Metasearch via Voting

Most existing approaches to visual search reranking predominantly focus on mining information within the initial search results. However, the initial ranked list cannot provide enough cues for reranking by itself due to the typically unsatisfying visual search performance. This paper presents a new method for visual search reranking called CrowdReranking, which is characterized by mining relevant visual patterns from image search results of multiple search engines which are available on the Internet. Observing that different search engines might have different data sources for indexing and methods for ranking, it is reasonable to assume that there exist different search results yet certain common visual patterns relevant to a given query among those results. We first construct a set of visual words based on the local image patches collected from multiple image search engines. We then explicitly detect two kinds of visual patterns, i.e., salient and concurrent patterns, among the visual words. Theoretically, we formalize reranking as an optimization problem on the basis of the mined visual patterns and propose a close-form solution. Empirically, we conduct extensive experiments on several real-world search engines and one benchmark dataset, and show that the proposed CrowdReranking is superior to the state-of-the-art works.

Frequently a user's information needs are stored in the databases of multiple search engines. It is inconvenient and inefficient for an ordinary user to invoke multiple search engines and identify useful documents from the returned results. To support unified access to multiple search engines, a metasearch engine can be constructed. When a metasearch engine receives a query from a user, it invokes the underlying search engines to retrieve useful information for the user. Metasearch engines have other benefits as a search tool such as increasing the search coverage of the Web and improving the scalability of the search. In this article, we survey techniques that have been proposed to tackle several underlying challenges for building a good metasearch engine. Among the main challenges, the database selection problem is to identify search engines that are likely to return useful documents to a given query. The document selection problem is to determine what documents to retrieve from each identified search engine. The result merging problem is to combine the documents returned from multiple search engines. We will also point out some problems that need to be further researched.

Relevance evaluation is an important topic in Web search engine research. Traditional evaluation methods resort to huge amount of human efforts which lead to an extremely time-consuming process in practice. With analysis on large scale user query logs and click-through data, we propose a performance evaluation method that fully automatically generates large scale Web search topics and answer sets under Cranfield framework. These query-to-answer pairs are directly utilized in relevance evaluation with several widely-adopted precision/recall-related retrieval performance metrics. Besides single search engine log analysis, we propose user behavior models on multiple search engines' click-through logs to reduce potential bias among different search engines. Experimental results show that the evaluation results are similar to those gained by traditional human annotation, and our method avoids the propensity and subjectivity of manual judgments by experts in traditional ways.

We propose UniScore as a metric for integrating and standardizing the outputs of multiple search engines in the analysis of data-dependent acquisition (DDA) data from LC/MS/MS-based bottom-up proteomics. UniScore is calculated from the annotation information attached to the product ions alone by matching the amino acid sequences of candidate peptides suggested by the search engine with the product ion spectrum. The acceptance criteria are controlled independently of the score values by using the false discovery rate based on the target-decoy approach. Compared to other rescoring methods that use deep learning-based spectral prediction, larger amounts of data can be processed using minimal computing resources. When applied to large-scale global proteome data and phosphoproteome data, the UniScore approach outperformed each of the conventional single search engines examined (Comet, X! Tandem, Mascot, and MaxQuant). Furthermore, UniScore could also be directly applied to peptide matching in chimeric spectra without any additional filters.

Voting is an important operation in multichannel computation paradigm and realization of ultrareliable and real-time control systems that arbitrates among the results of N redundant variants. These systems include N-modular redundant (NMR) hardware systems and diversely designed software systems based on N-version programming (NVP). Depending on the characteristics of the application and the type of selected voter, the voting algorithms can be implemented for either hardware or software systems. In this paper, a novel voting algorithm is introduced for real-time fault-tolerant control systems, appropriate for applications in which N is large. Then, its behavior has been software implemented in different scenarios of error-injection on the system inputs. The results of analyzed evaluations through plots and statistical computations have demonstrated that this novel algorithm does not have the limitations of some popular voting algorithms such as median and weighted; moreover, it is able to significantly increase the reliability and availability of the system in the best case to 2489.7% and 626.74%, respectively, and in the worst case to 3.84% and 1.55%, respectively.

N-version programming is one of the most common techniques which is used to improve the reliability of software by building in fault tolerance, redundancy and decreasing common cause failures. N different equivalent software versions are developed by N different and isolated workgroups by considering the same software specifications. The versions solve the same task and return results that have to be compared to determine the correct result. Decisions of N different versions are evaluated by a voting algorithm or the so-called voter. In this paper, two of the most commonly used software voting algorithms such as the majority voting algorithm and the consensus voting algorithm are studied. The distinctive features of Nversion programming with majority voting and N-version programming with consensus voting are described. These two algorithms make a decision about the correct result on the base of the agreement matrix. However, if the equivalence relation on the agreement matrix is not satisfied it is impossible to make a decision. It is shown that the agreement matrix can be transformed into an appropriate form by using the Boolean compositions when the equivalence relation is satisfied.

