Information Retrieval Tasks Research Articles

Scholarly Communication: A Discipline that should be promoted

Scholarly Communication is the name given to a wide field of activities related to the different ways in which the authors of academic works publish and disseminate their results. It includes tasks whose connections to this publication are not only direct, but mutually influenced. These are the tasks of dissemination, treatment, analysis, and information retrieval, such as those carried out by databases and other agents. But the issue goes even further and other activities must be added, equally interconnected, such as preservation and increased visibility, along with specialized actions carried out by university libraries. It is concluded that academic communication, despite its breadth and diversity (or precisely because of it), should be studied by a unified discipline that, thanks to a holistic vision, helps to overcome its main problems. This paper presents some of these problems and defends the need for both the field of activities itself and the discipline that studies them to receive a clearer and more determined impulse.

Comunicación académica: una disciplina que nos conviene impulsar

Comunicación académica es el nombre que recibe un amplio campo de actividades vinculado con las muy diversas formas en las que los autores de trabajos académicos publican y difunden sus resultados. Incluye tareas cuyas conexiones con esta publicación no solo son directas, sino que se influyen mutuamente. Se trata de las labores de difusión, tratamiento, análisis y recuperación de información, como las que llevan a cabo las bases de datos y otros agentes. Pero la cuestión aún va más lejos y a estas debemos añadir otras actividades, igualmente interconectadas, como la preservación y el aumento de la visibilidad, junto a actuaciones especializadas que realizan las bibliotecas universitarias. Se concluye que la comunicación académica pese a su amplitud y diversidad (o precisamente por ello) debería ser estudiada por una disciplina unificada que, gracias a una visión holística, ayude a superar sus principales problemas. En este trabajo se presentan algunos de estos problemas y se defiende la necesidad de que tanto el ámbito de actividades en sí como la disciplina que las estudia reciban un impulso más claro y decidido.

Extractive Arabic Text Summarization-Graph-Based Approach

With the noteworthy expansion of textual data sources in recent years, easy, quick, and precise text processing has become a challenge for key qualifiers. Automatic text summarization is the process of squeezing text documents into shorter summaries to facilitate verification of their basic contents, which must be completed without losing vital information and features. The most difficult information retrieval task is text summarization, particularly for Arabic. In this research, we offer an automatic, general, and extractive Arabic single document summarizing approach with the goal of delivering a sufficiently informative summary. The proposed model is based on a textual graph to generate a coherent summary. Firstly, the original text is converted to a textual graph using a novel formulation that takes into account sentence relevance, coverage, and diversity to evaluate each sentence using a mix of statistical and semantic criteria. Next, a sub-graph is built to reduce the size of the original text. Finally, unwanted and less weighted phrases are removed from the summarized sentences to generate a final summary. We used Recall-Oriented Research to Evaluate Main Idea (RED) as an evaluative metric to review our proposed technique and compare it with the most advanced methods. Finally, a trial on the Essex Arabic Summary Corpus (EASC) using the ROUGE index showed promising results compared with the currently available methods.

DATM: A Novel Data Agnostic Topic Modeling Technique With Improved Effectiveness for Both Short and Long Text

Topic modelling is important for tackling several data mining tasks in information retrieval. While seminal topic modelling techniques such as Latent Dirichlet Allocation (LDA) have been proposed, the ubiquity of social media and the brevity of its texts pose unique challenges for such traditional topic modelling techniques. Several extensions including <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">auxiliary aggregation</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">self aggregation</i> and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">direct learning</i> have been proposed to mitigate these challenges, however some still remain. These include a lack of consistency in the topics generated and the decline in model performance in applications involving disparate document lengths. There is a recent paradigm shift towards neural topic models, which are not suited for resource-constrained environments. This paper revisits LDA-style techniques, taking a theoretical approach to analyse the relationship between word co-occurrence and topic models. Our analysis shows that by altering the word co-occurrences within the corpus, topic discovery can be enhanced. Thus we propose a novel data transformation approach dubbed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DATM</i> to improve the topic discovery within a corpus. A rigorous empirical evaluation shows that DATM is not only powerful, but it can also be used in conjunction with existing benchmark techniques to significantly improve their effectiveness and their consistency by up to 2 fold.

