Hallucination remains a major obstacle to the domain generalizability and reliability of Large Language Models (LLMs). Recent approaches address this issue by integrating Retrieval-Augmented Generation (RAG) with stepwise reasoning processes to iteratively retrieve knowledge. However, indiscriminate incorporation of external knowledge may interfere with reasoning, increasing latency and amplifying error accumulation. Moreover, existing methods rely on a unidirectional flow of external knowledge into LLMs while neglecting internal–external knowledge synergy, limiting autonomous reasoning capability. To address these limitations, we propose the Reasoning–Evidence Interactive Augmented Graph (RE-IAG), a framework that couples reasoning with evidence through a staged triggering mechanism and structured interaction. RE-IAG performs localized refinement of intermediate reasoning via adaptive branching under uncertainty and selectively triggers retrieval when internal reasoning stagnates. Crucially, it organizes both internal reasoning and retrieved evidence into aligned graph structures, enabling structure-guided verification and fine-grained refinement of intermediate conclusions. This design transforms retrieval from passive augmentation into an active constraint on reasoning, reducing error propagation, alleviating knowledge conflicts, and improving knowledge integration for hallucination mitigation. Extensive experiments on four multi-hop QA benchmarks show that RE-IAG outperforms adaptive RAG baselines, achieves competitive or superior performance to RL-based approaches, and demonstrates strong robustness and generalization across model scales and architectures.

Accurately capturing a user's dynamic search intent based on her/his interactions with the system is crucial for improving the performance of session-based search. Existing methods often require the entire interaction sequence within a session to be recomputed continuously at each interaction step, and the token-level interactions are either captured within an overall transformer structure or simply ignored. As a consequence, the current approaches suffer from an increased computation burden and fall short of accurately capturing the dynamic evolution of user intent. In this paper, we propose a novel representation approach which treats both search intent and candidate documents as dimension-specific probability distributions of token embedding representations. Based on this representation, we propose an Dynamic Interaction-Driven intent Evolver (DIDE) for dynamically updating the user's search intent throughout a session with a lightweight similarity calculation method for document ranking. Comprehensive experimental results demonstrate that DIDE adeptly captures the dynamic nature of session-based search and significantly outperforms a range of strong baseline models across three different datasets.

Sentiment and emotion, which correspond to long-term and short-lived human feelings, are closely linked to each other, leading to the fact that sentiment analysis and emotion recognition are also two interdependent tasks in natural language processing (NLP). One task often leverages the shared knowledge from another task and performs better when solved in a joint learning paradigm. Conversational context dependency, multi-modal interaction, and multi-task correlation are three key factors that contribute to this joint paradigm. However, none of the recent approaches have considered them in a unified framework. To fill this gap, we propose a multi-modal, multi-task interactive graph attention network, termed M3GAT, to simultaneously solve the three problems. At the heart of the model is a proposed interactive conversation graph layer containing three core sub-modules, which are: (1) local-global context connection for modeling both local and global conversational context, (2) cross-modal connection for learning multi-modal complementary and (3) cross-task connection for capturing the correlation across two tasks. Comprehensive experiments on three benchmarking datasets, MELD, MEISD, and MSED, show the effectiveness of M3GAT over state-of-the-art baselines with the margin of 1.88%, 5.37%, and 0.19% for sentiment analysis, and 1.99%, 3.65%, and 0.13% for emotion recognition, respectively. In addition, we also show the superiority of multi-task learning over the single-task framework.

Large-scale pretrained language models have achieved compelling performance in a wide range of language understanding and information retrieval tasks. While their large scales ensure capacity, they also hinder deployment. Knowledge distillation offers an opportunity to compress a large language model to a small one, in order to reach a reasonable latency-performance tradeoff. However, for scenarios where the number of requests (e.g., queries submitted to a search engine) is highly variant, the static tradeoff attained by the compressed language model might not always fit. Once a model is assigned with a static tradeoff, it could be inadequate in that the latency is too high when the number of requests is large, or the performance is too low when the number of requests is small. To this end, we propose an elastic language model ( ElasticLM ) that elastically adjusts the tradeoff according to the request stream. The basic idea is to introduce a compute elasticity to the compressed language model, so that the tradeoff could vary on-the-fly along a scalable and controllable compute. Specifically, we impose an elastic structure to equip ElasticLM with compute elasticity and design an elastic optimization method to learn ElasticLM under compute elasticity. To serve ElasticLM , we apply an elastic schedule. Considering the specificity of information retrieval, we adapt ElasticLM to dense retrieval and reranking, and present an ElasticDenser and an ElasticRanker, respectively. Offline evaluation is conducted on a language understanding benchmark GLUE, and several information retrieval tasks including Natural Question, Trivia QA and MS MARCO. The results show that ElasticLM along with ElasticDenser and ElasticRanker can perform correctly and competitively compared with an array of static baselines. Furthermore, an online simulation with concurrency is also carried out. The results demonstrate that ElasticLM can provide elastic tradeoffs with respect to varying request stream.

