Query Patterns Research Articles

Wind turbine fault detection and identification through self-attention-based mechanism embedded with a multivariable query pattern

Condition monitoring (CM) of wind turbines (WTs) is commonly accepted as an effective way to increase the availability and reduce the operation and maintenance (O&M) costs of wind farms. CM methods based on data from Supervisory Control and Data Acquisition (SCADA) system have garnered a great deal of interest due to their accessibility and low cost. Existing studies based on SCADA data provide whether the WT is normal or abnormal, and further the fault locations if an anomaly occurs. Nevertheless, few studies have attempted to identify the underlying causes of the anomaly. In this paper, a novel WTCM method utilizing a self-attention-based mechanism embedded with a multivariable query pattern is proposed for the anomaly detection and underlying causes identification. Firstly, an anomaly detection model composed of multiple cascaded encoders and decoders with a multi-head self-attention mechanism is proposed to extract cross-variable correlations. Furthermore, the multivariable query pattern is designed to evaluate the influences of different features on the target, which is essential for fault identification. Finally, the relative anomaly index (RAI) is proposed for each feature to quantify the impact of each feature on the anomaly. RAIs are adopted to analyze the underlying causes of anomalies. Experimental results confirmed the effectiveness of the proposed method for the WT fault early detection and identification of anomaly causes.

Approximate Content-Addressable Memories: A Review

Content-addressable memory (CAM) has been part of the memory market for more than five decades. CAM can carry out a single clock cycle lookup based on the content rather than an address. Thanks to this attractive feature, CAM is utilized in memory systems where a high-speed content lookup technique is required. However, typical CAM applications only support exact matching, as opposed to approximate matching, where a certain Hamming distance (several mismatching characters between a query pattern and the dataset stored in CAM) needs to be tolerated. Recent interest in approximate search has led to the development of new CAM-based alternatives, accelerating the processing of large data workloads in the realm of big data, genomics, and other data-intensive applications. In this review, we provide an overview of approximate CAM and describe its current and potential applications that would benefit from approximate search computing.

Robust Query Driven Cardinality Estimation under Changing Workloads

Query driven cardinality estimation models learn from a historical log of queries. They are lightweight, having low storage requirements, fast inference and training, and are easily adaptable for any kind of query. Unfortunately, such models can suffer unpredictably bad performance under workload drift, i.e., if the query pattern or data changes. This makes them unreliable and hard to deploy. We analyze the reasons why models become unpredictable due to workload drift, and introduce modifications to the query representation and neural network training techniques to make query-driven models robust to the effects of workload drift. First, we emulate workload drift in queries involving some unseen tables or columns by randomly masking out some table or column features during training. This forces the model to make predictions with missing query information, relying more on robust features based on up-to-date DBMS statistics that are useful even when query or data drift happens. Second, we introduce join bitmaps, which extends sampling-based features to be consistent across joins using ideas from sideways information passing. Finally, we show how both of these ideas can be adapted to handle data updates. We show significantly greater generalization than past works across different workloads and databases. For instance, a model trained with our techniques on a simple workload (JOBLight-train), with 40 k synthetically generated queries of at most 3 tables each, is able to generalize to the much more complex Join Order Benchmark, which include queries with up to 16 tables, and improve query runtimes by 2× over PostgreSQL. We show similar robustness results with data updates, and across other workloads. We discuss the situations where we expect, and see, improvements, as well as more challenging workload drift scenarios where these techniques do not improve much over PostgreSQL. However, even in the most challenging scenarios, our models never perform worse than PostgreSQL, while standard query driven models can get much worse than PostgreSQL.

