Cross-language information retrieval models based on latent topic models trained with document-aligned comparable corpora

In this paper, we study different applications of cross-language latent topic models trained on comparable corpora. The first focus lies on the task of cross-language information retrieval (CLIR). The Bilingual Latent Dirichlet allocation model (BiLDA) allows us to create an interlingual, language-independent representation of both queries and documents. We construct several BiLDA-based document models for CLIR, where no additional translation resources are used. The second focus lies on the methods for extracting translation candidates and semantically related words using only per-topic word distributions of the cross-language latent topic model. As the main contribution, we combine the two former steps, blending the evidences from the per-document topic distributions and the per-topic word distributions of the topic model with the knowledge from the extracted lexicon. We design and evaluate the novel evidence-rich statistical model for CLIR, and prove that such a model, which combines various (only internal) evidences, obtains the best scores for experiments performed on the standard test collections of the CLEF 2001–2003 campaigns. We confirm these findings in an alternative evaluation, where we automatically generate queries and perform the known-item search on a test subset of Wikipedia articles. The main importance of this work lies in the fact that we train translation resources from comparable document-aligned corpora and provide novel CLIR statistical models that exhaustively exploit as many cross-lingual clues as possible in the quest for better CLIR results, without use of any additional external resources such as parallel corpora or machine-readable dictionaries.     

Web、语料库与双语平行语料库的建设

对Web和语料库以及多语语料库的关系进行辨析,针对Web上丰富的各类电子文本,从语言工程角度出发,提出"分步骤、按领域"建设大规模双语平行语料库的思路,即选定领域专一、语言可靠、格式规范的文本,逐次建设特定领域的语料库,最后汇总成高质量、大规模、全领域的"高大全"式双语平行语料库。同时,围绕一个实例介绍如何利用Web资源建设特定领域双语平行语料库。     

Leveraging comparable corpora for cross-lingual information retrieval in resource-lean language pairs

Cross-language information retrieval (CLIR) has so far been studied with the assumption that some rich linguistic resources such as bilingual dictionaries or parallel corpora are available. But creation of such high quality resources is labor-intensive and they are not always at hand. In this paper we investigate the feasibility of using only comparable corpora for CLIR, without relying on other linguistic resources. Comparable corpora are text documents in different languages that cover similar topics and are often naturally attainable (e.g., news articles published in different languages at the same time period). We adapt an existing cross-lingual word association mining method and incorporate it into a language modeling approach to cross-language retrieval. We investigate different strategies for estimating the target query language models. Our evaluation results on the TREC Arabic–English cross-lingual data show that the proposed method is effective for the CLIR task, demonstrating that it is feasible to perform cross-lingual information retrieval with just comparable corpora.     

Dictionary-Based Cross-Language Information Retrieval: Problems, Methods, and Research Findings

This paper reviews literature on dictionary-based cross-language information retrieval (CLIR) and presents CLIR research done at the University of Tampere (UTA). The main problems associated with dictionary-based CLIR, as well as appropriate methods to deal with the problems are discussed. We will present the structured query model by Pirkola and report findings for four different language pairs concerning the effectiveness of query structuring. The architecture of our automatic query translation and construction system is presented.     

实用型跨语言检索查询翻译接口性能改进研究

针对某网站在线运行的基于词典的查询翻译接口,对接口中实现的3种比较常用的查询翻译消歧算法进行改进,利用性能测试工具Loadrunner对改进前后的查询翻译接口的运行效率进行对比分析,最后总结影响基于词典的查询翻译的运行效率的3个主要因素。     

Using Statistical Term Similarity for Sense Disambiguation in Cross-Language Information Retrieval

With the increasing availability of machine-readable bilingual dictionaries, dictionary-based automatic query translation has become a viable approach to Cross-Language Information Retrieval (CLIR). In this approach, resolving term ambiguity is a crucial step. We propose a sense disambiguation technique based on a term-similarity measure for selecting the right translation sense of a query term. In addition, we apply a query expansion technique which is also based on the term similarity measure to improve the effectiveness of the translation queries. The results of our Indonesian to English and English to Indonesian CLIR experiments demonstrate the effectiveness of the sense disambiguation technique. As for the query expansion technique, it is shown to be effective as long as the term ambiguity in the queries has been resolved. In the effort to solve the term ambiguity problem, we discovered that differences in the pattern of word-formation between the two languages render query translations from one language to the other difficult.     