Neural text matching models have been used in a range of applications such as question answering and natural language inference, and have yielded a good performance. However, these neural models are of a limited adaptability, resulting in a decline in performance when encountering test examples from a different dataset or even a different task. The adaptability is particularly important in the few-shot setting: in many cases, there is only a limited amount of labeled data available for a target dataset or task, while we may have access to a richly labeled source dataset or task. However, adapting a model trained on the abundant source data to a few-shot target dataset or task is challenging. To tackle this challenge, we propose a Meta-Weight Regulator (MWR), which is a meta-learning approach that learns to assign weights to the source examples based on their relevance to the target loss. Specifically, MWR first trains the model on the uniformly weighted source examples, and measures the efficacy of the model on the target examples via a loss function. By iteratively performing a (meta) gradient descent, high-order gradients are propagated to the source examples. These gradients are then used to update the weights of source examples, in a way that is relevant to the target performance. As MWR is model-agnostic, it can be applied to any backbone neural model. Extensive experiments are conducted with various backbone text matching models, on four widely used datasets and two tasks. The results demonstrate that our proposed approach significantly outperforms a number of existing adaptation methods and effectively improves the cross-dataset and cross-task adaptability of the neural text matching models in the few-shot setting.

Language modeling is essential in Natural Language Processing and Information Retrieval related tasks. After the statistical language models, Quantum Language Model (QLM) has been proposed to unify both single words and compound terms in the same probability space without extending term space exponentially. Although QLM achieved good performance in ad hoc retrieval, it still has two major limitations: (1) QLM cannot make use of supervised information, mainly due to the iterative and non-differentiable estimation of the density matrix, which represents both queries and documents in QLM. (2) QLM assumes the exchangeability of words or word dependencies, neglecting the order or position information of words. This article aims to generalize QLM and make it applicable to more complicated matching tasks (e.g., Question Answering) beyond ad hoc retrieval. We propose a complex-valued neural network-based QLM solution called C-NNQLM to employ an end-to-end approach to build and train density matrices in a light-weight and differentiable manner, and it can therefore make use of external well-trained word vectors and supervised labels. Furthermore, C-NNQLM adopts complex-valued word vectors whose phase vectors can directly encode the order (or position) information of words. Note that complex numbers are also essential in the quantum theory. We show that the real-valued NNQLM (R-NNQLM) is a special case of C-NNQLM. The experimental results on the QA task show that both R-NNQLM and C-NNQLM achieve much better performance than the vanilla QLM, and C-NNQLM’s performance is on par with state-of-the-art neural network models. We also evaluate the proposed C-NNQLM on text classification and document retrieval tasks. The results on most datasets show that the C-NNQLM can outperform R-NNQLM, which demonstrates the usefulness of the complex representation for words and sentences in C-NNQLM.

An essential task of information retrieval (IR) is to compute the probability of relevance of a document given a query. If we regard a query term or n-gram fragment as a relevance matching unit, most retrieval models firstly calculate the relevance evidence between the given query and the candidate document separately, and then accumulate these evidences as the final document relevance prediction. This kind of approach obeys the the classical probability, which is not fully consistent with human cognitive rules in the actual retrieval process, due to the possible existence of interference effect between relevance matching units. In our work, we propose a Quantum Interference inspired Neural Matching model (QINM), which can apply the interference effects to guide the construction of additional evidence generated by the interaction between matching units in the retrieval process. Experimental results on two benchmark collections demonstrate that our approach outperforms the quantum-inspired retrieval models, and some well-known neural retrieval models in the ad-hoc retrieval task.

Synchronous collaborative search systems (SCSS) refer to systems which support two or more users with similar information need to search together simultaneously. Generally, SCSS provide a social engine to enable users to communicate. However, when the number of users in the social engine is insufficient to collaborate on the search task, the social engine will encounter the cold start problem and can not perform collaborative search well. In this paper, we present a novel Synchronous Collaborative Search System with a Live Interactive Engine (SCSS-LIE). SCSS-LIE proposes to apply a ring topology to add an intelligent auxiliary robot, Infobot, into the social engine to support real-time interaction between users and the search engine to address the cold start problem of the social engine. The reading comprehension model BiDAF (Bi-Directional Attention Flow) is employed in the Infobot in the process of interacting with the search engine to obtain answers to facilitate the acquisition of information. SCSS-LIE can not only allow users with similar information need to be grouped into one chat channel to communicate, but also enable them to conduct real-time interaction with the search engine to improve search efficiency.

It has been widely accepted that Long Short-Term Memory (LSTM) network, coupled with attention mechanism and memory module, is useful for aspect-level sentiment classification. However, existing approaches largely rely on the modelling of semantic relatedness of an aspect with its context words, while to some extent ignore their syntactic dependencies within sentences. Consequently, this may lead to an undesirable result that the aspect attends on contextual words that are descriptive of other aspects. In this paper, we propose a proximity-weighted convolution network to offer an aspect-specific syntax-aware representation of contexts. In particular, two ways of determining proximity weight are explored, namely position proximity and dependency proximity. The representation is primarily abstracted by a bidirectional LSTM architecture and further enhanced by a proximity-weighted convolution. Experiments conducted on the SemEval 2014 benchmark demonstrate the effectiveness of our proposed approach compared with a range of state-of-the-art models is available at https://github.com/GeneZC/PWCN.

Capturing the meaning of sentences has long been a challenging task. Current models tend to apply linear combinations of word features to conduct semantic composition for bigger-granularity units e.g. phrases, sentences, and documents. However, the semantic linearity does not always hold in human language. For instance, the meaning of the phrase “ivory tower” cannot be deduced by linearly combining the meanings of “ivory” and “tower”. To address this issue, we propose a new framework that models different levels of semantic units (e.g. sememe, word, sentence, and semantic abstraction) on a single Semantic Hilbert Space, which naturally admits a non-linear semantic composition by means of a complex-valued vector word representation. An end-to-end neural network 1 is proposed to implement the framework in the text classification task, and evaluation results on six benchmarking text classification datasets demonstrate the effectiveness, robustness and self-explanation power of the proposed model. Furthermore, intuitive case studies are conducted to help end users to understand how the framework works.