An important problem in fault-tolerant distributed computer systems is maintaining agreement between nonfaulty processes in the presence of undiagnosed faults. Approximate agreement defines a condition in which it is not necessary for the agreed values to be numerically identical. Rather, processes need only agree with each other to within a predefined numerical tolerance. Convergent voting algorithms which achieve approximate agreement have been studied in the context of two classes of systems, synchronous and asynchronous. Studies have also addressed both completely connected and partially connected systems. Together, the two properties of synchrony and connectivity yield 4 different voting domains. In all studies to date, each voting domain has been treated as a separate problem. This paper: shows that for at least one broad family of voting algorithms, the 4 domains are special cases of a more general convergent voting problem; analyzes convergent voting under the 3-mode hybrid fault model of Thambidurai and Park; and presents a set of unifying relations applicable to all 4 voting domains. These relations are used to specify voting algorithms which optimize fault-tolerance, convergence rate, or computational overhead in any given voting domain. The task of designing a voting algorithm for a particular fault-tolerant system is thus greatly simplified.

A TMR Genetic Voting Algorithm for Fault-tolerant Medical Robot

Estimation of solubility of acid gases in ionic liquids using different machine learning methods

Voting algorithms play a significant role in most fault-tolerant and control systems so these algorithms are continually in progress and with regard to specific situations in different systems, new voting algorithms are growing and increase the reliability and fault tolerability of systems. In this paper, a new combined algorithm using a weighted voter based on Markov chain is presented that is increased the reliability of system than both of median voter and average voter. In this paper, the proposed algorithm on average weather data of a year in Tehran was tested and simulation results show the effectiveness of the method than the other methods.

Retracted.

A metasearch engine is a system that supports uni ed access to multiple local search engines. Database selection is one of the main challenges in building a large-scale metasearch engine. The problem is to eAEciently and accurately determine a small number of potentially useful local search engines to invoke for each user query. In order to enable accurate selection, metadata that re ect the contents of each search engine need to be collected and used. In this paper, we propose a highly scalable and accurate database selection method. This method has several novel features. First, the metadata for representing the contents of all search engines are organized into a single integrated representative. Such a representative yields both computation eAEciency and storage eAEciency. Second, our selection method is based on a theory for ranking search engines optimally. Experimental results indicate that this new method is very e ective. An operational prototype system has been built based on the proposed approach.

Using data from the Pew Internet and American Life Project surveys, this article explores changing trends in reported sophistication and satisfaction with search skills and with search engines. We find that the proportion of Internet users searching online for answers to specific questions—as opposed to casual browsing—has grown significantly. Moreover, as users get more experience online, they increasingly become dependent on search engines, confident in their findings, and savvy about how search engines structure information, privilege paid results, and track users. When other factors are controlled, years of online experience is a strong predictor of the likelihood of a person doing specific searches on a daily basis, and experience can have an even stronger positive effect than education and income. We also find that years of online experience, frequency of use, and sophistication with multiple search engines can overcome socio-economic status in predicting how active a person is in searching across different topics.

