Mapping queries to the Linking Open Data cloud: A case study using DBpedia

We introduce the task of mapping search engine queries to DBpedia, a major linking hub in the Linking Open Data cloud. We propose and compare various methods for addressing this task, using a mixture of information retrieval and machine learning techniques. Specifically, we present a supervised machine learning-based method to determine which concepts are intended by a user issuing a query. The concepts are obtained from an ontology and may be used to provide contextual information, related concepts, or navigational suggestions to the user submitting the query. Our approach first ranks candidate concepts using a language modeling for information retrieval framework. We then extract query, concept, and search-history feature vectors for these concepts. Using manual annotations we inform a machine learning algorithm that learns how to select concepts from the candidates given an input query. Simply performing a lexical match between the queries and concepts is found to perform poorly and so does using retrieval alone, i.e., omitting the concept selection stage. Our proposed method significantly improves upon these baselines and we find that support vector machines are able to achieve the best performance out of the machine learning algorithms evaluated.     

Training query filtering for semi-supervised learning to rank with pseudo labels

Semi-supervised learning is a machine learning paradigm that can be applied to create pseudo labels from unlabeled data for learning a ranking model, when there is only limited or no training examples available. However, the effectiveness of semi-supervised learning in information retrieval (IR) can be hindered by the low quality pseudo labels, hence the need for the training query filtering that removes the low quality queries. In this paper, we assume two application scenarios with respect to the availability of human labels. First, for applications without any labeled data available, a clustering-based approach is proposed to select the high quality training queries. This approach selects the training queries following the empirical observation that the relevant documents of high quality training queries are highly coherent. Second, for applications with limited labeled data available, a classification-based approach is proposed. This approach learns a weak classifier to predict the retrieval performance gain of a given training query by making use of query features. The queries with high performance gains are selected for the following transduction process to create the pseudo labels for learning to rank algorithms. Experimental results on the standard LETOR dataset show that our proposed approaches outperform the strong baselines.     

双特征空间的实体排序学习

随着大规模知识图谱的出现以及企业高效管理领域知识图谱的需求,知识图谱中的自组织实体检索成为研究热点。给定知识图谱以及用户查询,实体检索的目标在于从给定的知识图谱中返回实体的排序列表。从匹配的角度来看,传统的实体检索模型大都将用户查询和实体统一映射到词的特征空间。这样做具有明显的缺点,例如,将同属于一个实体的两个词视为独立的。为此,该文提出将用户查询和实体同时映射到实体与词两个特征空间方法,称为双特征空间的排序学习。首先将实体抽象成若干个域。之后从词空间和实体空间两个维度分别抽取排序特征,最终应用于排序学习算法中。实验结果表明,在标准数据集上,双特征空间的实体排序学习模型性能显著优于当前先进的实体检索模型。     

面向微博搜索的时间敏感的排序学习方法

近年来微博检索已经成为信息检索领域的研究热点。相关的研究表明,微博检索具有时间敏感性。已有工作根据不同的时间敏感性假设,例如,时间越新文档越相关,或者时间越接近热点时刻文档越相关,得到多种不同的检索模型,都在一定程度上提高了检索效果。但是这些假设主要来自于观察,是一种直观简化的假设,仅能从某个方面反映时间因素影响微博排序的规律。该文验证了微博检索具有复杂的时间敏感特性,直观的简化假设并不能准确地描述这种特性。在此基础上提出了一个利用微博的时间特征和文本特征,通过机器学习的方式来构建一个针对时间敏感的微博检索的排序学习模型(TLTR)。在时间特征上,考察了查询相关的全局时间特征以及查询-文档对的局部时间特征。在TREC Microblog Track 20112012数据集上的实验结果表明,TLTR模型优于现有的其他时间敏感的微博排序方法。     