The recently proposed quantum language model (QLM) aimed at a principled approach to modeling term dependency by applying the quantum probability theory. The latest development for a more effective QLM has adopted word embeddings as a kind of global dependency information and integrated the quantum-inspired idea in a neural network architecture. While these quantum-inspired LMs are theoretically more general and also practically effective, they have two major limitations. First, they have not taken into account the interaction among words with multiple meanings, which is common and important in understanding natural language text. Second, the integration of the quantum-inspired LM with the neural network was mainly for effective training of parameters, yet lacking a theoretical foundation accounting for such integration. To address these two issues, in this paper, we propose a Quantum Many-body Wave Function (QMWF) inspired language modeling approach. The QMWF inspired LM can adopt the tensor product to model the aforesaid interaction among words. It also enables us to reveal the inherent necessity of using Convolutional Neural Network (CNN) in QMWF language modeling. Furthermore, our approach delivers a simple algorithm to represent and match text/sentence pairs. Systematic evaluation shows the effectiveness of the proposed QMWF-LM algorithm, in comparison with the state of the art quantum-inspired LMs and a couple of CNN-based methods, on three typical Question Answering (QA) datasets.

It has been shown that relevance judgment of documents is influenced by multiple factors beyond topicality. Some multidimensional user relevance models (MURM) proposed in literature have investigated the impact of different dimensions of relevance on user judgment. Our hypothesis is that a user might give more importance to certain relevance dimensions in a session which might change dynamically as the session progresses. This motivates the need to capture the weights of different relevance dimensions using feedback and build a model to rank documents for subsequent queries according to these weights. We propose a geometric model inspired by the mathematical framework of Quantum theory to capture the user's importance given to each dimension of relevance and test our hypothesis on data from a web search engine and TREC Session track

Under the notion that the document collection is a sample from a population, the observed per-topic metric (e.g., AP) value varies with different samples, leading to the per-topic variance. The results of the system comparison, such as comparing the ranking of systems according to the summary metric (e.g., MAP) or testing whether there is significant difference between two systems, are affected by the variability of per-topic metric values. In this paper, we study the effect of per-topic variance on the system comparison. To measure such effects, we employ two ranking-based methods, i.e., Error Rate (ER) and Kendall Rank Correlation Coefficient (KRCC), as well as two significance test based methods, namely Achieved Significance Level (ASL) and Estimated Difference (ED). We conduct empirical comparison of TREC participated systems on Robust and Adhoc track, which shows that the effect of per-topic variance on the ranking of systems is not obvious, while the significance test based comparisons are susceptible to the per-topic variance.

Recent research has shown the usefulness of using collective user interaction data (e.g., query logs) to recommend query modification suggestions for Intranet search. However, most of the query suggestion approaches for Intranet search follow an ``one size fits all'' strategy, whereby different users who submit an identical query would get the same query suggestion list. This is problematic, as even with the same query, different users may have different topics of interest, which may change over time in response to the user's interaction with the system.

Modeling multidimensional relevance in information retrieval (IR) has attracted much attention in recent years. However, most existing studies are conducted through relatively small‐scale user studies, which may not reflect a real‐world and natural search scenario. In this article, we propose to study the multidimensional user relevance model (MURM) on large scale query logs, which record users' various search behaviors (e.g., query reformulations, clicks and dwelling time, etc.) in natural search settings. We advance an existing MURM model (including five dimensions: topicality, novelty, reliability, understandability, and scope) by providing two additional dimensions, that is, interest and habit. The two new dimensions represent personalized relevance judgment on retrieved documents. Further, for each dimension in the enriched MURM model, a set of computable features are formulated. By conducting extensive document ranking experiments on Bing's query logs and TREC session Track data, we systematically investigated the impact of each dimension on retrieval performance and gained a series of insightful findings which may bring benefits for the design of future IR systems.

In many research and application areas, such as information retrieval and machine learning, we often encounter dealing with a probability distribution that is mixed by one distribution that is relevant to our task in hand and the other that is irrelevant and that we want to get rid of. Thus, it is an essential problem to separate the irrelevant distribution from the mixture distribution. This article is focused on the application in Information Retrieval, where relevance feedback is a widely used technique to build a refined query model based on a set of feedback documents. However, in practice, the relevance feedback set, even provided by users explicitly or implicitly, is often a mixture of relevant and irrelevant documents. Consequently, the resultant query model (typically a term distribution) is often a mixture rather than a true relevance term distribution, leading to a negative impact on the retrieval performance. To tackle this problem, we recently proposed a Distribution Separation Method (DSM), which aims to approximate the true relevance distribution by separating a seed irrelevance distribution from the mixture one. While it achieved a promising performance in an empirical evaluation with simulated explicit irrelevance feedback data, it has not been deployed in the scenario where one should automatically obtain the irrelevance feedback data. In this article, we propose a substantial extension of the basic DSM from two perspectives: developing a further regularization framework and deploying DSM in the automatic irrelevance feedback scenario. Specifically, in order to avoid the output distribution of DSM drifting away from the true relevance distribution when the quality of seed irrelevant distribution (as the input to DSM) is not guaranteed, we propose a DSM regularization framework to constrain the estimation for the relevance distribution. This regularization framework includes three algorithms, each corresponding to a regularization strategy incorporated in the objective function of DSM. In addition, we exploit DSM in automatic (i.e., pseudo) irrelevance feedback, by automatically detecting the seed irrelevant documents via three different document reranking methods. We have carried out extensive experiments based on various TREC datasets, in order to systematically evaluate the proposed methods. The experimental results demonstrate the effectiveness of our proposed approaches in comparison with various strong baselines.