A comparison of two text specificity measures analyzing a heterogenous text corpus

The article compares the performance of two term specificity measures, Cohen’s d and Z-score, when analyzing political and media discourses on Russia’s war in Ukraine in four languages and five countries. In addition to linguistic and stylistic heterogeneity, 3,347 texts included in the corpus have variable length. The two measures display convergent validity, as confirmed by various performance metrics. It is argued that the measures can be adapted to a broader range of tasks in information retrieval and digital humanities, in addition to their usefulness for text mining and content analysis.

Information Retrieval With Chessboard-Shaped Topology for Hyperspectral Target Detection

Given a priori knowledge, hyperspectral target detection aims to locate objects of interest within specific scenes by utilizing differences in spectral characteristics among various land covers. However, for those traditional model-driven detectors with monotonic analytical mode, they perform mediocrely in the disassembly of hyperspectral image (HSI) data, failing to cope with real scenes with complexity. The discrepancy between fixed model assumptions and HSI data severely reduces detection effects, leading to the inability of such methods to mine deep-level features and adapt to the variability of imaging scenes. To overcome the limitations of traditional methods, we propose a chessboard-shaped topological framework for high-dimensional data structures to disassemble an HSI from both spatial and spectral dimensions adaptively. With hyperspectral target detection is refined into an information retrieval task in a topological space, a target detection method based on chessboard-shaped topology (CTTD) is proposed. In the topological space, latent and hidden data features of original images are presented in an intuitive way. Therefore, the differences in both spatial and spectral dimensions between the two classes of objects, namely target and background, are specifically amplified and exploited to perform the information retrieval task with superior performance. Extensive experimental results on benchmark HSI data sets demonstrate that CTTD can efficiently adapt to the variability of real scenes while extracting abundant and detailed information for accurate target localization. Moreover, both detection effects and computational efficiency exhibited by the proposed method provide a strong support for its popularization in practical applications.

Multi-Modal Cross-Domain Alignment Network for Video Moment Retrieval

As an increasingly popular task in multimedia information retrieval, video moment retrieval (VMR) aims to localize the target moment from an untrimmed video according to a given language query. Most previous methods depend heavily on numerous manual annotations ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , moment boundaries), which are extremely expensive to acquire in practice. In addition, due to the domain gap between different datasets, directly applying these pre-trained models to an unseen domain leads to a significant performance drop. In this paper, we focus on a novel task: cross-domain VMR, where fully-annotated datasets are available in one domain (“source domain”), but the domain of interest (“target domain”) only contains unannotated datasets. As far as we know, we present the first study on cross-domain VMR. To address this new task, we propose a novel Multi-Modal Cross-Domain Alignment (MMCDA) network to transfer the annotation knowledge from the source domain to the target domain. However, due to the domain discrepancy between the source and target domains and the semantic gap between videos and queries, directly applying trained models to the target domain generally leads to a performance drop. To solve this problem, we develop three novel modules: (i) a domain alignment module is designed to align the feature distributions between different domains of each modality; (ii) a cross-modal alignment module aims to map both video and query features into a joint embedding space and to align the feature distributions between different modalities in the target domain; (iii) a specific alignment module tries to obtain the fine-grained similarity between a specific frame and the given query for optimal localization. By jointly training these three modules, our MMCDA can learn domain-invariant and semantic-aligned cross-modal representations. Extensive experiments on three challenging benchmarks (ActivityNet Captions, Charades-STA and TACoS) illustrate that our cross-domain method MMCDA outperforms all state-of-the-art single-domain methods. Impressively, MMCDA raises the performance by more than 7% in representative cases, which demonstrates its effectiveness.