S3QLRDF: distributed SPARQL query processing using Apache Spark—a comparative performance study

The proliferation of semantic data in the form of Resource Description Framework (RDF) triples demands an efficient, scalable, and distributed storage along with a highly available and fault-tolerant parallel processing strategy. There are three open issues with distributed RDF data management systems that are not well addressed altogether in existing work. First is the querying efficiency, second is that solutions are optimized for certain types of query patterns and don’t necessarily work well for all types, and third is concerned with reducing pre-processing cost. More precisely, the rapid growth of RDF data raises the need for an efficient partitioning strategy over distributed data management systems to improve SPARQL (SPARQL Protocol and RDF Query Language) query performance regardless of its pattern shape with minimized pre-processing time. In distributed RDF systems, both the data and the query processing are highly distributed. On the other hand, SPARQL workloads are dynamic and structurally diverse that can have different degrees of complexity. A complex SPARQL query over a large RDF graph in distributed systems requires combining a lot of distributed pieces of data through join operations. Therefore, designing an efficient data-partitioning schema and join strategy to minimize data transfer is the fundamental challenge in distributed RDF data management systems. In this context, we propose a new relational partitioning schema called Property Table Partitioning (PTP) for RDF data, that further partitions existing Property Table into multiple tables based on distinct properties (comprising of all subjects with non-null values for those distinct properties) in order to minimize input size and number of join operations of a query. This paper proposed a distributed RDF data management system called S3QLRDF, which is built on top of Spark and utilizes SQL to execute SPARQL queries over PTP schema. The experimental analysis with respect to preprocessing costs and query performance, using synthetic and real datasets shows that S3QLRDF outperforms state-of-the-art distributed RDF management systems.

High-Performance Row Pattern Recognition Using Joins

The SQL standard introduced MATCH_RECOGNIZE in 2016 for row pattern recognition. Since then, MATCH_RECOGNIZE has been supported by several leading relation systems, they implemented this function using Non-Deterministic Finite Automaton (NFA). While NFA is suitable for pattern recognition in streaming scenarios, the current uses of NFA by the relational systems for historical data analysis scenarios overlook important optimization opportunities. We propose a new approach to use Join to speed up row pattern recognition in historical analysis scenarios for relational systems. Implemented as a logical plan rewrite rule, the new approach first filters the input relation to MATCH_RECOGNIZE using Joins constructed based on a subset of symbols taken from the PATTERN expression, then run the NFA-based MATCH_RECOGNIZE on the filtered rows, reducing the net cost. The rule also includes a specialized cardinality model for the Joins and a cost model for the NFA-based MATCH_RECOGNIZE operator for choosing an appropriate symbol set. The rewrite rule is applicable when the query pattern's definition is self-contained and either the input table has no duplicates or there is a window condition. Applying the rewrite rule to a query benchmark with 1,800 queries spanning over 6 patterns and 3 pattern definitions, we observed median speedups of 5.4X on Trino (v373 with ORC files on Hive), 57.5X on SQL Server (2019) using column store and 41.6X on row store.

Visual and Quantitative Analyses of Virus Genomic Sequences using a Metric-based Algorithm

This work aims to study the virus RNAs using a novel accelerated algorithm to explore any-length repetitive genomic fragments in sequences using Hamming distance between the binary-expressed characters of an RNA and a query pattern. Primary attention is paid to the building and analyzing 1-D distributions (walks) of atg-patterns - codon-starting triplets in genomes. These triplets compose a distributed set called a word scheme of RNA. A complete genome map is built by plotting the mentioned atg-walks, trajectories of separate (a-, c-, g-, and t-symbols) nucleotides, and the lines designating the genomic words. The said map can be additionally equipped by gene’s designations making this tool pertinent for multi-scale genomic analyses. The visual examination of atg-walks is followed by calculating statistical parameters of genomic sequences, including estimating walk- geometry deviation of RNAs and fractal properties of word-length distributions. This approach is applied to the SARS CoV-2, MERS CoV, Dengue, and Ebola viruses, whose complete genomic sequences are taken from GenBank and GISAID. The relative stability of these walks for SARS CoV-2 and MERS CoV viruses was found, unlike the Dengue and Ebola distributions that showed an increased deviation of their geometrical and fractal characteristics. The developed approach can be useful in further studying mutations of viruses and building their phylogenic trees.