基于迭代方法的多层Markov网络信息检索模型

查询扩展是提高检索效果的有效方法,传统的查询扩展方法大都以单个查询词的相关性来扩展查询词,没有充分考虑词项之间、文档之间以及查询之间的相关性,使得扩展效果不佳。针对此问题,该文首先通过分别构造词项子空间和文档子空间的Markov网络,用于提取出最大词团和最大文档团,然后根据词团与文档团的映射关系将词团分为文档依赖和非文档依赖词团,并构建基于文档团依赖的Markov网络检索模型做初次检索,从返回的检索结果集合中构造出查询子空间的Markov网络,用于提取出最大查询团,最后,采用迭代的方法计算文档与查询的相关概率,并构建出最终的基于迭代方法的多层Markov网络信息检索模型。实验结果表明 该文的模型能较好地提高检索效果。     

Combining pre-retrieval query quality predictors using genetic programming

Predicting the effectiveness of queries plays an important role in information retrieval. In recent years, a number of methods are proposed for this task, however, there has been little work done on combining multiple predictors. Previous studies on combining multiple predictors rely on non-backtracking based machine learning methods. These studies show minor improvement over single predictors due to the limitation of non-backtracking. This paper discusses work on using machine learning to automatically generate an effective predictors’ combination for query performance prediction. This task is referred to as—learning to predict for query performance prediction in the field. In this paper, a learning method, PredGP, is presented to address this task. PredGP employs genetic programming to learn a predictor by combining various pre-retrieval predictors. The proposed method is evaluated using the TREC Chemical Prior-Art Retrieval Task dataset and found to be significantly better than single predictors.     

基于查询语义特性的稠密文本检索模型

针对现有的稠密文本检索模型（dense passage retrieval,DPR）存在的负采样效率低、易产生过拟合等问题,提出了一种基于查询语义特性的稠密文本检索模型（Q-DPR）。首先,针对模型的负采样过程,提出了一种基于近邻查询的负采样方法。该方法通过检索近邻查询,快速地构建高质量的负相关样本,以降低模型的训练成本。其次,针对模型易产生过拟合的问题,提出了一种基于对比学习的查询自监督方法。该方法通过建立查询间的自监督对比损失,缓解模型对训练标签的过拟合,从而提升模型的检索准确性。Q-DPR在面向开放领域问答的大型数据集MSMARCO上表现优异,取得了0.348的平均倒数排名以及0.975的召回率。实验结果证明,该模型成功地降低了训练的开销,同时也提升了检索的性能。     

NERank+: a graph-based approach for entity ranking in document collections

Most entity ranking research aims to retrieve a ranked list of entities from a Web corpus given a user query. The rank order of entities is determined by the relevance between the query and contexts of entities. However, entities can be ranked directly based on their relative importance in a document collection, independent of any queries. In this paper, we introduce an entity ranking algorithm named NERank+. Given a document collection, NERank+ first constructs a graph model called Topical Tripartite Graph, consisting of document, topic and entity nodes. We design separate ranking functions to compute the prior ranks of entities and topics, respectively. A meta-path constrained random walk algorithm is proposed to propagate prior entity and topic ranks based on the graph model.We evaluate NERank+ over real-life datasets and compare it with baselines. Experimental results illustrate the effectiveness of our approach.     

融合文本概念化与网络表示的观点检索

观点检索是自然语言处理领域中的一个热点研究课题.现有的观点检索模型在检索过程中往往无法根据上下文将词汇进行知识、概念层面的抽象,在语义层面忽略词汇之间的语义联系,观点层面缺乏观点泛化能力.因此,提出一种融合文本概念化与网络表示的观点检索方法.该方法首先利用知识图谱分别将用户查询和文本概念化到正确的概念空间,并利用网络表示将知识图谱中的词汇节点表示成低维向量,然后根据词向量推出查询和文本的向量并用余弦公式计算用户查询与文本的相关度,接着引入基于统计机器学习的分类方法挖掘文本的观点.最后利用概念空间、网络表示空间以及观点分析结果构建特征,并服务于观点检索模型,相关实验表明,本文提出的检索模型可以有效提高多种检索模型的观点检索性能.其中,基于统一相关模型的观点检索方法在两个实验数据集上相比基准方法在MAP评价指标上分别提升了6.1%和9.3%,基于排序学习的观点检索方法在两个实验数据集上相比于基准方法在MAP评价指标上分别提升了2.3%和14.6%.     