The Process and Algorithm Analysis of Text Mining System Based on Artificial Intelligence

The rapid development of the Internet leads to the rapid growth of network information, we call it information explosion. The Internet is full of information, and it is difficult for users to find this information and useful knowledge of the ocean. The Web has become the world's largest information repository, and there is an urgent need for efficient access to the valuable knowledge of vast amounts of web information. The purpose of this paper is to study the process and algorithm analysis of text mining system based on artificial intelligence. This paper presents an algorithm of document feature acquisition based on genetic algorithm. Selecting suitable features is an important task in specific text classification and information retrieval. Finding appropriate feature vectors to represent the text will undoubtedly help with subsequent sorting and grouping. Based on the genetic algorithm of variable length chromosome, this paper improves the crossover, mutation and selection operations, and proposes an algorithm to obtain text feature vectors. This method has a wide range of applications and good results.

A Comprehensive Survey of Music Genre Classification Using Audio Files

This survey extensively studies music genre classification, a critical task in music information retrieval, to automatically categorize audio recordings into various genres. It provides a comprehensive review of approaches, methodologies, and recent advancements in genre classification from audio data. Scholars and practitioners in the field will find this study to be a valuable resource as it covers various aspects of the discipline, including feature extraction, classification methods, dataset exploration, evaluation metrics, and recent developments. The survey aims to enhance the understanding of music genre classification and foster further research and progress in the field by critically evaluating state-of-the-art techniques discussed in research papers, discussing their strengths and limitations, and providing a comprehensive overview of the field.

Improving visual-semantic embeddings by learning semantically-enhanced hard negatives for cross-modal information retrieval

Visual Semantic Embedding (VSE) networks aim to extract the semantics of images and their descriptions and embed them into the same latent space for cross-modal information retrieval. Most existing VSE networks are trained by adopting a hard negatives loss function which learns an objective margin between the similarity of relevant and irrelevant image–description embedding pairs. However, the objective margin in the hard negatives loss function is set as a fixed hyperparameter that ignores the semantic differences of the irrelevant image–description pairs. To address the challenge of measuring the optimal similarities between image–description pairs before obtaining the trained VSE networks, this paper presents a novel approach that comprises two main parts: (1) finds the underlying semantics of image descriptions; and (2) proposes a novel semantically-enhanced hard negatives loss function, where the learning objective is dynamically determined based on the optimal similarity scores between irrelevant image–description pairs. Extensive experiments were carried out by integrating the proposed methods into five state-of-the-art VSE networks that were applied to three benchmark datasets for cross-modal information retrieval tasks. The results revealed that the proposed methods achieved the best performance and can also be adopted by existing and future VSE networks.

A neural harmonic-aware network with gated attentive fusion for singing melody extraction

Singing melody extraction from polyphonic musical audio is one of the most challenging tasks in music information retrieval (MIR). Recently, data-driven methods based on convolutional neural networks (CNNs) have achieved great success for this task. In the literature, harmonic relationship has been proven crucial for this task. However, few existing CNN-based singing melody extraction methods consider the harmonic relationship in the training stage. The state-of-the-art CNN based methods are not capable of capturing such long-dependency harmonic relationship due to limited receptive field and unacceptable computation cost. In this paper, we introduce a neural harmonic-aware network with gated attentive fusion (NHAN-GAF) for singing melody extraction. Specifically, in the 2-D spectrograms modeling branch, we propose to employ multiple parallel 1-D CNN kernels to capture the harmonic relations between 1–2 octaves along the frequency axis in the spectrogram. Considering the advantage of jointly using Time–Frequency (T-F) domain and time domain information, we use two-branch neural nets to learn discriminative representation for this task. A novel gated attentive fusion (GAF) network is suggested to encode potential correlations between the two branches and fuse the descriptors learned from raw waveform and T-F spectrograms. Moreover, the idea of GAF can be exploited to the multimedia applications with multimodal analysis. With the two proposed components, our proposed model is capable of learning the harmonic relationship in the spectrogram and better capturing the contextual but discriminative features for singing melody extraction. We use part of the vocal tracks of the RWC dataset and MIR-1 K dataset to train the model and evaluate the performance of the proposed model on the ADC 2004, MIREX 05 and MedleyDB datasets. The experimental results show that the proposed method outperforms the state-of-the-art ones.