Evidential Combination of Classifiers for Imbalanced Data

It remains an important research topic for the classification of imbalanced data. There exist some methods to solve this problem, such as hybrid-sampling, over-sampling, and under-sampling. Each method has its own advantage, and different methods generally provide some complementary knowledge. We want to combine these three methods at the decision level in an appropriate way for achieving as good as possible classification performance. Evidence theory is expert at representing and combining uncertain information. So a new method called an evidential combination of classifiers (ECC) is proposed for dealing with imbalanced data. The classification result generated by different strategies (i.e., hybrid-sampling, over-sampling, or under-sampling) may have different reliabilities for query patterns. A cautious reliability evaluation rule is developed for each classification result based on the close neighborhoods. After that, the classification result is revised with a new belief redistribution way according to the reliability evaluation, and the probability/belief of one class can be partially transferred to other classes as well as the total ignorance, which is defined by the whole frame of classes. By doing this, we can reduce the error risk of each classification method. Then, the revised classification results from different methods are combined by evidence theory to make the final class decision. The effectiveness of the ECC method has been demonstrated using several experiments, and it shows that ECC can effectively improve the classification performance comparing with other related methods.

A workload-driven method for designing aggregate-oriented NoSQL databases

Due to the scalability and availability problems with traditional relational database systems, a variety of NoSQL stores have emerged over the last decade to deal with big data. How data are structured in a NoSQL store has a large impact on the query and update performance and the storage usage. Thus, different from the traditional database design, not only the data structure but also the data access patterns need to be considered in the design of NoSQL database schemas. In this paper, we present a general workload-driven method for designing key-value, wide-column, and document NoSQL database schemas. We first present a generic logical model Query Path Graph (QPG) that can represent the data structures of the UML class diagram. We also define mappings from the SQL-based query patterns to QPG and from QPG to aggregate-oriented NoSQL schemas. We use a cost model to measure the query and update performance and optimize the QPG schemas. We evaluate the proposed method with several typical case studies by simulating workloads on databases with different schema designs. The results demonstrate that our method preserves the generality and the quality of the design.

A New Belief-Based Bidirectional Transfer Classification Method.

In pattern classification, we may have a few labeled data points in the target domain, but a number of labeled samples are available in another related domain (called the source domain). Transfer learning can solve such classification problems via the knowledge transfer from source to target domains. The source and target domains can be represented by heterogeneous features. There may exist uncertainty in domain transformation, and such uncertainty is not good for classification. The effective management of uncertainty is important for improving classification accuracy. So, a new belief-based bidirectional transfer classification (BDTC) method is proposed. In BDTC, the intraclass transformation matrix is estimated at first for mapping the patterns from source to target domains, and this matrix can be learned using the labeled patterns of the same class represented by heterogeneous domains (features). The labeled patterns in the source domain are transferred to the target domain by the corresponding transformation matrix. Then, we learn a classifier using all the labeled patterns in the target domain to classify the objects. In order to take full advantage of the complementary knowledge of different domains, we transfer the query patterns from target to source domains using the K-NN technique and do the classification task in the source domain. Thus, two pieces of classification results can be obtained for each query pattern in the source and target domains, but the classification results may have different reliabilities/weights. A weighted combination rule is developed to combine the two classification results based on the belief functions theory, which is an expert at dealing with uncertain information. We can efficiently reduce the uncertainty of transfer classification via the combination strategy. Experiments on some domain adaptation benchmarks show that our method can effectively improve classification accuracy compared with other related methods.

Hardware-Algorithm Codesign for Fast and Energy Efficient Approximate String Matching on FPGA for Computational Biology.

Myers bit-vector algorithm for approximate string matching (ASM) is a dynamic programming based approach that takes advantage of bit-parallel operations. It is one of the fastest algorithms to find the edit distance between two strings. In computational biology, ASM is used at various stages of the computational pipeline, including proteomics and genomics. The computationally intensive nature of the underlying algorithms for ASM operating on the large volume of data necessitates the acceleration of these algorithms. In this paper, we propose a novel ASM architecture based on Myers bit-vector algorithm for parallel searching of multiple query patterns in the biological databases. The proposed parallel architecture uses multiple processing engines and hardware/software codesign for an accelerated and energy-efficient design of ASM algorithm on hardware. In comparison with related literature, the proposed design achieves 22× better performance with a demonstrative energy efficiency of ∼500×109 cell updates per joule.

