Query Recall Research Articles

Underwater object detection method based on learnable query recall mechanism and lightweight adapter.

With the rapid development of ocean observation technology, underwater object detection has begun to occupy an essential position in the fields of aquaculture, environmental monitoring, marine science, etc. However, due to the problems unique to underwater images such as severe noise, blurred objects, and multi-scale, deep learning-based target detection algorithms lack sufficient capabilities to cope with these challenges. To address these issues, we improve DETR to make it well suited for underwater scenarios. First, a simple and effective learnable query recall mechanism is proposed to mitigate the effect of noise and can significantly improve the detection performance of the object. Second, for underwater small and irregular object detection, a lightweight adapter is designed to provide multi-scale features for the encoding and decoding stages. Third, the regression mechanism of the bounding box is optimized using the combination loss of smooth L1 and CIoU. Finally, we validate the designed network against other state-of-the-art methods on the RUOD dataset. The experimental results show that the proposed method is effective.

DistPreserv: Maintaining User Distribution for Privacy-Preserving Location-Based Services

Location-Based Services (LBSs) are one of the most frequently used mobile applications in the modern society. Geo-Indistinguishability (Geo-Ind) is a promising privacy protection model for LBSs since it can provide formal security guarantees for location privacy. However, Geo-Ind undermines the statistical location distribution of users on the LBS server because of perturbed locations, thereby disabling the server to provide distribution-based services (e.g., traffic congestion maps). To overcome this issue, we give a privacy definition, called DistPreserv, to enable the LBS server to acquire valid location distributions while providing users with strict location protection. Then we propose a privacy-preserving LBS scheme to benefit both users and the server, in which a location perturbation mechanism is designed to achieve the given definition under the guide of the incentive compatibility, and a retrieval area determination method is presented to ensure query accuracy of users by using the dynamic programming on the two-dimensional map plane. Finally, we theoretically prove that the designed mechanism can achieve the definition of DistPreserv and the property of incentive compatibility. Experimental explorations using a real-world dataset indicate that our proposal prominently improves the availability of users location distributions by over 90%, while providing high precision and recall of queries.

Hollow-tree: a metric access method for data with missing values

Similarity search is fundamental to store and retrieve large volumes of complex data required by many real world applications. A useful mechanism for such concept is the query-by-similarity. Based on their topological properties, metric similarity functions can be used to index sets of data which can be queried effectively and efficiently by the so-called metric access methods. However, data produced by various application domains and the varying data types handled often lead to missing data, hence, they do not follow the metric similarity requirements. As a consequence, missing data cause distortions in the index structure and yield bias in the query answer. In this paper, we propose the Hollow-tree, a novel access method aimed at successfully retrieving data with missing attribute values. It employs new strategies for indexing and searching data elements, capable of handling the missing data issues when the cause of missingness is ignorable. The indexing strategy is based on a family of distance functions that allow measuring the distance between elements with missing values, along with a set of policies able to organize the elements in the index without causing distortions to its internal structure. The searching strategy employs fractal dimension property of the data to achieve accurate query answer while considering data with missing values part of the response. Results from experiments performed on a variety of real and synthetic data sets showed that, while other metric access methods deteriorate with small amounts of missing values, the Hollow-tree maintains a remarkable performance with almost 100% of precision and recall for range queries and more than 90% for k-nearest neighbor queries, for up to 40% of missing values.

Effective Searching of RDF Knowledge Graphs

RDF knowledge graphs are typically searched using triple-pattern queries. Often, triple-pattern queries will return too many or too few results, making it difficult for users to find relevant answers to their information needs. To remedy this, we propose a general framework for effective searching of RDF knowledge graphs. Our framework extends both the searched knowledge graph and triple-pattern queries with keywords to allow users to form a wider range of queries. In addition, it provides result ranking based on statistical machine translation, and performs automatic query relaxation to improve query recall. Finally, we also define a notion of result diversity in the setting of RDF data and provide mechanisms to diversify RDF search results using Maximal Marginal Relevance. We evaluate the effectiveness of our retrieval framework using various carefully-designed user studies on DBpedia, a large and real-world RDF knowledge graph.

