Filters








87,680 Hits in 6.8 sec

Multimodal Event Processing: A Neural-Symbolic Paradigm for the Internet of Multimedia Things

Edward Curry, Dhaval Salwala, Praneet Dhingra, Felipe Arruda Pontes, Piyush Yadav
2022 IEEE Internet of Things Journal  
The content of multimodal streams is represented using Multimodal Event Knowledge Graphs to capture the semantic, spatial, and temporal content of the multimodal streams.  ...  seconds for video streams of 30 fps input rate.  ...  Once GNOSIS receives a multimodal stream, it is processed to generate a knowledge graph representation using query-aware edge-cloud processing pipelines.  ... 
doi:10.1109/jiot.2022.3143171 fatcat:3dez6jqpjbfebeamdda7qpu5dy

Efficient Temporal Reasoning on Streams of Events with DOTR [chapter]

Alessandro Margara, Gianpaolo Cugola, Dario Collavini, Daniele Dell'Aglio
2018 Lecture Notes in Computer Science  
The Stream Reasoning (SR) domain aims to combine the performance of stream/event processing and the reasoning expressiveness of knowledge representation systems by adopting Semantic Web standards to represent  ...  We implement the model in the DOTR system that defines ontological reasoning using Datalog rules and temporal reasoning using the TESLA Complex Event Processing language, which builds on metric temporal  ...  We perform all the experiments on a Intel Core i7 4850HQ machine with 16 GB of DDR3 RAM, running macOS 10.13.0. We use the processing time per input element as a performance indicator.  ... 
doi:10.1007/978-3-319-93417-4_25 fatcat:jbg774rbyvb7rmd2say6uie34y

Streaming graph partitioning

Zainab Abbas, Vasiliki Kalavri, Paris Carbone, Vladimir Vlassov
2018 Proceedings of the VLDB Endowment  
Online methods ingest edges or vertices as a stream, making partitioning decisions on the fly based on partial knowledge of the graph.  ...  Furthermore, we employ an experimental comparison across different applications and datasets, using a unified distributed runtime based on Apache Flink.  ...  Our results indicate that the majority of streaming graph partitioning algorithms are unsuitable for continuous processing of unbounded streams due to their reliance on a priori knowledge of graph properties  ... 
doi:10.14778/3236187.3236208 fatcat:ehsvpqdykbc35odmga2rcyvlfq

Low-Resolution Face Recognition in the Wild via Selective Knowledge Distillation

Shiming Ge, Shengwei Zhao, Chenyu Li, Jia Li
2019 IEEE Transactions on Image Processing  
are then used to regularize the fine-tuning process of the student stream.  ...  To avoid significant performance drop at the student stream, we then selectively distil the most informative facial features from the teacher stream by solving a sparse graph optimization problem, which  ...  a sparse graph optimization problem, which are then used to regularize the fine-tuning process of the student stream.  ... 
doi:10.1109/tip.2018.2883743 fatcat:32z2fr6vpzbn3esvl5uwqrgd6e

Zero-Shot Action Recognition with Three-Stream Graph Convolutional Networks

Nan Wu, Kazuhiko Kawamoto
2021 Sensors  
In this paper, in order to solve these problems, we propose a three-stream graph convolutional network that processes both types of data. Our model has two parts.  ...  Large datasets are often used to improve the accuracy of action recognition. However, very large datasets are problematic as, for example, the annotation of large datasets is labor-intensive.  ...  Two-Stream Graph Convolutional Networks Many recent ZSAR models, including TS-GCN, use knowledge graphs. The knowledge graph contains the relationship between words.  ... 
doi:10.3390/s21113793 pmid:34070872 fatcat:bcwm3f6bzzcndjzabjfelqsmh4

Video Based Person Re-Identification Through Selective Knowledge Distillation

Gudavalli Sai Abhilash, Kantheti Rajesh, Jangam Dileep Shaleem, Grandi Sai Sarath, Palli R Krishna Prasad
2019 International Journal of Scientific Research in Computer Science Engineering and Information Technology  
are then used to regularize the fine- tuning process of the student stream.  ...  To avoid significant performance drop at the student stream, we then selectively distil the most informative facial features from the teacher stream by solving a sparse graph optimization problem, which  ...  Chen [6] proposed a framework for learning compact and fast object detection networks with improved accuracy using knowledge distillation.  ... 
doi:10.32628/cseit1952179 fatcat:o75zy4c6vzdh3cxjcjrjmjwxeu

Skeletal Human Action Recognition using Hybrid Attention based Graph Convolutional Network [article]

Hao Xing, Darius Burschka
2022 arXiv   pre-print
Moreover, we design a new initial graph adjacency matrix that connects head, hands and feet, which shows visible improvement in terms of action recognition accuracy.  ...  The results demonstrate that our model has strong performance on both dataset.  ...  Compared to the HA-GCN (single T), the full HA-GCN model using multi-scale temporal layer has significant performance improvement for both input streams.  ... 
arXiv:2207.05493v1 fatcat:z4zgjodsvrfavadhfbhplcceju

Recursive Multi-Section on the Fly: Shared-Memory Streaming Algorithms for Hierarchical Graph Partitioning and Process Mapping [article]

Marcelo Fonseca Faraj, Christian Schulz
2022 arXiv   pre-print
In this work, we present a shared-memory streaming multi-recursive partitioning scheme that performs recursive multi-sections on the fly without knowing the overall input graph.  ...  A current trend for partitioning huge graphs are streaming algorithms, which use low computational resources.  ...  Thereby our algorithms obtains good process mappings using single pass through the input graph.  ... 
arXiv:2202.00394v1 fatcat:z3xresbvrvcopawubrexydn7ui

Towards Scalable Non-Monotonic Stream Reasoning via Input Dependency Analysis

Thu-Le Pham, Alessandra Mileo, Muhammad Intizar Ali
2017 2017 IEEE 33rd International Conference on Data Engineering (ICDE)  
The input dependency graph allows us to dynamically configure the streaming window size in order to maximise the scalability of the non-monotonic reasoner.  ...  We introduce an input dependency graph to represent the relationships between input events based on the structure of a given logical rule set.  ...  We compare their performance with the performance of processing the entire input window W .  ... 
doi:10.1109/icde.2017.226 dblp:conf/icde/PhamMA17 fatcat:vy5l6qbcpjbuxprelyv42zc6eu

Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network

Junpeng Kang, Jing Zhang, Wensheng Li, Li Zhuo
2021 IET Image Processing  
The existing crowd activity recognition mainly uses visual information, rarely fully exploiting and utilizing the correlation or external knowledge between crowd content.  ...  as edges. (3) Crowd activity in live video streaming is recognised by combining the output of ReGCN, the deep spatiotemporal features and the crowd motion intensity as external knowledge.  ...  Compared with using ReGCN alone, the performance improvement is not obvious.  ... 
doi:10.1049/ipr2.12239 fatcat:yzmmqohxynhonaxbeqbbuxvqpm

NOUS: Construction and Querying of Dynamic Knowledge Graphs

Sutanay Choudhury, Khushbu Agarwal, Sumit Purohit, Baichuan Zhang, Meg Pirrung, Will Smith, Mathew Thomas
2017 2017 IEEE 33rd International Conference on Data Engineering (ICDE)  
The ability to construct domain specific knowledge graphs (KG) and perform question-answering or hypothesis generation is a transformative capability.  ...  We propose an end-to-end framework for developing custom knowledge graph driven analytics for arbitrary application domains.  ...  We view the Knowledge Graph construction as an incremental process and develop a family of algorithms designed for dynamic graphs. 2.  ... 
doi:10.1109/icde.2017.228 dblp:conf/icde/ChoudhuryAPZPST17 fatcat:g7kzogldv5ahdig7i7u5epmgda

NOUS: Construction and Querying of Dynamic Knowledge Graphs [article]

Sutanay Choudhury and Khushbu Agarwal and Sumit Purohit and Baichuan Zhang and Meg Pirrung and Will Smith and Mathew Thomas
2016 arXiv   pre-print
The ability to construct domain specific knowledge graphs (KG) and perform question-answering or hypothesis generation is a transformative capability.  ...  We propose an end-to-end framework for developing custom knowledge graph driven analytics for arbitrary application domains.  ...  We view the Knowledge Graph construction as an incremental process and develop a family of algorithms designed for dynamic graphs. 2.  ... 
arXiv:1606.02314v1 fatcat:vgm6fiorafbpfp56uafxyyj2fe

Efficient RDF stream reasoning with graphics processingunits (GPUs)

Chang Liu, Jacopo Urbani, Guilin Qi
2014 Proceedings of the 23rd International Conference on World Wide Web - WWW '14 Companion  
In this paper, we study the problem of stream reasoning and propose a reasoning approach over large amounts of RDF data, which uses graphics processing units (GPU) to improve the performance.  ...  First, we show how the problem of stream reasoning can be reduced to a temporal reasoning problem. Then, we describe a number of algorithms to perform stream reasoning with GPUs.  ...  In a typical stream reasoning scenario, we are required to perform inference on a RDF graph composed by the background knowledge base and all RDF triples in a window of a RDF stream over time .  ... 
doi:10.1145/2567948.2577323 dblp:conf/www/LiuUQ14 fatcat:4sq6re3ilfcwbgirsydidrc32u

OPTIMIZED FEATURE SELECTION BASED PREDICTIVE ROUND ROBIN SCHEDULING (OFS-PRRS)

N. Arunadevi
2017 International Journal of Advanced Research in Computer Science  
LASSO function in big data analytics is used based on assumption of linear dependency between input features and output value.  ...  Optimized Feature Selection based Predictive Round Robin Scheduling (OFS-PRRS) Technique for stream data in big data analytics with higher prediction accuracy and lesser scheduling time.  ...  Predictive resource scheduling is used to improve the performance by leveraging big data analytics.  ... 
doi:10.26483/ijarcs.v8i8.4773 fatcat:r2glkqwhkngsfd2r4sqnybqqq4

CPU load shedding for binary stream joins

Bugra Gedik, Kun-Lung Wu, Philip S. Yu, Ling Liu
2006 Knowledge and Information Systems  
In contrast to the conventional approach of dropping tuples from the input streams, we explore the concept of selective processing for load shedding.  ...  We support such dynamic selective processing through three forms of runtime adaptations: adaptation to input stream rates, adaptation to time correlation between the streams and adaptation to join directions  ...  Table 1 summarizes the notations used throughout the paper. A windowed stream join is performed by fetching tuples from the input streams and processing them against tuples in the opposite window.  ... 
doi:10.1007/s10115-006-0044-4 fatcat:j77kfgtklndmtlizkfqkl7cgwq
« Previous Showing results 1 — 15 out of 87,680 results