Expressive querying for accelerating visual analytics

Data visualization is the primary means by which data analysts explore patterns, trends, and insights in their data. Unfortunately, existing visual analytics tools offer limited expressiveness and scalability when it comes to searching for visualizations over large datasets, making visual data exploration labor-intensive and time-consuming. In this work, we introduce the problem of visualization search and highlight two underlying challenges of search enumeration and visualization matching. To address them, we first present our work on Zenvisage that helps enumerate large collections of visualizations and supports simple visualization matching with the help of an interactive interface and an expressive visualization query language. For more finegrained and flexible visualization matching, including search for underspecified and approximate patterns, we extend Zenvisage to develop ShapeSearch. ShapeSearch supports a novel shape querying algebra that helps express a large class of pattern queries that are hard to specify with existing systems. ShapeSearch exposes multiple specification mechanisms: sketch, natural-language, and visual regular expressions that help users easily issue shape queries, while applying query-aware and perceptually-aware optimizations to efficiently execute them within interactive response times. To conclude, we discuss a number of open research problems to further improve the usability and performance of both Zenvisage and ShapeSearch.

FPMBot: Discovering the frequent pattern of IoT-botnet domain queries in large-scale network

With the rapid development of Internet of Things (IoT), new types of security issues has emerged, and one of the most severe one is the IoT-based botnet. A number of traditional works dedicate in analyzing DNS traffic for its wide misuse in botnets. However, most works are limited to specific behaviors, such as regularity in domains queries and periodicity in C&C(Command and Control) processes, which are often sheltered by attackers. Moreover, ever-growing numbers of domains and queries caused by IoT devices make most approaches not effective anymore since their architectures are hard to confront traffic explosion. In this paper, we illustrate the essential property of botnet, i.e., the frequent pattern of DNS request relationships, and propose a generic and scalable IoT-botnet detection system, named FPMBot, to detect bots domain queries in large-scale DNS traffic. The key insight is that bots in the same botnet inevitably query same sets of domains or servers whenever they try to conduct attacks or connect to C&C servers, and form frequent patterns in a bipartite graph of requesters and responsers. This frequent pattern in domain queries is an essential behavior for bots since a significant advantage of bots is the ability to launch large-scale attacks synchronously. We utilize a frequent pattern mining algorithm to detect such patterns and implement FPMBot based on Apache Spark parallel computing architecture to handle the daily increasing traffic. The experiment result on more than 14 billion records in four days real-world DNS logs shows that FPMBot has a high detection precision over 95% and performs well in large-scale network.

CORE

Complex Event Recognition (CER) systems are a prominent technology for finding user-defined query patterns over large data streams in real time. CER query evaluation is known to be computationally challenging, since it requires maintaining a set of partial matches, and this set quickly grows super-linearly in the number of processed events. We present CORE, a novel COmplex event Recognition Engine that focuses on the efficient evaluation of a large class of complex event queries, including time windows as well as the partition-by event correlation operator. This engine uses a novel automaton-based evaluation algorithm that circumvents the super-linear partial match problem: under data complexity, it takes constant time per input event to maintain a data structure that compactly represents the set of partial matches and, once a match is found, the query results may be enumerated from the data structure with output-linear delay. We experimentally compare CORE against state-of-the-art CER systems on real-world data. We show that (1) CORE's performance is stable with respect to both query and time window size, and (2) CORE outperforms the other systems by up to five orders of magnitude on different workloads.

Brain Tumor Infodemiology: Worldwide Online Health-Seeking Behavior Using Google Trends and Wikipedia Pageviews.

