Mate Tomoskozi
Research Subjects
Mate Tomoskozi has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Frank H. P. Fitzek
Deutsche Telekom Chair of Communication Networks, Technische Universitt Dresden, Dresden, Germany
5 shared publications - Huanzhuo Wu
Nokia, Munich, Germany
2 shared publications - Martin Reisslein
School of Electrical, Computer, and Energy Engineering, Arizona State University, Tempe, AZ, USA
2 shared publications - Dongho You
Department of Information and Communication Engineering, Hannam University, Daejeon, South Korea
1 shared publication - Elif Tasdemir
Centre for Tactile Internet with Human-in-the-Loop (CeTI), Technische Universitt Dresden, Dresden, Germany
1 shared publication - Frank Gabriel
G Lab Germany, Deutsche Telekom Chair, Technische Universität Dresden, Dresden, Germany
1 shared publication
Latest Publications
- 14 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1217-1222
You Only Hear Once: Lightweight In-Network AI Design for Multi-object Anomaly Detection
With the rapid evolution of programmable network devices, the transformation of network infrastructure to a programmable computing platform has become a reality. As a result, such programmable networks make the deployment of in-network AI possible for achieving predictive maintenance, which in turn enables the much desired (sound-based) time-critical anomaly detection in modern smart factories. However, such abnormal sound detection systems face high computational complexity due to their high density and multi-source nature. In this paper, we present our novel You only hear once (Yoho) design for multi-target anomalous sound detection and end-to-end feature extraction. Yoho can detect anomalies directly from a mixed multi-source acoustic signal without recovering each source data. In addition, with a lightweight dual-path backbone design (Yoho-Dual-Path), we achieve low model complexity and high detection accuracy. Our experiments show that the proposed solution can reduce the inference time by 77% and improve the detection accuracy by 51%. This enables the integration of intelligent anomalous sound detection systems into the network.
- 14 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1217-1222
You Only Hear Once: Lightweight In-Network AI Design for Multi-object Anomaly Detection
With the rapid evolution of programmable network devices, the transformation of network infrastructure to a programmable computing platform has become a reality. As a result, such programmable networks make the deployment of in-network AI possible for achieving predictive maintenance, which in turn enables the much desired (sound-based) time-critical anomaly detection in modern smart factories. However, such abnormal sound detection systems face high computational complexity due to their high density and multi-source nature. In this paper, we present our novel You only hear once (Yoho) design for multi-target anomalous sound detection and end-to-end feature extraction. Yoho can detect anomalies directly from a mixed multi-source acoustic signal without recovering each source data. In addition, with a lightweight dual-path backbone design (Yoho-Dual-Path), we achieve low model complexity and high detection accuracy. Our experiments show that the proposed solution can reduce the inference time by 77% and improve the detection accuracy by 51%. This enables the integration of intelligent anomalous sound detection systems into the network.
- 19 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 24 (1) , 698-740
Packet Header Compression: A Principle-Based Survey of Standards and Recent Research Studies
Many emerging communication applications transmit packets with small payloads, yet require rich metadata context for the proper functioning of the various involved protocol layers. Packet header compression becomes vital to efficiently support these emerging communication applications in resource-constrained environments, such as the Internet of Things (IoT). As a basis for the survey of the existing packet header compression standards and recent packet header compression research studies, we first introduce a novel set of five principles of packet header compression, namely the principles of identification, definition, placement, compression, and control. Based on these five principles, we survey the major header compression standards, spanning from the Thinwire protocol from 1984, via Robust Header Compression (RoHC) version 1 (2001) and version 2 (2008) to Static Context Header Compression (SCHC, 2020) and QPACK currently in draft status. We also comprehensively survey the recent header compression research studies in the past six years (2016–2021), following the introduced five principles as taxonomy. The survey of header compression research includes novel compression concepts and evaluations in various modern environments, such as in wireless mesh networks. We conclude the survey by outlining major open packet header compression research challenges along the five principles.
- 19 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Communications Surveys & Tutorials
- Vol. 24 (1) , 698-740
Packet Header Compression: A Principle-Based Survey of Standards and Recent Research Studies
Many emerging communication applications transmit packets with small payloads, yet require rich metadata context for the proper functioning of the various involved protocol layers. Packet header compression becomes vital to efficiently support these emerging communication applications in resource-constrained environments, such as the Internet of Things (IoT). As a basis for the survey of the existing packet header compression standards and recent packet header compression research studies, we first introduce a novel set of five principles of packet header compression, namely the principles of identification, definition, placement, compression, and control. Based on these five principles, we survey the major header compression standards, spanning from the Thinwire protocol from 1984, via Robust Header Compression (RoHC) version 1 (2001) and version 2 (2008) to Static Context Header Compression (SCHC, 2020) and QPACK currently in draft status. We also comprehensively survey the recent header compression research studies in the past six years (2016–2021), following the introduced five principles as taxonomy. The survey of header compression research includes novel compression concepts and evaluations in various modern environments, such as in wireless mesh networks. We conclude the survey by outlining major open packet header compression research challenges along the five principles.