Effective searching of RDF knowledge graphs

RDF knowledge graphs are typically searched using triple-pattern queries. Often, triple-pattern queries will return too many or too few results, making it difficult for users to find relevant answers to their information needs. To remedy this, we propose a general framework for effective searching of RDF knowledge graphs. Our framework extends both the searched knowledge graph and triple-pattern queries with keywords to allow users to form a wider range of queries. In addition, it provides result ranking based on statistical machine translation, and performs automatic query relaxation to improve query recall. Finally, we also define a notion of result diversity in the setting of RDF data and provide mechanisms to diversify RDF search results using Maximal Marginal Relevance. We evaluate the effectiveness of our retrieval framework using various carefully-designed user studies on DBpedia, a large and real-world RDF knowledge graph.

Building and Querying RDF/OWL Database of Semantically Annotated Nuclear Medicine Images.

As the use of positron emission tomography-computed tomography (PET-CT) has increased rapidly, there is a need to retrieve relevant medical images that can assist image interpretation. However, the images themselves lack the explicit information needed for query. We constructed a semantically structured database of nuclear medicine images using the Annotation and Image Markup (AIM) format and evaluated the ability the AIM annotations to improve image search. We created AIM annotation templates specific to the nuclear medicine domain and used them to annotate 100 nuclear medicine PET-CT studies in AIM format using controlled vocabulary. We evaluated image retrieval from 20 specific clinical queries. As the gold standard, two nuclear medicine physicians manually retrieved the relevant images from the image database using free text search of radiology reports for the same queries. We compared query results with the manually retrieved results obtained by the physicians. The query performance indicated a 98% recall for simple queries and a 89% recall for complex queries. In total, the queries provided 95% (75 of 79 images) recall, 100% precision, and an F1 score of 0.97 for the 20 clinical queries. Three of the four images missed by the queries required reasoning for successful retrieval. Nuclear medicine images augmented using semantic annotations in AIM enabled high recall and precision for simple queries, helping physicians to retrieve the relevant images. Further study using a larger data set and the implementation of an inference engine may improve query results for more complex queries.

Information Behavior within the Humanities: Searching or Browsing, Recall or Precision? Researching the Information Needs of Academics: the Case Study of the Faculty of History of the University of Warsaw


 PURPOSE/THESIS: The aim of this research was to discover the characteristics of information behavior among the scholars whose academic focus lies within the field of humanities. Recognizing the importance of local conditions and their impact on information behavior, the researcher conducted her empirical study on a group of Polish academics, i.e. the employees of the Faculty of History at the University of Warsaw. Particular emphasis was put on establishing (1) how academics retrieve required literature, (2) what factors may influence the degree of satisfaction from the query, (3) how crucial was the level of results’ relevance and (4) whether academics assign greater importance to query recall or precision. APPROACH/METHODS: Empirical research was preceded by the analysis of subject literature enabling the author of the paper to outline several internationally observed tendencies in information behavior of academics, and to formulate theses with regard to Polish academics. These were verified on the basis of 54 in-depth interviews with the employees of the Faculty of History at the University of Warsaw and their ordered queries of academic literature. RESULTS AND CONCLUSIONS: Research results have indicated that humanities scholars prefer browsing document sets characterized by a high level of recall as a method of retrieving required subject literature. Due to the difficulty in formulating query instructions with sufficient precision, they less frequently choose the method of searching for document sets characterized by high precision factor. The need to browse through irrelevant titles was found to have no significant impact on the level of user satisfaction. Additional research has also indicated that academic texts databases are still insufficiently adjusted to the needs of humanities scholars. ORIGINALITY/VALUE: The research has allowed to describe information behavior of humanities scholars with regard to the use of digital technologies in the research process, which so far has not taken place in the case of Polish academic community.
 

소아 및 종양 핵의학 영상판독에서 RDF/OWL 데이터의 유용성

Recently, the structured data format in RDF/OWL has played an increasingly vital role in the semantic web. We converted pediatric and oncologic nuclear medicine imaging reports in free text into RDF/OWL format and evaluated the usefulness of nuclear medicine imaging reports in RDF/OWL by comparing SPARQL query results with the manually retrieved results by physicians from the reports in free text. SPARQL query showed 95% recall for simple queries and 91% recall for dedicated queries. In total, SPARQL query retrieved 93% (51 lesions of 55) recall and 100% precision for 20 clinical query items. All query results missed by SPARQL query were of some inference. Nuclear medicine imaging reports in the format of RDF/OWL were very useful for retrieving simple and dedicated query results using SPARQL query. Further study using more number of cases and knowledge for inference is warranted.