Optimal High-Performance Parallel Text Retrieval via Fat-Trees

We present here a high-performance parallel free-text retrieval method for multiple text queries using the Vector Space Model. Our method employs the fat-tree area universal routing network for connecting the processors of a parallel machine, however (in its general form) it could also be efficiently applied over any other high-bandwidth network of processors or workstations. We provide a theoretical analysis of our technique which shows it is excessively efficient and clearly superior (concerning both the amortized processing times and the average waiting times per query) to parallel text retrieval methods for single queries (e.g., based on binary trees). Moreover, we prove our method to be optimal with respect to the execution of all the implied communication tasks on ideal fat-trees. We also experimentally demonstrate the high performance and superiority of our technique via suitable embeddings of ideal fat-trees on realistic two-dimensional mesh-oriented parallel machines (e.g., the GCel Parsytec machine) and via the use of the large-scale TREC document collections. Note that the fat-tree can simulate any other network of the same hardware with only a polylogarithmic loss of efficiency. Received May 1997, and in final form April 21, 1999.     

QoRank: A query‐dependent ranking model using LSE‐based weighted multiple hyperplanes aggregation for information retrieval

Ranking is a core problem for information retrieval since the performance of the search system is directly impacted by the accuracy of ranking results. Ranking model construction has been the focus of both the fields of information retrieval and machine learning, and learning to rank in particular has attracted much interest. Many ranking models have been proposed, for example, RankSVM is a state‐of‐the‐art method for learning to rank and has been empirically demonstrated to be effective. However, most of the proposed methods do not consider about the significant differences between queries, only resort to a single function in ranking. In this paper, we present a novel ranking model named QoRank, which performs the learning task dependent on queries. We also propose a LSE (least‐squares estimation) ‐based weighted method to aggregate the ranking lists produced by base decision functions as the final ranking. Comparison of QoRank with other ranking techniques is conducted, and several evaluation criteria are employed to evaluate its performance. Experimental results on the LETOR OHSUMED data set show that QoRank strikes a good balance of accuracy and complexity, and outperforms the baseline methods. © 2010 Wiley Periodicals, Inc.     

Beyond Single-Page Web Search Results

Given a user keyword query, current Web search engines return a list of individual Web pages ranked by their "goodness" with respect to the query. Thus, the basic unit for search and retrieval is an individual page, even though information on a topic is often spread across multiple pages. This degrades the quality of search results, especially for long or uncorrelated (multitopic) queries (in which individual keywords rarely occur together in the same document), where a single page is unlikely to satisfy the user's information need. We propose a technique that, given a keyword query, on the fly generates new pages, called composed pages, which contain all query keywords. The composed pages are generated by extracting and stitching together relevant pieces from hyperlinked Web pages and retaining links to the original Web pages. To rank the composed pages, we consider both the hyperlink structure of the original pages and the associations between the keywords within each page. Furthermore, we present and experimentally evaluate heuristic algorithms to efficiently generate the top composed pages. The quality of our method is compared to current approaches by using user surveys. Finally, we also show how our techniques can be used to perform query-specific summarization of Web pages.     