Traditional information retrieval systems rank documents according to their relevance to users' input queries. State of the art commercial search engines (SEs) train ranking models and suggest query refinements by exploiting collective intelligence implicitly using global users' query logs. However, they do not provide an explicit channel for users to communicate with each other in the search process. By asking or discussing with other users on the fly, a user could find relevant information more conveniently and gain a better search experience. In this paper, we present a demo of novel Search Engine with a live Chat Channel (SECC). SECC can group users automatically based on their input queries and allow them to communicate with each other in real time through a chat interface.

Formulating and reformulating reliable textual queries have been recognized as a challenging task in Information Retrieval (IR), even for experienced users. Most existing query expansion methods, especially those based on implicit relevance feedback, utilize the user's historical interaction data, such as clicks, scrolling and viewing time on documents, to derive a refined query model. It is further expected that the user's search experience would be largely improved if we could dig out user's latent query intention, in real-time, by capturing the user's current interaction at the term level directly. In this paper, we propose a real-time eye tracking based query expansion method, which is able to: (1) automatically capture the terms that the user is viewing by utilizing eye tracking techniques; (2) derive the user's latent intent based on the eye tracking terms and by using the Latent Dirichlet Allocation (LDA) approach. A systematic user study has been carried out and the experimental results demonstrate the effectiveness of our proposed methods.

The quantum probabilistic framework has recently been applied to Information Retrieval (IR). A representative is the Quantum Language Model (QLM), which is developed for the ad-hoc retrieval with single queries and has achieved significant improvements over traditional language models. In QLM, a density matrix, defined on the quantum probabilistic space, is estimated as a representation of user's search intention with respect to a specific query. However, QLM is unable to capture the dynamics of user's information need in query history. This limitation restricts its further application on the dynamic search tasks, e.g., session search. In this paper, we propose a Session-based Quantum Language Model (SQLM) that deals with multi-query session search task. In SQLM, a transformation model of density matrices is proposed to model the evolution of user's information need in response to the user's interaction with search engine, by incorporating features extracted from both positive feedback (clicked documents) and negative feedback (skipped documents). Extensive experiments conducted on TREC 2013 and 2014 session track data demonstrate the effectiveness of SQLM in comparison with the classic QLM.

Recent research has shown that the improvement of mean retrieval effectiveness (e.g., MAP) may sacrifice the retrieval stability across queries, implying a tradeoff between effectiveness and stability. The evaluation of both effectiveness and stability are often based on a baseline model, which could be weak or biased. In addition, the effectiveness-stability tradeoff has not been systematically or quantitatively evaluated over TREC participated systems. The above two problems, to some extent, limit our awareness of such tradeoff and its impact on developing future IR models. In this paper, motivated by a recently proposed bias-variance based evaluation, we adopt a strong and unbiased "baseline", which is a virtual target model constructed by the best performance (for each query) among all the participated systems in a retrieval task. We also propose generalized bias-variance metrics, based on which a systematic and quantitative evaluation of the effectiveness-stability tradeoff is carried out over the participated systems in the TREC Ad-hoc Track (1993-1999) and Web Track (2010-2012). We observe a clear effectiveness-stability tradeoff, with a trend of becoming more obvious in more recent years. This implies that when we pursue more effective IR systems over years, the stability has become problematic and could have been largely overlooked.

Recent research has shown that the performance of search engines can be improved by enriching a user's personal profile with information about other users with shared interests. In the existing approaches, groups of similar users are often statically determined, e.g., based on the common documents that users clicked. However, these static grouping methods are query-independent and neglect the fact that users in a group may have different interests with respect to different topics. In this paper, we argue that common interest groups should be dynamically constructed in response to the user's input query. We propose a personalisation framework in which a user profile is enriched using information from other users dynamically grouped with respect to an input query. The experimental results on query logs from a major commercial web search engine demonstrate that our framework improves the performance of the web search engine and also achieves better performance than the static grouping method.

This paper reports on an approach to the analysis of form (layout and formatting) during genre recognition recorded using eye tracking. The researchers focused on eight different types of e-mail, such as calls for papers, newsletters and spam, which were chosen to represent different genres. The study involved the collection of oculographic behavior data based on the scanpath duration and scanpath length based metric, to highlight the ways in which people view the features of genres. We found that genre analysis based on purpose and form (layout features, etc.) was an effective means of identifying the characteristics of these e-mails. The research, carried out on a group of 24 participants, highlighted their interaction and interpretation of the e-mail texts and the visual cues or features perceived. In addition, the ocular strategies of scanning and skimming, they employed for the processing of the texts by block, genre and representation were evaluated.

It has been recognized that, when an information retrieval (IR) system achieves improvement in mean retrieval effectiveness (e.g. mean average precision (MAP)) over all the queries, the performance (e.g., average precision (AP)) of some individual queries could be hurt, resulting in retrieval instability. Some stability/robustness metrics have been proposed. However, they are often defined separately from the mean effectiveness metric. Consequently, there is a lack of a unified formulation of effectiveness, stability and overall retrieval quality (considering both). In this paper, we present a unified formulation based on the bias-variance decomposition. Correspondingly, a novel evaluation methodology is developed to evaluate the effectiveness and stability in an integrated manner. A case study applying the proposed methodology to evaluation of query language modeling illustrates the usefulness and analytical power of our approach.