A user-centric approach to identifying best deployment strategies for language tools: the impact of content and access language on Web user behaviour and attitudes

The number of Web users whose first language is not English continues to grow, as does the amount of content provided in languages other than English. This poses new challenges for actors on the Web, such as in which language(s) content should be offered, how search tools should deal with mono- and multilingual content, and how users can make the best use of navigation and search options, suited to their individual linguistic skills. How should these challenges be dealt with? Technological approaches to non-English (or in general, cross-language) Web search have made large progress; however, translation remains a hard problem. This precludes a low-cost but high-quality blanket all-language coverage of the whole Web. In this paper, we propose a user-centric approach to answering questions of where to best concentrate efforts and investments. Drawing on linguistic research, we describe data on the availability of content and access to it in first and second languages across the Web. We then present three studies that investigated the impact of the availability (or not) of first-language content and access forms on user behaviour and attitudes. The results indicate that non-English languages are under-represented on the Web and that this is partly due to content-creation, link-setting and link-following behaviour. They also show that user satisfaction is influenced both by the cognitive effort of searching and the availability of alternative information in that language. These findings suggest that more cross-language tools are desirable. However, they also indicate that context (such as user groups’ domain expertise or site type) should be considered when tradeoffs between information quality and multilinguality need to be taken into account.     

Cross-lingual training of summarization systems using annotated corpora in a foreign language

The increasing trend of cross-border globalization and acculturation requires text summarization techniques to work equally well for multiple languages. However, only some of the automated summarization methods can be defined as “language-independent,” i.e., not based on any language-specific knowledge. Such methods can be used for multilingual summarization, defined in Mani (Automatic summarization. Natural language processing. John Benjamins Publishing Company, Amsterdam, 2001) as “processing several languages, with a summary in the same language as input”, but, their performance is usually unsatisfactory due to the exclusion of language-specific knowledge. Moreover, supervised machine learning approaches need training corpora in multiple languages that are usually unavailable for rare languages, and their creation is a very expensive and labor-intensive process. In this article, we describe cross-lingual methods for training an extractive single-document text summarizer called MUSE (MUltilingual Sentence Extractor)—a supervised approach, based on the linear optimization of a rich set of sentence ranking measures using a Genetic Algorithm. We evaluated MUSE’s performance on documents in three different languages: English, Hebrew, and Arabic using several training scenarios. The summarization quality was measured using ROUGE-1 and ROUGE-2 Recall metrics. The results of the extensive comparative analysis showed that the performance of MUSE was better than that of the best known multilingual approach (TextRank) in all three languages. Moreover, our experimental results suggest that using the same sentence ranking model across languages results in a reasonable summarization quality, while saving considerable annotation efforts for the end-user. On the other hand, using parallel corpora generated by machine translation tools may improve the performance of a MUSE model trained on a foreign language. Comparative evaluation of an alternative optimization technique—Multiple Linear Regression—justifies the use of a Genetic Algorithm.     

Studying machine translation technologies for large-data CLIR tasks: a patent prior-art search case study

Prior-art search in patent retrieval is concerned with finding all existing patents relevant to a patent application. Since patents often appear in different languages, cross-language information retrieval (CLIR) is an essential component of effective patent search. In recent years machine translation (MT) has become the dominant approach to translation in CLIR. Standard MT systems focus on generating proper translations that are morphologically and syntactically correct. Development of effective MT systems of this type requires large training resources and high computational power for training and translation. This is an important issue for patent CLIR where queries are typically very long sometimes taking the form of a full patent application, meaning that query translation using MT systems can be very slow. However, in contrast to MT, the focus for information retrieval (IR) is on the conceptual meaning of the search words regardless of their surface form, or the linguistic structure of the output. Thus much of the complexity of MT is not required for effective CLIR. We present an adapted MT technique specifically designed for CLIR. In this method IR text pre-processing in the form of stop word removal and stemming are applied to the MT training corpus prior to the training phase. Applying this step leads to a significant decrease in the MT computational and training resources requirements. Experimental application of the new approach to the cross language patent retrieval task from CLEF-IP 2010 shows that the new technique to be up to 23 times faster than standard MT for query translations, while maintaining IR effectiveness statistically indistinguishable from standard MT when large training resources are used. Furthermore the new method is significantly better than standard MT when only limited translation training resources are available, which can be a significant issue for translation in specialized domains. The new MT technique also enables patent document translation in a practical amount of time with a resulting significant improvement in the retrieval effectiveness.     