BackgroundSearching the internet for information on common neurologic symptoms and diseases has been increasing in recent times. It is postulated that online search volume data could be utilized to gauge public awareness and real-world epidemiological data regarding brain tumors.ObjectivesThe goal of this study was to describe the pattern of online search queries of keywords related to neoplasms of the central nervous system (CNS).MethodsUsing Google Trends, search activity from January 2004 – January 2021 was quantified using relative search volume (RSV). The average RSV for the first 3 years was compared with the final 3 years to account for percent change. Wikipedia article views from July 2007 – January 2021 were generated using Pageviews. Peaks in RSV and page views were then matched for related news.Results“Brain tumor”, “brain cancer”, “glioblastoma”, and “glioma” had the highest search volume. RSV from Google Trends and views of Wikipedia pages reflected comparable data in terms of known prevalence rankings of tumor subtypes. There were no observable trends that could correlate to the rising numbers of brain tumor cases worldwide. However, headlines of personalities being diagnosed with glioblastomas were mostly responsible for temporary increases in public interest.ConclusionsTransient rises in online search volume mirror public awareness of more aggressive CNS neoplasms that have a high burden of disease. Worldwide interest in brain tumors may not necessarily correspond to clinical occurrence, but may signify an unmet gap in providing accurate online information to neuro-oncologic patients.

Attention-based Local Mean [formula omitted]-Nearest Centroid Neighbor Classifier

Among classic data mining algorithms, the K-Nearest Neighbor (KNN)-based methods are effective and straightforward solutions for the classification tasks. However, most KNN-based methods do not fully consider the impact across different training samples in classification tasks, which leads to performance decline. To address this issue, we propose a method named Attention-based Local Mean K-Nearest Centroid Neighbor Classifier (ALMKNCN), bridging the nearest centroid neighbor computing with the attention mechanism, which fully considers the influence of each training query sample. Specifically, we first calculate the local centroids of each class with the given query pattern. Then, our ALMKNCN introduces the attention mechanism to calculate the weight of pseudo-distance between the test sample to each class centroid. Finally, based on attention coefficient, the distances between the query sample and local mean vectors are weighted to predict the classes for query samples. Extensive experiments are carried out on real data sets and synthetic data sets by comparing ALMKNCN with the state-of-art KNN-based methods. The experimental results demonstrate that our proposed ALMKNCN outperforms the compared methods with large margins.

Belief Combination of Classifiers for Incomplete Data

Data with missing values, or incomplete information, brings some challenges to the development of classification, as the incompleteness may significantly affect the performance of classifiers. In this paper, we handle missing values in both training and test sets with uncertainty and imprecision reasoning by proposing a new belief combination of classifier (BCC) method based on the evidence theory. The proposed BCC method aims to improve the classification performance of incomplete data by characterizing the uncertainty and imprecision brought by incompleteness. In BCC, different attributes are regarded as independent sources, and the collection of each attribute is considered as a subset. Then, multiple classifiers are trained with each subset independently and allow each observed attribute to provide a sub-classification result for the query pattern. Finally, these sub-classification results with different weights (discounting factors) are used to provide supplementary information to jointly determine the final classes of query patterns. The weights consist of two aspects: global and local. The global weight calculated by an optimization function is employed to represent the reliability of each classifier, and the local weight obtained by mining attribute distribution characteristics is used to quantify the importance of observed attributes to the pattern classification. Abundant comparative experiments including seven methods on twelve datasets are executed, demonstrating the out-performance of BCC over all baseline methods in terms of accuracy, precision, recall, F1 measure, with pertinent computational costs.

Query Optimization for Distributed Spatio-Temporal Sensing Data Processing

The unprecedented development of Internet of Things (IoT) technology produces humongous amounts of spatio-temporal sensing data with various geometry types. However, processing such datasets is often challenging due to high-dimensional sensor data geometry characteristics, complex anomalistic spatial regions, unique query patterns, and so on. Timely and efficient spatio-temporal querying significantly improves the accuracy and intelligence of processing sensing data. Most existing query algorithms show their lack of supporting spatio-temporal queries and irregular spatial areas. In this paper, we propose two spatio-temporal query optimization algorithms based on SpatialHadoop to improve the efficiency of query spatio-temporal sensing data: (1) spatio-temporal polygon range query (STPRQ), which aims to find all records from a polygonal location in a time interval; (2) spatio-temporal k nearest neighbors query (STkNNQ), which directly searches the query point’s k closest neighbors. To optimize the STkNNQ algorithm, we further propose an adaptive iterative range optimization algorithm (AIRO), which can optimize the iterative range of the algorithm according to the query time range and avoid querying irrelevant data partitions. Finally, extensive experiments based on trajectory datasets demonstrate that our proposed query algorithms can significantly improve query performance over baseline algorithms and shorten response time by 81% and 35.6%, respectively.