Cross-media retrieval using query dependent search methods

The content-based cross-media retrieval is a new type of multimedia retrieval in which the media types of query examples and the returned results can be different. In order to learn the semantic correlations among multimedia objects of different modalities, the heterogeneous multimedia objects are analyzed in the form of multimedia document (MMD), which is a set of multimedia objects that are of different media types but carry the same semantics. We first construct an MMD semi-semantic graph (MMDSSG) by jointly analyzing the heterogeneous multimedia data. After that, cross-media indexing space (CMIS) is constructed. For each query, the optimal dimension of CMIS is automatically determined and the cross-media retrieval is performed on a per-query basis. By doing this, the most appropriate retrieval approach for each query is selected, i.e. different search methods are used for different queries. The query dependent search methods make cross-media retrieval performance not only accurate but also stable. We also propose different learning methods of relevance feedback (RF) to improve the performance. Experiment is encouraging and validates the proposed methods.     

A new method for fuzzy information retrieval based on fuzzy hierarchical clustering and fuzzy inference techniques

In this paper, we extend the work of Kraft et al. to present a new method for fuzzy information retrieval based on fuzzy hierarchical clustering and fuzzy inference techniques. First, we present a fuzzy agglomerative hierarchical clustering algorithm for clustering documents and to get the document cluster centers of document clusters. Then, we present a method to construct fuzzy logic rules based on the document clusters and their document cluster centers. Finally, we apply the constructed fuzzy logic rules to modify the user's query for query expansion and to guide the information retrieval system to retrieve documents relevant to the user's request. The fuzzy logic rules can represent three kinds of fuzzy relationships (i.e., fuzzy positive association relationship, fuzzy specialization relationship and fuzzy generalization relationship) between index terms. The proposed fuzzy information retrieval method is more flexible and more intelligent than the existing methods due to the fact that it can expand users' queries for fuzzy information retrieval in a more effective manner.     

An adaptive learning automata-based ranking function discovery algorithm

Due to the massive amount of heterogeneous information on the web, insufficient and vague user queries, and use of the same query by different users for different aims, the information retrieval process deals with a huge amount of uncertainty and doubt. Under such circumstances, designing an efficient retrieval function and ranking algorithm by which the most relevant results are provided is of the greatest importance. In this paper, a learning automata-based ranking function discovery algorithm in which different sources of information are combined is proposed. In this method, the learning automaton is used to adjust the portion of the final ranking that is assigned to each source of evidence based on the user feedback. All sources of information are first given the same importance. The proportion of a given source increases, if the documents provided by this source are reviewed by the user and decreases otherwise. As the proposed algorithm proceeds, the probability of appearance of each source in the final ranking gets proportional to its relevance to the user queries. Several simulation experiments are conducted on well-known data collections and query types to show the performance of the proposed algorithm. The obtained results demonstrate that the proposed algorithm outperforms several existing methods in terms of precision at position n, mean average precision, and normalized discount cumulative gain.     

Cross-language plagiarism detection

Cross-language plagiarism detection deals with the automatic identification and extraction of plagiarism in a multilingual setting. In this setting, a suspicious document is given, and the task is to retrieve all sections from the document that originate from a large, multilingual document collection. Our contributions in this field are as follows: (1) a comprehensive retrieval process for cross-language plagiarism detection is introduced, highlighting the differences to monolingual plagiarism detection, (2) state-of-the-art solutions for two important subtasks are reviewed, (3) retrieval models for the assessment of cross-language similarity are surveyed, and, (4) the three models CL-CNG, CL-ESA and CL-ASA are compared. Our evaluation is of realistic scale: it relies on 120,000 test documents which are selected from the corpora JRC-Acquis and Wikipedia, so that for each test document highly similar documents are available in all of the six languages English, German, Spanish, French, Dutch, and Polish. The models are employed in a series of ranking tasks, and more than 100 million similarities are computed with each model. The results of our evaluation indicate that CL-CNG, despite its simple approach, is the best choice to rank and compare texts across languages if they are syntactically related. CL-ESA almost matches the performance of CL-CNG, but on arbitrary pairs of languages. CL-ASA works best on “exact” translations but does not generalize well.     