Deep Learning Techniques for Pattern Recognition in EEG Audio Signal-Processing-Based Eye-Closed and Eye-Open Cases

Recently, pattern recognition in audio signal processing using electroencephalography (EEG) has attracted significant attention. Changes in eye cases (open or closed) are reflected in distinct patterns in EEG data, gathered across a range of cases and actions. Therefore, the accuracy of extracting other information from these signals depends significantly on the prediction of the eye case during the acquisition of EEG signals. In this paper, we use deep learning vector quantization (DLVQ), and feedforward artificial neural network (F-FANN) techniques to recognize the case of the eye. The DLVQ is superior to traditional VQ in classification issues due to its ability to learn a code-constrained codebook. On initialization by the k-means VQ approach, the DLVQ shows very promising performance when tested on an EEG-audio information retrieval task, while F-FANN classifies EEG-audio signals of eye state as open or closed. The DLVQ model achieves higher classification accuracy, higher F score, precision, and recall, as well as superior classification abilities as compared to the F-FANN.

Dealing with textual noise for robust and effective BERT re-ranking

The pre-trained language models (PLMs), such as BERT, have been successfully employed in two-phases ranking pipeline for information retrieval (IR). Meanwhile, recent studies have reported that BERT model is vulnerable to imperceptible textual perturbations on quite a few natural language processing (NLP) tasks. As for IR tasks, current established BERT re-ranker is mainly trained on large-scale and relatively clean dataset, such as MS MARCO, but actually noisy text is more common in real-world scenarios, such as web search. In addition, the impact of within-document textual noises (perturbations) on retrieval effectiveness remains to be investigated, especially on the ranking quality of BERT re-ranker, considering its contextualized nature. To mitigate this gap, we carry out exploratory experiments on the MS MARCO dataset in this work to examine whether BERT re-ranker can still perform well when ranking text with noise. Unfortunately, we observe non-negligible effectiveness degradation of BERT re-ranker over a total of ten different types of synthetic within-document textual noise. Furthermore, to address the effectiveness losses over textual noise, we propose a novel noise-tolerant model, De-Ranker, which is learned by minimizing the distance between the noisy text and its original clean version. Our evaluation on the MS MARCO and TREC 2019–2020 DL datasets demonstrates that De-Ranker can deal with synthetic textual noise more effectively, with 3%–4% performance improvement over vanilla BERT re-ranker. Meanwhile, extensive zero-shot transfer experiments on a total of 18 widely-used IR datasets show that De-Ranker can not only tackle natural noise in real-world text, but also achieve 1.32% improvement on average in terms of cross-domain generalization ability on the BEIR benchmark.

Conversational Agents for Information Retrieval in the Education Domain

Text-based conversational agents (CAs) are widely deployed across a number of daily tasks, including information retrieval. However, most existing agents follow a default design that disregards user needs and preferences, ultimately leading to a lack of usage and an unsatisfying user experience. To better understand how CAs can be designed in order to lead to effective system use, we deduced relevant design requirements from both literature and 13 user interviews. We built and tested a question-answering, text-based CA for an information retrieval task in an education scenario. Results from our experimental test with 41 students indicate that following a user-centered design has a significant positive effect on enjoyment and trust in a CA as opposed to deploying a default CA. If not designed with the user in mind, CAs are not necessarily more beneficial than traditional question-answering systems. Beyond practical implications for effective CA design, this paper points towards key challenges and potential research avenues when deploying social cues for CAs.

Evaluating the Impact on Clinical Task Efficiency of a Natural Language Processing Algorithm for Searching Medical Documents: Prospective Crossover Study.

