbourne, Vic., Australia, 1998), ACM Press.

....fashion, using a dictionary, a machine translation system, or a similar resource. Typically, any given word may have multiple possible translations, so significant e#ort has been devoted to disambiguating the resulting translations, either through the use of context [19] statistical co occurrence [3, 19], triangulated translation [8] or a number of similar techniques. In addition, researchers found that in many cases it is helpful to include words that are not direct translations of any query word, but are closely related to the meaning of the query. This observation led to the common use of ....

....documents. The use of additional disambiguation techniques such as those based on phrases and co occurrence measures (e.g. mutual information) still have to be explored within our framework. These techniques have been put to good use in approaches that do not rely on language models (e.g. [3, 7]) 3. MATHEMATICAL FRAMEWORK Let Q = e1 . ek be the query in the source language and let RQ be the set of target documents that are relevant to that query. Lavrenko and Croft [12] suggest that e#ective ranking of target documents could be achieved if we had a way of estimating the relevance ....

L. A. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In W. B. Croft, A. Mo#at, C. J. van Rijsbergen, R. Wilkinson, and J. Zobel, editors, Proceedings of the Twenty-First Annual International ACM-SIGIR Conference on pages 64--71, Melbourne, Australia, August 1998. ACM Press.

No context found.

Ballesteros, L., Croft, W.: Resolving ambiguity for cross-language retrieval. In: Research and Development in Information Retrieval. (1998) 64--71.

No context found.

Ballesteros, L. and Croft, W.B. Resolving ambiguity for cross-language retrieval. SIGIR '98, 64-71, 1998.

No context found.

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In Proceedings of the 21st Annual International ACM SIGIR Conference on pages 64--71, 1998.

No context found.

L. Ballesteros and W. B. Croft. Resolving ambiguity for cross-language retrieval. In W. B. Croft, A. Moffat, and C. V. Rijsbergen, editors, Proceedings of the 21st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 64--71. ACM Press, Aug. 1998.

No context found.

Ballestreros, L., Croft, W.B.: Resolving Ambiguity for Cross-language Retrieval. In Proceedings of the 20 th International ACM SIGIR Conference on Research and Development in Information Retrieval, Philadelphia, PA, pp 84--91, 1997.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC