Scaling Parallel Rule-Based Reasoning

Peters, Martin; Brink, Christopher; Sachweh, Sabine; Zündorf, Albert

doi:10.1007/978-3-319-07443-6_19

Cited by 13 publications

(11 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These results are still not suitable for the real-time system at this stage. However, the optimisation opportunities such as multithread safe reasoning [46], ADL threads management, parallel programming, partitioning workload to graphics processing units (GPUs) [47], and using a Table 5: Summary of recent KB approaches machine with higher number of cores (i.e., quad-core, octa-core CPU or higher)…”

Section: Discussionmentioning

confidence: 99%

A semantics-based approach to sensor data segmentation in real-time Activity Recognition

Triboan

Chen

et al. 2019

Future Generation Computer Systems

View full text Add to dashboard Cite

Activity Recognition (AR) is key in context-aware assistive living systems. One challenge in AR is the segmentation of observed sensor events when interleaved or concurrent activities of daily living (ADLs) are performed. Several studies have proposed methods of separating and organising sensor observations and recognise generic ADLs performed in a simple or composite manner. However, little has been explored in semantically distinguishing individual sensor events directly and passing it to the relevant ongoing/new atomic activities. This paper proposes Semiotic theory inspired ontological model, capturing generic knowledge and inhabitant-specific preferences for conducting ADLs to support the segmentation process. A multithreaded decision algorithm and system prototype were developed and evaluated against 30 use case scenarios where each event was simulated at 10sec interval on a machine with i7 2.60GHz CPU, 2 cores and 8GB RAM. The result suggests that all sensor events were adequately segmented with 100% accuracy for single ADL scenarios and minor improvement of 97.8% accuracy for composite ADL scenario. However, the performance has suffered to segment each event with the average classification time of 3971ms and 62183ms for single and composite ADL scenarios, respectively.

show abstract

Section: Discussionmentioning

confidence: 99%

A semantics-based approach to sensor data segmentation in real-time Activity Recognition

Triboan

Chen

et al. 2019

Future Generation Computer Systems

View full text Add to dashboard Cite

show abstract

“…Diamond [21] also uses a Rete network to evaluate SPARQL queries on RDF data sets. Also, Peters et al [23] proposed a Rete-based, parallel rule engine, which utilizes GPUs to parallelize transformations on RDF data.…”

Section: Semantic Technologiesmentioning

confidence: 99%

Evaluation of Optimization Strategies for Incremental Graph Queries

Szárnyas

Maginecz²,

Varró

2017

Period. Polytech. Elec. Eng. Comp. Sci.

View full text Add to dashboard Cite

Keywords graph queries, relational algebra, query optimization IntroductionThe key components of Big Data are often defined as variety, velocity and volume [28] of data. Applications operating on continuously changing graphs are a prime example: the semi-structured graph-like nature introduces a high variety, changes happen at high velocity, and datasets are often high-volume. Such applications include fraud detection in financial transactions [27], validation of engineering models [3], and static analysis of source code repositories [35]. These use cases provide a set of complex queries that need to be evaluated continuously on each change of the underlying graph.Traditional approaches need to reevaluate each query upon each change, which often takes minutes on a large dataset. In contrast, incremental query evaluation caches interim results, hence it only requires reevaluation on a small fragment of the dataset impacted by the change. This leads to significant speedup for large and continuously changing data. Although several approaches exist for incremental query evaluation [9,20] in the context of expert systems, incremental query evaluation is not in widespread use in graph databases.In order to predict query performance at runtime, relational databases synthesize and evaluate different query plans which impose a certain ordering on relational algebraic operations prescribed by the query. Optimizing query plans is a challenging task, since a wide variety of query plans may exist even for simple queries with different costs. Database engines use heuristics-based optimization techniques and evaluate a cost function for the different query plans [10].Query plans have been adapted for graph query engines using a local-search based query evaluation strategy where it is called the search plan. Optimization techniques may exploit the type and multiplicity information defined in the graph schema (or metamodel) [29,22] or rely upon runtime statistics of the instance graph [11,38,39].In case of incremental graph query engines, the structure and the content of caches have the most significant impact on query performance. Therefore, optimization is directed to reduce execution time and memory consumption imposed by a complex network of caches [37].

show abstract

“…However, several systems have implemented their own flavor of RETE [18]. Furthermore, RETE has recently been ported to GPU architectures [23].…”

Section: Related Work and Analysismentioning

confidence: 99%

“…However, the price tag of such a supercomputer, i.e., several million dollars, does not fit every wallet. On a more reasonable budget, recent experiments of the RETE algorithm on high-end GPUs [23] report a throughput of 2.7 million triples/sec. for ρdf and 1.4 million triples/sec.…”

Section: Memory Access Patternsmentioning

confidence: 99%

Inferray

et al. 2016

View full text Add to dashboard Cite

The advent of semantic data on the Web requires efficient reasoning systems to infer RDF and OWL data. The linked nature and the huge volume of data entail efficiency and scalability challenges when designing productive inference systems. This paper presents Inferray, an implementation of RDFS, ρDF, and RDFS-Plus inference with improved performance over existing solutions. The main features of Inferray are 1) a storage layout based on vertical partitioning that guarantees sequential access and efficient sort-merge join inference; 2) efficient sorting of pairs of 64-bit integers using ad-hoc optimizations on MSD radix and a custom counting sort; 3) a dedicated temporary storage to perform efficient graph closure computation. Our measurements on synthetic and real-world datasets show improvements over competitors on RDFS-Plus, and up to several orders of magnitude for transitivity closure.

show abstract

Scaling Parallel Rule-Based Reasoning

Cited by 13 publications

References 12 publications

A semantics-based approach to sensor data segmentation in real-time Activity Recognition

A semantics-based approach to sensor data segmentation in real-time Activity Recognition

Evaluation of Optimization Strategies for Incremental Graph Queries

Inferray

Contact Info

Product

Resources

About