Finding relevant data in the biomedical literature can be difficult sometime. To select a few neuroscience examples, suppose you would like to know (a) which serotonin receptor subunits are expressed in dentate gyrus mossy cells; (b) the average volume of the amygdala in adult male chimpanzees; (c) whether there is an EEG signature of CreutzfeldJacob disease; or (d) studies reporting bilateral fMRI activity in Brodmann area 38. These are fairly simple questions, for which standard search engines should fare relatively well. Yet even for these kinds of questions, securing the relevant information takes much longer than googling up the local weather forecast for the week-end or tomorrow’s commuter train schedule. The actual data required to build biologically realistic computational models are often more detailed: what is the time constant of the excitatory synaptic current from a specified pair of neuron types? Finding the answer in this case might require many hours or even days of queries over multiple search engines. Most importantly, the results of these queries must be typically followed by at least cursory reading of dozens of papers. When the graduate student triumphantly brings to the lab the needed reference, the adviser could mumble without lifting the eyes from the keyboard “that’s in young animals, and it was recorded at room temperature”. Another unfortunate major limitation is that, until and unless a definitive answer is found, it is usually impossible to know whether the information is available or not. In other words, existing biomedical search engines are ill-equipped to inform users that something is not yet known. Standard search algorithms such as PubMed are less than ideal to deal with data identification, because they are ultimately based on matching strings or concepts that appear in the title, abstract, and the keywords. These texts, however, are written with narrow scientific agendas in mind. The authors cannot possibly provide a list of keywords that would encompass all research projects for which some data in their articles might be relevant. If the topic of a report is the molecular phenotyping of a new genetic model of schizophrenia, the technical details of the deconvolution algorithm to deblur the optical micrographs would be nearly impossible to pick up through keyword searches. Could we devise a procedure to interrogate the scientific literature so as to extract accurately and efficiently most if not all of the relevant data? Is there a literature mining protocol that could give us the confidence that, if the query returns a blank, it means that the sought data is not yet available? Although common to all of biomedical science, this issue is particularly critical in neuroscience because of its unmatched diversity of dimensions, scales, questions, approaches, and techniques. Thus, effective tagging of publications with relevant metadata remains an outstanding neuroinformatics challenge. Full text searches provide half of the solution, in that they eliminate many of the issues related to false negatives. Many of the helpful terms to identify relevant data, for example, appear in the Materials and Methods sections of published articles rather than in their titles, abstracts, and keywords. Search engines scanning through the entire main text of publications include early visionary projects such as Textpresso, which started within the limited domain of C. elegans, then expanded to a

We would like to respond to the article entitled ‘‘Validity and reliability of a systematic database search strategy to identify publications resulting from pharmacy residency research projects.’’ We agree with the authors’ novel approach to confirm publication via a survey and to identify resident publications via a systematic approach to validate the utility of the database. The authors suggest that this approach can be used to compare publication success between residency programs. Additionally, this information can be used for marketing purposes for the individual residency program. One limitation noted by the authors was the use of only PubMed and EMBASE as the search engines to identify resident publications in the study and suggest other search engines, such as International Pharmaceutical Abstracts (IPA) and Google Scholar, may be used as alternatives. We aim to describe the methods and results of our project in which we assessed the utility of 5 different search engines to identify resident publications. Briefly, we identified the publication rate of abstracts presented at the 2003 (n 1⁄4 171), 2005 (n 1⁄4 221), and 2007 (n 1⁄4 273) Great Lakes Pharmacy Resident Conference (GLPRC) was 11.4% (n 1⁄4 76). We used a similar search strategy in Kwak et al to identify resident publication using the search engines (1) Scopus, (2) IPA, and (3) MEDLINE (PubMed). For this project, we used information from the abstracts that were previously confirmed for publications (n 1⁄4 76) to identify the proportion of these publications found in each of the following search engines: (1) Scopus, (2) IPA, (3) MEDLINE (PubMed), (4) EMBASE, and (5) Google Scholar. Two investigators (SMV and PMS) identified published research projects, and any disagreement between investigators was reviewed for consensus. A kappa statistic was calculated to determine interrater reliability for each search engine. The proportion of the published research was compared for each of the 5 search engines. Additionally, all pairs of search engines were assessed to determine the proportion published using combinations of searches. The proportions of published works were identified in the following search engines: Scopus (90.8%), IPA (75%), MEDLINE (PubMed; 80.3%), EMBASE (88.2%), and Google Scholar (90.8%; Table 1). Additionally, a higher proportion of nonpharmacy journals were identified using the search strategy compared to pharmacy journals (Table 1) in each individual search engine. After assessing various combinations, only Scopus plus IPA or Google Scholar plus IPA identified all 76 publications (Table 2). Based on our results, no single search engine was able to identify all publications. All 9 potential pairs were assessed. Only the combination use of either Scopus and IPA or Google Scholar and IPA resulted in identifying all 76 publications. The use of MEDLINE (PubMed) and EMBASE resulted in the poorest combination of search engines, as it only identified 88.2% (n 1⁄4 67) of publications. Kwak et al used the combination of MEDLINE (PubMed) and EMBASE with great success; however, the utility of this combination may differ in a general population of residency programs. Residents of certain geographic areas may submit to journals indexed to different search engines which may have contributed to the differences between our study and Kwak et al, as we assessed abstracts from the Great Lakes region. Additionally, confirmed publications in our previous search strategy used Scopus, IPA, and MEDLINE (PubMed) and not a validated measure which may have influenced our results. In conclusion, we affirm that a validated measure approach, as used by Kwak et al, is the gold standard for confirming the utility of various search engines; however, it may be difficult to replicate in a larger, generalized population of residents. The use of a single search engine is likely not adequate to identify all published residents’ studies. Regional differences may exist in the use of search engines, as the combination of MEDLINE (PubMed) and EMBASE resulted in a high sensitivity in Kwak