Information retrieval (IR) from the free text within electronic health records (EHRs) is time consuming and complex. We hypothesize that natural language processing (NLP)-enhanced search functionality for EHRs can make clinical workflows more efficient and reduce cognitive load for clinicians. This study aimed to evaluate the efficacy of 3 levels of search functionality (no search, string search, and NLP-enhanced search) in supporting IR for clinical users from the free text of EHR documents in a simulated clinical environment. A clinical environment was simulated by uploading 3 sets of patient notes into an EHR research software application and presenting these alongside 3 corresponding IR tasks. Tasks contained a mixture of multiple-choice and free-text questions. A prospective crossover study design was used, for which 3 groups of evaluators were recruited, which comprised doctors (n=19) and medical students (n=16). Evaluators performed the 3 tasks using each of the search functionalities in an order in accordance with their randomly assigned group. The speed and accuracy of task completion were measured and analyzed, and user perceptions of NLP-enhanced search were reviewed in a feedback survey. NLP-enhanced search facilitated more accurate task completion than both string search (5.14%; P=.02) and no search (5.13%; P=.08). NLP-enhanced search and string search facilitated similar task speeds, both showing an increase in speed compared to the no search function, by 11.5% (P=.008) and 16.0% (P=.007) respectively. Overall, 93% of evaluators agreed that NLP-enhanced search would make clinical workflows more efficient than string search, with qualitative feedback reporting that NLP-enhanced search reduced cognitive load. To the best of our knowledge, this study is the largest evaluation to date of different search functionalities for supporting target clinical users in realistic clinical workflows, with a 3-way prospective crossover study design. NLP-enhanced search improved both accuracy and speed of clinical EHR IR tasks compared to browsing clinical notes without search. NLP-enhanced search improved accuracy and reduced the number of searches required for clinical EHR IR tasks compared to direct search term matching.

Automatic signboard detection and localization in densely populated developing cities

Most city establishments of developing cities are digitally unlabeled because of the lack of automatic annotation systems. Hence location and trajectory services such as Google Maps, Uber etc remain underutilized in such cities. Accurate signboard detection in natural scene images is the foremost task for error-free information retrieval from such city streets. Yet, developing accurate signboard localization system is still an unresolved challenge because of its diverse appearances that include textual images and perplexing backgrounds. We present a novel object detection approach that can detect signboards automatically and is suitable for such cities. We use Faster R-CNN based localization by incorporating two specialized pretraining methods and a run time efficient hyperparameter value selection algorithm. We have taken an incremental approach in reaching our final proposed method through detailed evaluation and comparison with baselines using our constructed SVSO (Street View Signboard Objects) signboard dataset containing signboard natural scene images of six developing countries. We demonstrate state-of-the-art performance of our proposed method on both SVSO dataset and Open Image Dataset. Our proposed method can detect signboards accurately (even if the images contain multiple signboards with diverse shapes and colors in a noisy background) achieving 0.90 mAP (mean average precision) score on SVSO independent test set. Our implementation is available at: https://cutt.ly/dhr1D2K.

Construction of Fuzzy Linguistic Approximate Concept Lattice in an Incomplete Fuzzy Linguistic Formal Context

Uncertainty research is one of the critical problems in artificial intelligence. In an uncertain environment, a large quantity of information is expressed in linguistic values. Aiming at the missing linguistic-valued information, we first propose incomplete fuzzy linguistic formal context and then discuss the fuzzy linguistic approximate concept. Our proposal can describe the attributes of objects from two aspects simultaneously. One is an object's essential attributes, and another includes the essential and possible attributes. As a result, more object-related information can be obtained to reduce information loss effectively. We design a similarity metric for correcting the errors caused by the initial complement operation. We then construct a corresponding fuzzy linguistic approximate concept lattice for the task of approximate information retrieval. Finally, we illustrate the applicability and feasibility of the proposed approach with concrete examples, which clearly show that our approach can better deal with the linguistic-valued information in an uncertain environment.