Modern search engines have been moving away from simplistic interfaces that aimed at satisfying a user's need with a single‐shot query. Interactive features are now integral parts of web search engines. However, generating good query modification suggestions remains a challenging issue. Query log analysis is one of the major strands of work in this direction. Although much research has been performed on query logs collected on the web as a whole, query log analysis to enhance search on smaller and more focused collections has attracted less attention, despite its increasing practical importance. In this article, we report on a systematic study of different query modification methods applied to a substantial query log collected on a local website that already uses an interactive search engine. We conducted experiments in which we asked users to assess the relevance of potential query modification suggestions that have been constructed using a range of log analysis methods and different baseline approaches. The experimental results demonstrate the usefulness of log analysis to extract query modification suggestions. Furthermore, our experiments demonstrate that a more fine‐grained approach than grouping search requests into sessions allows for extraction of better refinement terms from query log files.

The classical bag-of-word models for information retrieval (IR) fail to capture contextual associations between words. In this article, we propose to investigate pure high-order dependence among a number of words forming an unseparable semantic entity, that is, the high-order dependence that cannot be reduced to the random coincidence of lower-order dependencies. We believe that identifying these pure high-order dependence patterns would lead to a better representation of documents and novel retrieval models. Specifically, two formal definitions of pure dependence—unconditional pure dependence (UPD) and conditional pure dependence (CPD)—are defined. The exact decision on UPD and CPD, however, is NP-hard in general. We hence derive and prove the sufficient criteria that entail UPD and CPD, within the well-principled information geometry (IG) framework, leading to a more feasible UPD/CPD identification procedure. We further develop novel methods for extracting word patterns with pure high-order dependence. Our methods are applied to and extensively evaluated on three typical IR tasks: text classification and text retrieval without and with query expansion.

Most of the state-of-art approaches to Query-by-Example (QBE) video retrieval are based on the Bag-of-visual-Words (BovW) representation of visual content. It, however, ignores the spatial-temporal information, which is important for similarity measurement between videos. Direct incorporation of such information into the video data representation for a large scale data set is computationally expensive in terms of storage and similarity measurement. It is also static regardless of the change of discriminative power of visual words for different queries. To tackle these limitations, in this paper, we propose to discover Spatial-Temporal Correlations (STC) imposed by the query example to improve the BovW model for video retrieval. The STC, in terms of spatial proximity and relative motion coherence between different visual words, is crucial to identify the discriminative power of the visual words. We develop a novel technique to emphasize the most discriminative visual words for similarity measurement, and incorporate this STC-based approach into the standard inverted index architecture. Our approach is evaluated on the TRECVID2002 and CC\_WEB\_VIDEO datasets for two typical QBE video retrieval tasks respectively. The experimental results demonstrate that it substantially improves the BovW model as well as a state of the art method that also utilizes spatial-temporal information for QBE video retrieval.

A concept hierarchy created from a document collection can be used for query recommendation on Intranets by ranking terms according to the strength of their links to the query within the hierarchy. A major limitation is that this model produces the same recommendations for identical queries and rebuilding it from scratch periodically can be extremely inefficient due to the high computational costs. We propose to adapt the model by incorporating query refinements from search logs. Our intuition is that the concept hierarchy built from the collection and the search logs provide complementary conceptual views on the same search domain, and their integration should continually improve the effectiveness of recommended terms. Two adaptation approaches using query logs with and without click information are compared. We evaluate the concept hierarchy models (static and adapted versions) built from the Intranet collections of two academic institutions and compare them with a state-of-the-art log-based query recommender, the Query Flow Graph, built from the same logs. Our adaptive model significantly outperforms its static version and the query flow graph when tested over a period of time on data (documents and search logs) from two institutions' Intranets.

The "Bag of Visual Words" (BoW) framework has been widely used in query-by-example video retrieval to model the visual content by a set of quantized local feature descriptors. In this paper, we propose a novel technique to enhance BoW by the selection of Word-of-Interest (WoI) that utilizes the quantified temporal motion coherence of the visual words between the adjacent frames in the query example. Experiments carried out using TRECVID datasets show that our technique improves the retrieval performance of the classical BoW-based approach.

Both similarity-based and popularity-based document ranking functions have been successfully applied to information retrieval (IR) in general. However, the dimension of semantic granularity also should be considered for effective retrieval. In this article, we propose a semantic granularity-based IR model that takes into account the three dimensions, namely similarity, popularity, and semantic granularity, to improve domain-specific search. In particular, a concept-based computational model is developed to estimate the semantic granularity of documents with reference to a domain ontology. Semantic granularity refers to the levels of semantic detail carried by an information item. The results of our benchmark experiments confirm that the proposed semantic granularity based IR model performs significantly better than the similarity-based baseline in both a bio-medical and an agricultural domain. In addition, a series of user-oriented studies reveal that the proposed document ranking functions resemble the implicit ranking functions exercised by humans. The perceived relevance of the documents delivered by the granularity-based IR system is significantly higher than that produced by a popular search engine for a number of domain-specific search tasks. To the best of our knowledge, this is the first study regarding the application of semantic granularity to enhance domain-specific IR.

In image retrieval, most existing approaches that incorporate local features produce high dimensional vectors, which lead to a high computational and data storage cost. Moreover, when it comes to the retrieval of generic real-life images, randomly generated patches are often more discriminant than the ones produced by corner/blob detectors. In order to tackle these problems, we propose a novel method incorporating local features with a hybrid sampling (a combination of detector-based and random sampling). We take three large data collections for the evaluation: MIRFlickr, ImageCLEF, and a collection from British National Geological Survey. The overall performance of the proposed approach is better than the performance of global features and comparable with the current state-of-the-art methods in content-based image retrieval. One of the advantages of our method when compared with others is its easy implementation and low computational cost. Another is that hybrid sampling can improve the performance of other methods based on the ``bag of visual words'' approach.

While much research has been performed on query logs collected for major Web search engines, query log analysis to enhance search on smaller and more focused collections has attracted less attention. Our hypothesis is that an intranet search engine can be enhanced by adapting the search system to real users' search behaviour through exploiting its query logs. In this work we describe how a constantly adapting domain model can be used to identify and capture changes in intranet users' search requirements over time. We employ an algorithm that dynamically builds a domain model from query modifications taken from an intranet query log and employs a decay measure, as used in Machine Learning and Optimisation methods, to promote more recent terms. This model is used to suggest query refinements and additions to users and to elevate seasonally relevant terms. A user evaluation using models constructed from a substantial university intranet query log is provided. Statistical evidence demonstrates the system's ability to suggest seasonally relevant terms over three different academic trimesters. We conclude that log files of an intranet search engine are a rich resource to build adaptive domain models, and in our experiments these models significantly outperform sensible baselines.

A fundamental goal of unsupervised feature selection is denoising, which aims to identify and reduce noisy features that are not discriminative. Due to the lack of information about real classes, denoising is a challenging task. The noisy features can disturb the reasonable distance metric and result in unreasonable feature spaces, i.e., the feature spaces in which common clustering algorithms cannot effectively find real classes. To overcome the problem, we make a primary observation that the relevance of features is intrinsic and independent of any metric scaling on the feature space. This observation implies that feature selection should be invariant, at least to some extent, with respect to metric scaling. In this paper, we clarify the necessity of considering the metric invariance in unsupervised feature selection and propose a novel model incorporating metric invariance. Our proposed method is motivated by the following observations: if the statistic that guides the unsupervised feature selection process is invariant with respect to possible metric scaling, the solution of this model will also be invariant. Hence, if a metric-invariant model can distinguish discriminative features from noisy ones in a reasonable feature space, it will also work on the unreasonable counterpart transformed from the reasonable one by metric scaling. A theoretical justification of the metric invariance of our proposed model is given and the empirical evaluation demonstrates its promising performance.

Pseudo relevance feedback (PRF), which has been widely applied in IR, aims to derive a distribution from the top n pseudo relevant documents D. However, these documents are often a mixture of relevant and irrelevant documents. As a result, the derived distribution is actually a mixture model, which has long been limiting the performance of PRF. This is particularly the case when we deal with difficult queries where the truly relevant documents in D are very sparse. In this situation, it is often easier to identify a small number of seed irrelevant documents, which can form a seed irrelevant distribution. Then, a fundamental and challenging problem arises: solely based on the mixed distribution and a seed irrelevance distribution, how to automatically generate an optimal approximation of the true relevance distribution? In this paper, we propose a novel distribution separation model (DSM) to tackle this problem. Theoretical justifications of the proposed algorithm are given. Evaluation results from our extensive simulated experiments on several large scale TREC data sets demonstrate the effectiveness of our method, which outperforms a well respected PRF Model, the Relevance Model (RM), as well as the use of RM on D with the seed negative documents directly removed.

Abstract Expert finding is a key task in enterprise search and has recently attracted lots of attention from both research and industry communities. Given a search topic, a prominent existing approach is to apply some information retrieval (IR) system to retrieve top ranking documents, which will then be used to derive associations between experts and the search topic based on cooccurrences. However, we argue that expert finding is more sensitive to multiple levels of associations and document features that current expert finding systems insufficiently address, including (a) multiple levels of associations between experts and search topics, (b) document internal structure, and (c) document authority. We propose a novel approach that integrates the above‐mentioned three aspects as well as a query expansion technique in a two‐stage model for expert finding. A systematic evaluation is conducted on TREC collections to test the performance of our approach as well as the effects of multiple windows, document features, and query expansion. These experimental results show that query expansion can dramatically improve expert finding performance with statistical significance. For three well‐known IR models with or without query expansion, document internal structures help improve a single window‐based approach but without statistical significance, while our novel multiple window‐based approach can significantly improve the performance of a single window‐based approach both with and without document internal structures.

With the widespread applications of electronic learning (e-Learning) technologies to education at all levels, increasing number of online educational resources and messages are generated from the corresponding e-Learning environments. Nevertheless, it is quite difficult, if not totally impossible, for instructors to read through and analyze the online messages to predict the progress of their students on the fly. The main contribution of this paper is the illustration of a novel concept map generation mechanism which is underpinned by a fuzzy domain ontology extraction algorithm. The proposed mechanism can automatically construct concept maps based on the messages posted to online discussion forums. By browsing the concept maps, instructors can quickly identify the progress of their students and adjust the pedagogical sequence on the fly. Our initial experimental results reveal that the accuracy and the quality of the automatically generated concept maps are promising. Our research work opens the door to the development and application of intelligent software tools to enhance e-Learning.

We propose a novel probabilistic method based on the Hidden Markov Model (HMM) to learn the structure of a Latent Variable Model (LVM) for query language modeling. In the proposed LVM, the combinations of query terms are viewed as the latent variables and the segmented chunks from the feedback documents are used as the observations given these latent variables. Our extensive experiments shows that our method significantly outperforms a number of strong baselines in terms of both effectiveness and robustness.

We argue that expert finding is sensitive to multiple document features in an organization, and therefore, can benefit from the incorporation of these document features. We propose a unified language model, which integrates multiple document features, namely, multiple levels of associations, PageRank, indegree, internal document structure, and URL length. Our experiments on two TREC Enterprise Track collections, i.e., the W3C and CSIRO datasets, demonstrate that the natures of the two organizational intranets and two types of expert finding tasks, i.e., key contact finding for CSIRO and knowledgeable person finding for W3C, influence the effectiveness of different document features. Our work provides insights into which document features work for certain types of expert finding tasks, and helps design expert finding strategies that are effective for different scenarios.

In an adaptive information retrieval (IR) setting, the information seekers' beliefs about which terms are relevant or nonrelevant will naturally fluctuate. This article investigates how the theory of belief revision can be used to model adaptive IR. More specifically, belief revision logic provides a rich representation scheme to formalize retrieval contexts so as to disambiguate vague user queries. In addition, belief revision theory underpins the development of an effective mechanism to revise user profiles in accordance with information seekers' changing information needs. It is argued that information retrieval contexts can be extracted by means of the information-flow text mining method so as to realize a highly autonomous adaptive IR system. The extra bonus of a belief-based IR model is that its retrieval behavior is more predictable and explanatory. Our initial experiments show that the belief-based adaptive IR system is as effective as a classical adaptive IR system. To our best knowledge, this is the first successful implementation and evaluation of a logic-based adaptive IR model which can efficiently process large IR collections.

Dimensionality reduction plays an important role in efficient similarity search, which is often based on k-nearest neighbor (k-NN) queries over a high-dimensional feature space. In this paper, we introduce a novel type of k-NN query, namely conditional k-NN (ck-NN), which considers dimension-specific constraint in addition to the inter-point distances. However, existing dimensionality reduction methods are not applicable to this new type of queries. We propose a novel Mean-Std (standard deviation) guided Dimensionality Reduction (MSDR) to support a pruning based efficient ck-NN query processing strategy. Our preliminary experimental results on 3D protein structure data demonstrate that the MSDR method is promising.

Domain specific information retrieval has become in demand. Not only domain experts, but also average non-expert users are interested in searching domain specific (e.g., medical and health) information from online resources. However, a typical problem to average users is that the search results are always a mixture of documents with different levels of readability. Non-expert users may want to see documents with higher readability on the top of the list. Consequently the search results need to be re-ranked in a descending order of readability. It is often not practical for domain experts to manually label the readability of documents for large databases. Computational models of readability needs to be investigated. However, traditional readability formulas are designed for general purpose text and insufficient to deal with technical materials for domain specific information retrieval. More advanced algorithms such as textual coherence model are computationally expensive for re-ranking a large number of retrieved documents. In this paper, we propose an effective and computationally tractable concept-based model of text readability. In addition to textual genres of a document, our model also takes into account domain specific knowledge, i.e., how the domain-specific concepts contained in the document affect the document's readability. Three major readability formulas are proposed and applied to health and medical information retrieval. Experimental results show that our proposed readability formulas lead to remarkable improvements in terms of correlation with users' readability ratings over four traditional readability measures.

Language Modeling (LM) has been successfully applied to Information Retrieval (IR). However, most of the existing LM approaches only rely on term occurrences in documents, queries and document collections. In traditional unigram based models, terms (or words) are usually considered to be independent. In some recent studies, dependence models have been proposed to incorporate term relationships into LM, so that links can be created between words in the same sentence, and term relationships (e.g. synonymy) can be used to expand the document model. In this study, we further extend this family of dependence models in the following two ways: (1) Term relationships are used to expand query model instead of document model, so that query expansion process can be naturally implemented; (2) We exploit more sophisticated inferential relationships extracted with Information Flow (IF). Information flow relationships are not simply pairwise term relationships as those used in previous studies, but are between a set of terms and another term. They allow for context-dependent query expansion. Our experiments conducted on TREC collections show that we can obtain large and significant improvements with our approach. This study shows that LM is an appropriate framework to implement effective query expansion.

In recent years, statistical language models are being proposed as alternative to the vector space model. Viewing documents as language samples introduces the issue of defining a joint probability distribution over the terms. The present paper models a document as the result of a Markov process. It argues that this process is ergodic, which is theoretically plausible, and easy to verify in practice. The theoretical result is that the joint distribution can be easily obtained. This can also be applied for search resolutions other than the document level. We verified this in an experiment on query expansion demonstrating both the validity and the practicability of the method. This holds a promise for general language models.