Effective Retrival of Data from E-mail Corpus for Digital Investigations

In digital investigations the textual evidence is very important. It has a vast majority during investigations. All this is due to a very great deal of stored digital data which is linguistic in nature (e.g. human languages, programming languages). Some of the most important text-based evidences are SMS, chat logs, emails, word processing documents, spreadsheets, address books, calendar appointments and system logs. Here, the one who investigates is flooded with data and he has to spend valuable investigation time, scanning through the noisy search results and going through irrelevant search results. The current digital forensic text string search tools use matching or indexing algorithms for searching digital evidences at a physical level to locate specific text strings. All these are designed in such a way that it achieves 100% query recall (i.e. it finds all instances relating to text strings). Here the nature of the data set is given, which leads to a highly extreme incidence of hits that are totally irrelevant to investigative objectives. There is a text string search tool that fails to group or order search hits in a manner that considerably improves the investigator's ability to get to the relevant hits first or at least in a quick manner. Thus, the text mining has been taken up as a new initiative for digital forensics. This type of text mining approach will enhance the IIR (Intelligent Information Retrieval) effectiveness of digital forensic text string searching. Henceforth, the technology for text mining can be scalable up to large datasets in Gigabytes or Terabytes. Here the software that has been developed consists of text analysis, data extraction, and Meta data correlation and visualization features and Information retrieval. The main aim of the system that was developed is to analyze the various transcripts like SMS, Email and word letters, event logs and chat transcripts. There are specific keywords that are searched by the system, which are weighted by the user in compliance with domain specific analysis. The correlated data are ranked by the system and hence is displayed by the system. For the investigators to analyze further this helps it by providing user graphs and charts about the ranked data.

SOS: A Distributed Mobile Q&A SystemBased on Social Networks

Recently, emerging research efforts have been focused on question and answer (Q&A) systems based on social networks. The social-based Q&A systems can answer non-factual questions, which cannot be easily resolved by web search engines. These systems either rely on a centralized server for identifying friends based on social information or broadcast a user's questions to all of its friends. Mobile Q&A systems, where mobile nodes access the Q&A systems through Internet, are very promising considering the rapid increase of mobile users and the convenience of practical use. However, such systems cannot directly use the previous centralized methods or broadcasting methods, which generate high cost of mobile Internet access, node overload, and high server bandwidth cost with the tremendous number of mobile users. We propose a distributed Social-based mObile Q&A System (SOS) with low overhead and system cost as well as quick response to question askers. SOS enables mobile users to forward questions to potential answerers in their friend lists in a decentralized manner for a number of hops before resorting to the server. It leverages lightweight knowledge engineering techniques to accurately identify friends who are able to and willing to answer questions, thus reducing the search and computation costs of mobile nodes. The trace-driven simulation results show that SOS can achieve a high query precision and recall rate, a short response latency and low overhead. We have also deployed a pilot version of SOS for use in a small group in Clemson University. The feedback from the users shows that SOS can provide high-quality answers.

Spatially aware feature selection and weighting for object retrieval

Many recent image retrieval methods are based on the “bag-of-words” (BoW) model with some additional spatial consistency checking. This paper proposes a more accurate similarity measurement that takes into account spatial layout of visual words in an offline manner. The similarity measurement is embedded in the standard pipeline of the BoW model, and improves two features of the model: i) latent visual words are added to a query based on spatial co-occurrence, to improve query recall; and ii) weights of reliable visual words are increased to improve the precision. The combination of these methods leads to a more accurate measurement of image similarity. This is similar in concept to the combination of query expansion and spatial verification, but does not require query time processing, which is too expensive to apply to full list of ranked results. Experimental results demonstrate the effectiveness of our proposed method on three public datasets.

Intelligent search in social communities of smartphone users

Social communities of smartphone users have recently gained significant interest due to their wide social penetration. The applications in this domain, however, currently rely on centralized or cloud-like architectures for data sharing and searching tasks, introducing both data-disclosure and performance concerns. In this paper, we present a distributed search architecture for intelligent search of objects in a mobile social community. Our framework, coined SmartOpt, is founded on an in-situ data storage model, where captured objects remain local on smartphones and searches then take place over an intelligent multi-objective lookup structure we compute dynamically. Our MO-QRT structure optimizes several conflicting objectives, using a multi-objective evolutionary algorithm that calculates a diverse set of high quality non-dominated solutions in a single run. Then a decision-making subsystem is utilized to tune the retrieval preferences of the query user. We assess our ideas both using trace-driven experiments with mobility and social patterns derived by Microsoft's GeoLife project, DBLP and Pics `n' Trails but also using our real Android SmartP2P ( http://smartp2p.cs.ucy.ac.cy/ ) system deployed over our SmartLab ( http://smartlab.cs.ucy.ac.cy/ ) testbed of 40+ smartphones. Our study reveals that SmartOpt yields high query recall rates of 95 %, with one order of magnitude less time and two orders of magnitude less energy than its competitors.

BloomCast: Efficient and Effective Full-Text Retrieval in Unstructured P2P Networks

Efficient and effective full-text retrieval in unstructured peer-to-peer networks remains a challenge in the research community. First, it is difficult, if not impossible, for unstructured P2P systems to effectively locate items with guaranteed recall. Second, existing schemes to improve search success rate often rely on replicating a large number of item replicas across the wide area network, incurring a large amount of communication and storage costs. In this paper, we propose BloomCast, an efficient and effective full-text retrieval scheme, in unstructured P2P networks. By leveraging a hybrid P2P protocol, BloomCast replicates the items uniformly at random across the P2P networks, achieving a guaranteed recall at a communication cost of O(√N), where N is the size of the network. Furthermore, by casting Bloom Filters instead of the raw documents across the network, BloomCast significantly reduces the communication and storage costs for replication. We demonstrate the power of BloomCast design through both mathematical proof and comprehensive simulations based on the query logs from a major commercial search engine and NIST TREC WT10G data collection. Results show that BloomCast achieves an average query recall of 91 percent, which outperforms the existing WP algorithm by 18 percent, while BloomCast greatly reduces the search latency for query processing by 57 percent.

Output URL bidding

Output URL bidding is a new bidding mechanism for sponsored search, where advertisers bid on search result URLs, as opposed to keywords in the input query. For example, an advertiser may want his ad to appear whenever the search result includes the sites www.imdb.com and en.wikipedia.org, instead of bidding on keywords that lead to these sites, e.g., movie titles or actor names. In this paper we study the tradeoff between the simplicity and the specification power of output bids and we explore their utility for advertisers. We first present a model to derive output bids from existing keyword bids. Then, we use the derived bids to experimentally study output bids and contrast them to input query bids. Our main results are the following: (1) Compact output bids that mix both URLs and hosts have the same specification power as more lengthy input bids; (2) Output bidding can increase the recall of relevant queries; and (3) Output and input biding can be combined into a hybrid mechanism that combines the benefits of both.

EXTRUCT

Users who are unfamiliar with database query languages can search XML data sets using keyword queries. Previous work has shown that current XML keyword search methods, although intuitive, do not effectively use the data's structural information and provide poor precision, recall, and ranking for most queries. Based on an extension of the concept of information theory, we have developed principled frameworks called normalized total correlation (NTC) and normalized term presence correlation (NTPC) to measure the relevance of candidate answers to keyword queries. We demonstrate EXTRUCT, an XML keyword search interface that uses NTC and NTPC. An extensive empirical evaluation over two real-world XML DBs has shown that EX-TRUCT has better precision and recall and provides better ranking than all previous approaches. We demonstrate EXTRUCT, along with seven other keyword search systems for four real-world XML data sets, using prepared queries as well as queries from the audience. The demonstration shows that using deep structural information increases the effectiveness of XML keyword search systems considerably.

Structural consistency: enabling XML keyword search to eliminate spurious results consistently

XML keyword search is a user-friendly way to query XML data using only keywords. In XML keyword search, to achieve high precision without sacrificing recall, it is important to remove spurious results not intended by the user. Efforts to eliminate spurious results have enjoyed some success using the concepts of LCA or its variants, SLCA and MLCA. However, existing methods still could find many spurious results. The fundamental cause for the occurrence of spurious results is that the existing methods try to eliminate spurious results locally without global examination of all the query results and, accordingly, some spurious results are not consistently eliminated. In this paper, we propose a novel keyword search method that removes spurious results consistently by exploiting the new concept of structural consistency. We define structural consistency as a property that is preserved if there is no query result having an ancestor-descendant relationship at the schema level with any other query results. A naive solution to obtain structural consistency would be to compute all the LCAs (or variants) and then to remove spurious results according to structural consistency. Obviously, this approach would always be slower than existing LCA-based ones. To speed up structural consistency checking, we must be able to examine the query results at the schema level without generating all the LCAs. However, this is a challenging problem since the schema-level query results do not homomorphically map to the instance-level query results, causing serious false dismissal. We present a comprehensive and practical solution to this problem and formally prove that this solution preserves structural consistency at the schema level without incurring false dismissal. We also propose a relevance-feedback-based solution for the problem where our method has low recall, which occurs when it is not the user’s intention to find more specific results. This solution has been prototyped in a full-fledged object-relational DBMS Odysseus developed at KAIST. Experimental results using real and synthetic data sets show that, compared with the state-of-the-art methods, our solution significantly (1) improves precision while providing comparable recall for most queries and (2) enhances the query performance by removing spurious results early.

A recall-based cluster formation game in peer-to-peer systems

In many large-scale content sharing applications, participants or peers are grouped together forming clusters based on their content or interests. In this paper, we deal with the maintenance of such clusters in the presence of updates. We model the evolution of the system as a strategic game, where peers determine their cluster membership based on a utility function of the query recall. Peers are guided either by selfish or altruistic motives: selfish peers aim at improving the recall of their own queries, whereas altruistic peers aim at improving the recall of the queries of other peers. We study the evolution of such clusters both theoretically and experimentally under a variety of conditions. We show that, in general, local decisions made independently by each peer enable the system to adapt to changes and maintain the overall recall of the query workload.

An adaptive overlay network inspired by social behaviour

Nature is a great source of inspiration for scientists, because natural systems seem to be able to find the best way to solve a given problem by using simple and robust mechanisms. Studying complex natural systems, scientists usually find that simple local dynamics lead to sophisticated macroscopic structures and behaviour. It seems that some kind of local interaction rules naturally allow the system to auto-organize itself as an efficient and robust structure, which can easily solve different tasks. Examples of such complex systems are social networks, where a small set of basic interaction rules leads to a relatively robust and efficient communication structure. In this paper, we present PROSA , a semantic peer-to-peer (P2P) overlay network inspired by social dynamics. The way queries are forwarded and links among peers are established in PROSA resemble the way people ask other people for collaboration, help or information. Behaving as a social network of peers, PROSA naturally evolves to a small world, where all peers can be reached in a fast and efficient way. The underlying algorithm used for query forwarding, based only on local choices, is both reliable and effective: peers sharing similar resources are eventually connected with each other, allowing queries to be successfully answered in a really small amount of time. The resulting emergent structure can guarantee fast responses and good query recall.

Difficulty-Aware Hybrid Search in Peer-to-Peer Networks

By combining an unstructured protocol with a DHT-based index, hybrid Peer-to-Peer (P2P) improves search efficiency in terms of query recall and response time. The key challenge in hybrid search is to estimate the number of peers that can answer a given query. Existing approaches assume that such a number can be directly obtained by computing item popularity. In this work, we show that such an assumption is not always valid, and previous designs cannot distinguish whether items related to a query are distributed in many peers or are in a few peers. To address this issue, we propose QRank, a difficulty-aware hybrid search, which ranks queries by weighting keywords based on term frequency. Using rank values, QRank selects proper search strategies for queries. We conduct comprehensive trace-driven simulations to evaluate this design. Results show that QRank significantly improves the search quality as well as reducing system traffic cost compared with existing approaches.

Digital forensic text string searching: Improving information retrieval effectiveness by thematically clustering search results

Current digital forensic text string search tools use match and/or indexing algorithms to search digital evidence at the physical level to locate specific text strings. They are designed to achieve 100% query recall (i.e. find all instances of the text strings). Given the nature of the data set, this leads to an extremely high incidence of hits that are not relevant to investigative objectives. Although Internet search engines suffer similarly, they employ ranking algorithms to present the search results in a more effective and efficient manner from the user's perspective. Current digital forensic text string search tools fail to group and/or order search hits in a manner that appreciably improves the investigator's ability to get to the relevant hits first (or at least more quickly). This research proposes and empirically tests the feasibility and utility of post-retrieval clustering of digital forensic text string search results – specifically by using Kohonen Self-Organizing Maps, a self-organizing neural network approach. This paper is presented as a work-in-progress. A working tool has been developed and experimentation has begun. Findings regarding the feasibility and utility of the proposed approach will be presented at DFRWS 2007, as well as suggestions for follow-on research.