Shallow pooling for sparse labels

Recent years have seen enormous gains in core information retrieval tasks, including document and passage ranking. Datasets and leaderboards, and in particular the MS MARCO datasets, illustrate the dramatic improvements achieved by modern neural rankers. When compared with traditional information retrieval test collections, such as those developed by TREC, the MS MARCO datasets employ substantially more queries—thousands vs. dozens – with substantially fewer known relevant items per query—often just one. For example, 94% of the nearly seven thousand queries in the MS MARCO passage ranking development set have only a single known relevant passage, and no query has more than four. Given the sparsity of these relevance labels, the MS MARCO leaderboards track improvements with mean reciprocal rank (MRR). In essence, the known relevant item is treated as the “right answer” or “best answer”, with rankers scored on their ability to place this item as high in the ranking as possible. In working with these sparse labels, we have observed that the top items returned by a ranker often appear superior to judged relevant items. Others have reported the same observation. To test this observation, we employed crowdsourced workers to make preference judgments between the top item returned by a modern neural ranking stack and a judged relevant item for the nearly seven thousand queries in the passage ranking development set. The results support our observation. If we imagine a hypothetical perfect ranker under MRR, with a score of 1 on all queries, our preference judgments indicate that a searcher would prefer the top result from a modern neural ranking stack more frequently than the top result from the hypothetical perfect ranker, making our neural ranker “better than perfect”. To understand the implications for the leaderboard, we pooled the top document from available runs near the top of the passage ranking leaderboard for over 500 queries. We employed crowdsourced workers to make preference judgments over these pools and re-evaluated the runs. Our results support our concerns that current MS MARCO datasets may no longer be able to recognize genuine improvements in rankers. In future, if rankers are measured against a single answer, this answer should be the best answer or most preferred answer, and maintained with ongoing judgments. Since only the best known answer is required, this ongoing maintenance might be performed with shallow pooling. When a previously unjudged document is surfaced as the top item in a ranking, it can be directly compared with the previous best known answer.

GCKG: Novel Gated Convolutional embedding model for Knowledge Graphs

With the aim of constructing a low-dimensional representation space, knowledge graph embedding has gradually become a hot spot in various information retrieval and machine learning tasks. However, most existing knowledge graphs suffer from incompleteness due to the fact that it is unable to collect a complete world knowledge. For knowledge graph completion (KGC), this paper introduces a novel embedding method, named gated Convolution for Knowledge Graph (GCKG), to automatically predict missing links for large-scale Knowledge Graphs. The model GCKG advances the existing models by employing a gated mechanism on convolutional neural network, so that it can capture global relationships and attribute characteristics between entities and relations in knowledge graphs. First, three embedding vectors of each triple are combined into three different matrices. Then, the matrices are fed to three convolutional layers. Finally, the network generates three different maps which are controlled to transmit the information by a gated mechanism. By performing extensive and comprehensive experiments, we evaluate GCKG with two benchmark datasets, FB15k-237 and WN18RR. Experiments show that GCKG achieves substantial improvements against state-of-the-art baselines, and it is respectively improved about 37% in MRR and 4% in Hits@10 on FB15k-237 database compared with CapsE.

Relevance Feedback and Deep Neural Network-Based Semantic Method for Query Expansion

Machine learning techniques have been widely used in almost every area of arts, science, and technology for the last two decades. Document analysis and query expansion also use machine learning techniques at a broad scale for information retrieval tasks. The state-of-the-art models like the Bo1 model, Bo2 model, KL divergence model, and chi-square model are probabilistic, and they work on DFR-based retrieval models. These models are much focused on term frequency and do not care about the semantic relationship among the terms. The proposed model applies the semantic method to find the semantic similarity among the terms to expand the query. The proposed method uses the relevance feedback method that selects a user-assisted most relevant document from top “ k ” initially retrieved documents and then applies deep neural network technique to select the most informative terms related to original query terms. The results are evaluated at FIRE 2011 ad hoc English test collection. The mean average precision of the proposed method is 0.3568. The proposed method also compares the state-of-the-art models. The proposed model observed 19.77% and 8.05% improvement on the mean average precision (MAP) parameter with respect to the original query and Bo1 model, respectively.