Applying Belief Revision logic to model adaptive information retrieval is appealing since it provides a rigorous theoretical foundation to model partiality and uncertainty inherent in any information retrieval (IR) processes. In particular, a retrieval context can be formalised as a belief set and the formalised context is used to disambiguate vague user queries. Belief revision logic also provides a robust computational mechanism to revise an IR system's beliefs about the users' changing information needs. In addition, information flow is proposed as a text mining method to automatically acquire the initial IR contexts. The advantage of a belief-based IRsystem is that its IR behaviour is more predictable and explanatory. However, computational efficiency is often a concern when the belief revision formalisms are applied to large real-life applications. This paper describes our belief-based adaptive IR system which is underpinned by an efficient belief revision mechanism. Our initial experiments show that the belief-based symbolic IR model is more effective than a classical quantitative IR model. To our best knowledge, this is the first successful empirical evaluation of a logic-based IR model based on large IR benchmark collections.

No abstract available.

Abstract Humans can make hasty, but generally robust judgements about what a text fragment is, or is not, about. Such judgements are termed information inference. This article furnishes an account of information inference from a psychologistic stance. By drawing on theories from nonclassical logic and applied cognition, an information inference mechanism is proposed that makes inferences via computations of information flow through an approximation of a conceptual space. Within a conceptual space information is represented geometrically. In this article, geometric representations of words are realized as vectors in a high dimensional semantic space, which is automatically constructed from a text corpus. Two approaches were presented for priming vector representations according to context. The first approach uses a concept combination heuristic to adjust the vector representation of a concept in the light of the representation of another concept. The second approach computes a prototypical concept on the basis of exemplar trace texts and moves it in the dimensional space according to the context. Information inference is evaluated by measuring the effectiveness of query models derived by information flow computations. Results show that information flow contributes significantly to query model effectiveness, particularly with respect to precision. Moreover, retrieval effectiveness compares favorably with two probabilistic query models, and another based on semantic association. More generally, this article can be seen as a contribution towards realizing operational systems that mimic text‐based human reasoning.

The language modelling approach to information retrieval can also be used to compute query models. A query model can be envisaged as an expansion of an initial query. The more prominent query models in the literature have a probabilistic basis. This paper introduces an alternative, non-probabilistic approach to query modelling whereby the strength of information flow is computed between a query Q and a term w. Information flow is a reflection of how strongly w is informationally contained within the query Q. The information flow model is based on Hyperspace Analogue to Language (HAL) vector representations, which reflects the lexical co-occurrence information of terms. Research from cognitive science has demonstrated the cognitive compatibility of HAL representations with human processing. Query models computed from TREC queries by HAL-based information flow are compared experimentally with two probabilistic query language models. Experimental results are provided showing the HAL-based information flow model be superior to query models computed via Markov chains, and seems to be as effective as a probabilistically motivated relevance model.

This paper presents an informational inference mechanism realized via the use of a high dimensional conceptual space. More specifically, we claim to have operationalized important aspects of Gärdenforss recent three-level cognitive model. The connectionist level is primed with the Hyperspace Analogue to Language (HAL) algorithm which produces vector representations for use at the conceptual level. We show how inference at the symbolic level can be implemented by employing Barwise and Seligmans theory of information flow. This article also features heuristics for enhancing HAL-based representations via the use of quality properties, determining concept inclusion and computing concept composition. The worth of these heuristics in underpinning informational inference are demonstrated via a series of experiments. These experiments, though small in scale, show that informational inference proposed in this article has a very different character to the semantic associations produced by the Minkowski distance metric and concept similarity computed via the cosine coefficient. In short, informational inference generally uncovers concepts that are carried, or, in some cases, implied by another concept, (or combination of concepts).

Experimental approaches are widely employed to benchmark the performance of an information retrieval (IR) system. Measurements in terms of recall and precision are computed as performance indicators. Although they are good at assessing the retrieval effectiveness of an IR system, they fail to explore deeper aspects such as its underlying functionality and explain why the system shows such performance. Recently, inductive (i.e., theoretical) evaluation of IR systems has been proposed to circumvent the controversies of the experimental methods. Several studies have adopted the inductive approach, but they mostly focus on theoretical modeling of IR properties by using some metalogic. In this article, we propose to use inductive evaluation for functional benchmarking of IR models as a complement of the traditional experiment-based performance benchmarking. We define a functional benchmark suite in two stages: the evaluation criteria based on the notion of "aboutness," and the formal evaluation methodology using the criteria. The proposed benchmark has been successfully applied to evaluate various well-known classical and logic-based IR models. The functional benchmarking results allow us to compare and analyze the functionality of the different IR models.

Information retrieval (IR) is driven by a process which decides whether a document is about a query. Recent attempts \nspawned from logic-based information retrieval theory have formalized properties characterizing “aboutness”, \nbut no consensus has yet been reached. The proposed properties are largely determined by the underlying \nframework within which aboutness is defined. In addition, some properties are only sound within the context of a \ngiven IR model, but are not sound from the perspective of the user. For example, a common form of aboutness, \nnamely overlapping aboutness, implies precision degrading properties such as compositional monotonicity. Therefore, \nthe motivating question for this paper is: Independent of any given IR model, and examined within an information- \nbased, abstract framework, what are commonsense properties of aboutness (and its dual, non-aboutness)? \nWe propose a set of properties characterizing aboutness and non-aboutness from a commonsense perspective. \nSpecial attention is paid to the rules prescribing conservative behaviour of aboutness with respect to information \ncomposition. The interaction between aboutness and non-aboutness is modeled via normative rules. The completeness, \nsoundness and consistency of the aboutness proof systems are analyzed and discussed. A case study based on \nmonotonicity shows that many current IR systems are either monotonic or non-monotonic. An interesting class of \nIR models, namely those that are conservatively monotonic, is identified.

No abstract available.