- 8 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 939-940
Bitteiler: Demonstration of Efficient and Private Massive Industrial IoT Communications
The increasingly difficult challenges in managing big data are inciting for revolutionary and fundamental techniques, as well as for technologies that can handle the complexity and the never-before-seen volume of data. Our goal is to enable high-density industrial networks to transmit and store production-related information efficiently and confidentially while retaining any existing network infrastructure. As the deployment of sensors leads to both the benefits and the challenges of industrial big data, our Bitteiler solution represents a means to reduce the data volume at the early stages of the data cycle to curb its impact on communications and storage systems. Bitteiler is a plug-and-play solution that allows customers to leverage smart IoT solutions to the fullest to achieve an industry 4.0-ready production process that is both efficient and secure. Our practical demonstration provides a hands-on experience with the Bitteiler technology – including in-network compression and coding – using a number of live IoT devices that can be manipulated by the audience.
- 8 January 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 939-940
Bitteiler: Demonstration of Efficient and Private Massive Industrial IoT Communications
The increasingly difficult challenges in managing big data are inciting for revolutionary and fundamental techniques, as well as for technologies that can handle the complexity and the never-before-seen volume of data. Our goal is to enable high-density industrial networks to transmit and store production-related information efficiently and confidentially while retaining any existing network infrastructure. As the deployment of sensors leads to both the benefits and the challenges of industrial big data, our Bitteiler solution represents a means to reduce the data volume at the early stages of the data cycle to curb its impact on communications and storage systems. Bitteiler is a plug-and-play solution that allows customers to leverage smart IoT solutions to the fullest to achieve an industry 4.0-ready production process that is both efficient and secure. Our practical demonstration provides a hands-on experience with the Bitteiler technology – including in-network compression and coding – using a number of live IoT devices that can be manipulated by the audience.
- Open Access
- 21 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 168567-168586
SpaRec: Sparse Systematic RLNC Recoding in Multi-Hop Networks
Sparse Random Linear Network Coding (RLNC) reduces the computational complexity of the RLNC decoding through a low density of the non-zero coding coefficients, which can be achieved through sending uncoded (systematic) packets. However, conventional recoding of sparse RLNC coded packets at an intermediate node in a multi-hop network increases the density of non-zero coding coefficients. We develop and evaluate sparsity-preserving recoding (SpaRec) strategies that preserve the low density of non-zero coding coefficients of sparse RLNC with systematic packets. We develop SpaRec strategies with and without decoding at the intermediate nodes, with and without a specified coding rate, as well as with finite and infinite recoding window lengths. We evaluate the SpaRec strategies in multi-hop networks in terms of packet loss, packet delivery delay, as well as recoding and decoding (computation) throughput. We find that the SpaRec strategies substantially improve the RLNC performance compared to conventional recoding.
- 21 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 168567-168586
Open AccessSpaRec: Sparse Systematic RLNC Recoding in Multi-Hop Networks
Sparse Random Linear Network Coding (RLNC) reduces the computational complexity of the RLNC decoding through a low density of the non-zero coding coefficients, which can be achieved through sending uncoded (systematic) packets. However, conventional recoding of sparse RLNC coded packets at an intermediate node in a multi-hop network increases the density of non-zero coding coefficients. We develop and evaluate sparsity-preserving recoding (SpaRec) strategies that preserve the low density of non-zero coding coefficients of sparse RLNC with systematic packets. We develop SpaRec strategies with and without decoding at the intermediate nodes, with and without a specified coding rate, as well as with finite and infinite recoding window lengths. We evaluate the SpaRec strategies in multi-hop networks in terms of packet loss, packet delivery delay, as well as recoding and decoding (computation) throughput. We find that the SpaRec strategies substantially improve the RLNC performance compared to conventional recoding.
- 1 January 2017
- Published by University of Szeged in Acta Cybernetica
- Vol. 23 (2) , 737-756
Prediction of RoHCv1 and RoHCv2 Compressor Utilities for VoIP
Modern cellular networks utilising the long{term evolution (LTE) and the coming 5G set of standards face an ever{increasing demand for low{latency mobile data from connected devices. Header compression is employed to min- imise the overhead for IP{based cellular network trac, thereby decreasing the overall bandwidth usage and, subsequently, transmission delays. Here, we employ machine learning approaches for the prediction of Robust Header Compression version 1's and version 2's compression utility for VoIP transmis- sions, which allows the compression to dynamically adapt to varying channel conditions. We evaluate various regression models employing r2 and mean square error scores next to complexity (number of coecients) based on an RTP specic training data set and separately captured live VoIP audio calls. We nd that the proposed weighted Ridge regression model explains about at least 50 % of the observed results and the accuracy score may be as high as 94 % for some of the VoIP transmissions.
- 1 January 2017
- Published by University of Szeged in Acta Cybernetica
- Vol. 23 (2) , 737-756
Prediction of RoHCv1 and RoHCv2 Compressor Utilities for VoIP
Modern cellular networks utilising the long{term evolution (LTE) and the coming 5G set of standards face an ever{increasing demand for low{latency mobile data from connected devices. Header compression is employed to min- imise the overhead for IP{based cellular network trac, thereby decreasing the overall bandwidth usage and, subsequently, transmission delays. Here, we employ machine learning approaches for the prediction of Robust Header Compression version 1's and version 2's compression utility for VoIP transmis- sions, which allows the compression to dynamically adapt to varying channel conditions. We evaluate various regression models employing r2 and mean square error scores next to complexity (number of coecients) based on an RTP specic training data set and separately captured live VoIP audio calls. We nd that the proposed weighted Ridge regression model explains about at least 50 % of the observed results and the accuracy score may be as high as 94 % for some of the VoIP transmissions.