Text plagiarism classification using syntax based linguistic features

The proposed work models document level text plagiarism detection as a binary classification problem, where the task is to distinguish a given suspicious-source document pair as plagiarized or non-plagiarized. The objective is to explore the potency of syntax based linguistic features extracted using shallow natural language processing techniques for plagiarism classification task. Shallow syntactic features, viz., part of speech tags and chunks are utilized after effective pre-processing and filtrations for pruning the irrelevant information. The work further proposes the modelling of this classification phase as an intermediate stage, which will be post candidate source retrieval and before exhaustive passage level detections. A two-phase feature selection approach is proposed, which improves the effectiveness of classification by selecting appropriate set of features as the input to machine learning based classifiers. The proposed approach is evaluated on smaller and larger test conditions using the corpus of plagiarized short answers (PSA) and plagiarism instances collected from PAN corpus respectively. Under both the test conditions, performances are evaluated using general as well as advanced classification metrics. Another main contribution of the current work is the analysis of dependencies and impact of the extracted features, upon the type and complexity of plagiarism imposed in the documents. The proposed results are compared with the two state-of-the-art approaches and they outperform the baseline approaches significantly. This in turn reflects the cogency of syntactic linguistic features in document level plagiarism classification, especially for the instances close to manual or real plagiarism scenarios.     

Random walk based rank aggregation to improving web search

In Web search, with the aid of related query recommendation, Web users can revise their initial queries in several serial rounds in pursuit of finding needed Web pages. In this paper, we address the Web search problem on aggregating search results of related queries to improve the retrieval quality. Given an initial query and the suggested related queries, our search system concurrently processes their search result lists from an existing search engine and then forms a single list aggregated by all the retrieved lists. We specifically propose a generic rank aggregation framework which consists of three steps. First we build a so-called Win/Loss graph of Web pages according to a competition rule, and then apply the random walk mechanism on the Win/Loss graph. Last we sort these Web pages by their ranks using a PageRank-like rank mechanism. The proposed framework considers not only the number of wins that an item won in competitions, but also the quality of its competitor items in calculating the ranking of Web page items. Experimental results show that our search system can clearly improve the retrieval quality in a parallel manner over the traditional search strategy that serially returns result lists. Moreover, we also provide empirical evidences as to demonstrate how different rank aggregation methods affect the retrieval quality.     

基于对抗式数据增强的深度文本检索重排序

在信息检索领域的排序任务中, 神经网络排序模型已经得到广泛使用. 神经网络排序模型对于数据的质量要求极高, 但是, 信息检索数据集通常含有较多噪音, 不能精确得到与查询不相关的文档. 为了训练一个高性能的神经网络排序模型, 获得高质量的负样本, 则至关重要. 借鉴现有方法doc2query的思想, 本文提出了深度、端到端的模型AQGM, 通过学习不匹配查询文档对, 生成与文档不相关、原始查询相似的对抗查询, 增加了查询的多样性,增强了负样本的质量. 本文利用真实样本和AQGM模型生成的样本, 训练基于BERT的深度排序模型, 实验表明,与基线模型BERT-base对比, 本文的方法在MSMARCO和TrecQA数据集上, MRR指标分别提升了0.3％和3.2％.     

Improving retrievability with improved cluster-based pseudo-relevance feedback selection

High findability of documents within a certain cut-off rank is considered an important factor in recall-oriented application domains such as patent or legal document retrieval. Findability is hindered by two aspects, namely the inherent bias favoring some types of documents over others introduced by the retrieval model, and the failure to correctly capture and interpret the context of conventionally rather short queries. In this paper, we analyze the bias impact of different retrieval models and query expansion strategies. We furthermore propose a novel query expansion strategy based on document clustering to identify dominant relevant documents. This helps to overcome limitations of conventional query expansion strategies that suffer strongly from the noise introduced by imperfect initial query results for pseudo-relevance feedback documents selection. Experiments with different collections of patent documents suggest that clustering based document selection for pseudo-relevance feedback is an effective approach for increasing the findability of individual documents and decreasing the bias of a retrieval system.