基于词汇链的关键词自动标引方法

关键词自动标引是一种识别有意义且具有代表性片段或词汇的自动化技术。关键词自动标引可以为自动摘要、自动分类、自动聚类、机器翻译等应用提供辅助作用。本文利用基于知网的词语语义相关度算法对词汇链的构建算法进行了改进,并结合词频和词的位置等统计信息,进行关键词的自动标引。实验证明,该方法可以有效的进行关键词的自动标引。     

基于单词规则的英汉机器翻译和消歧研究

本文比较各种机器翻译方法的特点，借鉴短语结构、ＧＰＳＧ、ＨＰＳＧ和语料库等计算语言学理论，采用以单词为核心的方法，建立一组以１０００单词为背景的英汉翻译规则，并实现以单词规则、通用规则为准，进行英汉翻译的机译系统。本文还对机译系统构成、设计、词典结构等技术问题进行了探讨，并提出一套规则和结点的评价方法，为消除机器翻译中的句法歧义做了有益的尝试。     

神经网络语言模型在统计机器翻译中的应用

近两年来,神经机器翻译（Neural Machine Translation, NMT）模型主导了机器翻译的研究,但是统计机器翻译（Statistical Machine Translation, SMT）在很多应用场合（尤其是专业领域）仍有较强的竞争力。如何利用深度学习技术提升现有统计机器翻译的水平成为研究者们关注的主要问题。由于语言模型是统计机器翻译中最核心的模块之一,本文主要从语言模型的角度入手,探索神经网络语言模型在统计机器翻译中的应用。本文分别探讨了基于词和基于短语的神经网络语言模型,在汉语到英语和汉语到日语的翻译实验表明神经网络语言模型能够显著改善统计机器翻译的译文质量。     

Word co-occurrences on Webpages as a measure of the relatedness of organizations: A new Webometrics concept

Web hyperlink analysis has been a key topic of Webometric research. However, inlink data collection from commercial search engines has been limited to only one source in recent years, which is not a promising prospect for the future development of the field. We need to tap into other Web data sources and to develop new methods. Toward this end, we propose a new Webometrics concept that is based on words rather than inlinks on Webpages. We propose that word co-occurrences on Webpages can be a measure of the relatedness of organizations. Word co-occurrence data can be collected from both general search engines and blog search engines, which expands data sources greatly. The proposed concept is tested in a group of companies in the LTE and WiMax sectors of the telecommunications industry. Data on the co-occurrences of company names on Webpages were collected from Google and Google Blog. The co-occurrence matrices were analyzed using MDS. The resulting MDS maps were compared with industry reality and with the MDS maps from co-link analysis. Results show that Web co-word analysis could potentially be as useful as Web co-link analysis. Google Blog seems to be a better source than Google for co-word data collection.     