ReportAGE: Automatically extracting the exact age of Twitter users based on self-reports in tweets.

Advancing the utility of social media data for research applications requires methods for automatically detecting demographic information about social media study populations, including users’ age. The objective of this study was to develop and evaluate a method that automatically identifies the exact age of users based on self-reports in their tweets. Our end-to-end automatic natural language processing (NLP) pipeline, ReportAGE, includes query patterns to retrieve tweets that potentially mention an age, a classifier to distinguish retrieved tweets that self-report the user’s exact age (“age” tweets) and those that do not (“no age” tweets), and rule-based extraction to identify the age. To develop and evaluate ReportAGE, we manually annotated 11,000 tweets that matched the query patterns. Based on 1000 tweets that were annotated by all five annotators, inter-annotator agreement (Fleiss’ kappa) was 0.80 for distinguishing “age” and “no age” tweets, and 0.95 for identifying the exact age among the “age” tweets on which the annotators agreed. A deep neural network classifier, based on a RoBERTa-Large pretrained transformer model, achieved the highest F1-score of 0.914 (precision = 0.905, recall = 0.942) for the “age” class. When the age extraction was evaluated using the classifier’s predictions, it achieved an F1-score of 0.855 (precision = 0.805, recall = 0.914) for the “age” class. When it was evaluated directly on the held-out test set, it achieved an F1-score of 0.931 (precision = 0.873, recall = 0.998) for the “age” class. We deployed ReportAGE on a collection of more than 1.2 billion tweets, posted by 245,927 users, and predicted ages for 132,637 (54%) of them. Scaling the detection of exact age to this large number of users can advance the utility of social media data for research applications that do not align with the predefined age groupings of extant binary or multi-class classification approaches.

Question Answering Over Knowledge Graphs: A Case Study in Tourism

Over the recent years, a large number of knowledge graphs (KGs) have been developed to store and present small and medium enterprises’ data in the form of subject-predicate-object triples. The KGs are not easily accessible to end-users because they need an understanding of query languages and KGs’ structures. To assist end-users in accessing these KGs, question answering over KGs targets to provide answers for natural language questions (NLQs). This paper proposes an approach to answer questions over small and medium scaled KGs based on graph isomorphism in two phases: (1) offline phase and (2) semantic parsing phase. In the offline phase, a semi-automated solution is proposed to generate NLQs and their answers, which are used to train machine learning models employed in the next phase. In the semantic parsing phase, a given input NLQ is mapped into a query pattern according to its grammatical structure. Each query pattern contains some slots that need to be filled with corresponding entities, classes and relations from the KG. While string and semantic similarity metrics are applied to identify the entities and classes, the probability distribution of the relations is used to extract the relations. The Cartesian product of the identified entities, classes and relations is utilized to fill the slots, derive SPARQL queries and finally retrieve the answers. To evaluate the proposed approach, we use SalzburgerLand KG, a real KG describing touristic entities of the region of Salzburg, Austria. Our results show that the approach improves the end-to-end user experience in terms of interactive question answering and performance.

A General Cardinality Estimation Framework for Subgraph Matching in Property Graphs

We introduce a framework for cardinality estimation of query patterns over property graph databases. This framework makes it possible to analyze, compare and combine different cardinality estimation approaches. It consists of three phases: obtaining a set of estimates for some subqueries, extending this set and finally combining the set into a single cardinality estimate for the query. We show that (parts of) many existing cardinality estimation approaches can be used as techniques in one of the phases from our framework. The phases are loosely coupled, making it possible to combine (parts of) current cardinality estimation approaches. We created a graph version of the Join Order Benchmark to perform experiments with different combinations of techniques. The results showed that query patterns without property constraints can be accurately estimated using synopses for small patterns. Accurate estimation of query patterns with property constraints require new estimation techniques to be developed that capture correlations between the property constraints and the topology in graph databases.