与自然语言查询表述相关的词语分析

考察在信息检索过程中用户以自然语言表述的查询语句中的词语使用情况。以一个信息需求描述颗粒度不等的查询表述语料库为素材,辅以汉语通用语料作为对照,通过词频以及词语的文本覆盖率等统计数据,按照是否需要在目标文本中直接或以其他形式出现,将查询表述语句中的词语区分为对汉语文本处理具有普遍意义的通用停用词、服务于信息检索表述用的专用停用词和与特定需求相关的信息内容词语。区分词语使用的不同性质,能为信息系统前端的自然语言查询处理增加一道剥离工序,防止将整个查询语句的分词结果全部作为检索项所造成的效率和准确率的退化。     

基于枢轴语言的平行语料构建方法

平行语料库的规模对于统计机器翻译性能的提高具有重要作用,但是平行语料库的人工构建成本很高。针对这个问题,本文提出了一种低成本高效率的平行语料构建方法,利用枢轴语言作为桥梁,借助已有的机器翻译技术并融合主动学习方法构建目标语言对的大规模高质量平行语料库。本文通过以英语作为枢轴语言构建日汉平行语料库的实例研究,利用成熟的基于短语的统计机器翻译技术,描述了基于译文自动评测的良好译文选择方法、基于主动学习的语料选取方法、以及翻译系统的更新迭代和评价实验。实验结果表明,本文提出的方法能够快速构建日汉平行语料,并有效提高日汉翻译系统的性能。     

基于主题图的英汉跨语言检索模型构建

针对现有跨语言检索模型普遍存在的翻译准确性差、效率低、成本高等不足,在深入分析主题图技术在揭示语词概念之间的语义关系和多语言支持等方面的优越性能的基础上,提出一个基于主题图的英汉跨语言检索模型,该模型采用索引翻译的策略来实现跨语言检索。该模型的突出特点是能够在提高翻译准确性的同时,有效降低翻译成本。此外,实现起来也比较简单。     

基于复合词生成的网络热点话题识别及描述算法

[目的/意义]在海量网络新闻和微博等新媒体文本中自动识别网络热点话题并抽取有意义词串来描述热点事件,对自动识别和描述网络舆情具有重要的研究意义。[方法/过程]在现有热点描述词抽取方法中,利用关联规则或多元词组合方法在抽取过程中存在噪音词较多和特征词语义被放大或转移等问题。本文提出一种基于复合词生成的描述词抽取方法,在所提取的语义更为精确的描述词集合上使用一趟聚类算法对新闻文本进行聚类,自动识别网络热点话题并对热点话题进行排名。[结果/结论]对腾讯新闻事件文本数据集所做的实验结果表明,本文所提出的方法较传统的词特征抽取方法在聚类结果上具有更好的话题簇识别能力和簇描述能力。     

Multilingual Information Retrieval Using Machine Translation, Relevance Feedback and Decompounding

Multilingual retrieval (querying of multiple document collections each in a different language) can be achieved by combining several individual techniques which enhance retrieval: machine translation to cross the language barrier, relevance feedback to add words to the initial query, decompounding for languages with complex term structure, and data fusion to combine monolingual retrieval results from different languages. Using the CLEF 2001 and CLEF 2002 topics and document collections, this paper evaluates these techniques within the context of a monolingual document ranking formula based upon logistic regression. Each individual technique yields improved performance over runs which do not utilize that technique. Moreover the techniques are complementary, in that combining the best techniques outperforms individual technique performance. An approximate but fast document translation using bilingual wordlists created from machine translation systems is presented and evaluated. The fast document translation is as effective as query translation in multilingual retrieval. Furthermore, when fast document translation is combined with query translation in multilingual retrieval, the performance is significantly better than that of query translation or fast document translation.     

Dictionary-Based Cross-Language Information Retrieval: Learning Experiences from CLEF 2000–2002

In this study the basic framework and performance analysis results are presented for the three year long development process of the dictionary-based UTACLIR system. The tests expand from bilingual CLIR for three language pairs Swedish, Finnish and German to English, to six language pairs, from English to French, German, Spanish, Italian, Dutch and Finnish, and from bilingual to multilingual. In addition, transitive translation tests are reported. The development process of the UTACLIR query translation system will be regarded from the point of view of a learning process. The contribution of the individual components, the effectiveness of compound handling, proper name matching and structuring of queries are analyzed. The results and the fault analysis have been valuable in the development process. Overall the results indicate that the process is robust and can be extended to other languages. The individual effects of the different components are in general positive. However, performance also depends on the topic set and the number of compounds and proper names in the topic, and to some extent on the source and target language. The dictionaries used affect the performance significantly.