Multimodal Extreme Scale Data Analytics for Smart Cities Environments

Informazioni relative al progetto

MARVEL

ID dell’accordo di sovvenzione: 957337

DOI 10.3030/957337

Progetto chiuso

Data della firma CE 22 Luglio 2020

Data di avvio 1 Gennaio 2021

Data di completamento 31 Dicembre 2023

Finanziato da

• INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT)

Costo totale € 5 998 086,25

Contributo UE € 5 998 086,25

Coordinato da

IDRYMA TECHNOLOGIAS KAI EREVNAS
Greece

Risultati finali

Demonstrators, pilots, prototypes (18)

MARVEL Integrated framework - final version

This deliverable is the final version of the MARVEL Integrated framework.

Security assurance and acceleration in E2F2C framework - initial version

This deliverable is the initial version of the security assurance and acceleration in E2F2C framework.

Efficient deployment of AI-optimised ML/DL models - final version

This deliverable is the final version of the efficient deployment of AI-optimised ML/DL models.

MARVEL Minimum Viable Product

This deliverable will report on the MARVEL Minimum Viable Product.

Multimodal and privacy-aware audio-visual intelligence - final version

This deliverable is the final version of the multimodal and privacy-aware audio-visual intelligence.

Management and distribution Toolkit - initial version

This deliverable is the initial version of the management and distribution Toolkit.

Optimal audio-visual capturing, analysis and voice anonymization

This deliverable is the final version of the optimal audio-visual capturing, analysis and voice anonymization.

MARVEL's decision-making toolkit - final version

This deliverable is the final version of the MARVEL’s decision-making toolkit.

Multimodal and privacy-aware audio-visual intelligence - initial version

This deliverable is the initial version of the multimodal and privacy-aware audio-visual intelligence.

Management and distribution Toolkit - final version

This deliverable is the final version of the management and distribution Toolkit.

Demonstrators execution - final version

This deliverable is the final version of the demonstrators execution.

Project website

This deliverable will report on the MARVEL website.

Demonstrators execution - initial version

This deliverable is the initial version of the demonstrators execution.

E2F2C Privacy preservation mechanisms

This deliverable will report on the E2F2C Privacy preservation mechanisms.

Efficient deployment of AI-optimised ML/DL models - initial version

This deliverable is the initial version of the efficient deployment of AI-optimised ML/DL models.

Security assurance and acceleration in E2F2C framework - final version

This deliverable is the final version of the security assurance and acceleration in E2F2C framework.

Optimal audio-visual capturing, analysis and voice anonymization - initial version

This deliverable is the initial version of the optimal audio-visual capturing, analysis and voice anonymization.

MARVEL's decision-making toolkit - initial version

This deliverable is the initial version of the MARVEL’s decision-making toolkit.

Documents, reports (10)

Market analysis and preliminary business modelling

This deliverable will report on the market analysis and preliminary business modelling

Technical evaluation and progress against benchmarks - initial version

This deliverable is the initial version of the technical evaluation and progress against benchmarks

Final assessment report and impact analysis

This deliverable will contain the final assessment report and impact analysis for MARVEL.

HPC infrastructure and resource management for audio-visual data analytics - initial version

This deliverable is the initial version of the HPC infrastructure and resource management for audiovisual data analytics

Evaluation report

This deliverable is the report on the MARVEL Evaluation.

MARVEL's framework large scale deployment

This deliverable will report on the MARVEL’s framework large scale deployment.

Technical evaluation and progress against benchmarks

This deliverable is the final version of the technical evaluation and progress against benchmarks.

HPC infrastructure and resource management for audio-visual data analytics - final version

This deliverable is the final version of the HPC infrastructure and resource management for audio-visual data analytics.

Architecture definition for MARVEL framework

This deliverable will document the architecture of MARVELs framework

Collection and analysis of experimental data

This deliverable will document the collection and analysis of experimental data

Pubblicazioni

Peer reviewed articles (25)

Single-Layer Transformers for More Accurate Early Exits with Less Overhead

Autori: Arian Bakhtiarnia; Qi Zhang; Alexandros Iosifidis
Pubblicato in: Neural Networks, Numero Volume 153, September 2022, Pages 461-473, 2022, Pagina/e 461-473, ISSN 0893-6080
Editore: Pergamon Press Ltd.
DOI: 10.1016/j.neunet.2022.06.038

Enriched Music Representations with Multiple Cross-modal Contrastive Learning

Autori: Andres Ferraro; Xavier Favory; Konstantinos Drossos; Yuntae Kim; Dmitry Bogdanov
Pubblicato in: IEEE Signal Processing Letters, Numero Volume 28, 05 April 2021, 2021, Pagina/e 733-737, ISSN 1070-9908
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/LSP.2021.3071082

Enhancing Emotion Recognition through Federated Learning: A Multimodal Approach with Convolutional Neural Networks

Autori: Simic, Nikola; Suzic, Sinisa; Miloševic, Nemanja; Stanojev, Vuk; Nosek, Tijana; Popovic, Branislav; Bajovic, Dragana
Pubblicato in: Applied Sciences, Numero Volume 14, Numero 4, 2024, 2024, Pagina/e 1-14, ISSN 1539-9087
Editore: Association for Computing Machinary, Inc.
DOI: 10.3390/app14041325

Federated Feature Selection for Cyber-Physical Systems of Systems

Autori: Cassará, Pietro; Gotta, Alberto; Valerio, Lorenzo
Pubblicato in: IEEE Transactions on Vehicular Technology, Numero to appear, 2022, Pagina/e 1 - 1, ISSN 0018-9545
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tvt.2022.3178612

FaaS execution models for edge applications

Autori: Claudio Cicconetti; Marco Conti; Andrea Passarella
Pubblicato in: Pervasive and mobile computing (Print) 86 (2022). doi:10.1016/j.pmcj.2022.101689, Numero Volume 86, October 2022, 2022, Pagina/e 1-15, ISSN 1574-1192
Editore: Elsevier BV
DOI: 10.1016/j.pmcj.2022.101689

Nonlinear consensus+innovations under correlated heavy-tailed noises: Mean square convergence rate and asymptotics

Autori: Vukovic, Manojlo; Jakovetic, Dusan; Bajovic, Dragana; Kar, Soummya
Pubblicato in: SIAM Journal on Control and Optimization, Numero Vol. 62, Iss. 1 (2024), 2024, Pagina/e 376-399, ISSN 1095-7138
Editore: SIAM
DOI: 10.1137/22m1543197

A 2-Phase Strategy for Intelligent Cloud Operations

Autori: Giacomo Lanciano; Remo Andreoli; Tommaso Cucinotta; Davide Bacciu University of Pisa, Italy ; Andrea Passarella
Pubblicato in: IEEE Access, Numero vol. 11, 2023, Pagina/e 96841-96853, ISSN 2169-3536
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/access.2023.3312218

Efficient High-Resolution Deep Learning: A Survey

Autori: Bakhtiarnia, Arian; Zhang, Qi; Iosifidis, Alexandros
Pubblicato in: ACM Computing Surveys journal, Numero 31, 2024, Pagina/e 1-35, ISSN 0360-0300
Editore: Association for Computing Machinary, Inc.
DOI: 10.1145/3645107

Stateful Function as a Service at the Edge

Autori: Carlo Puliafito; Claudio Cicconetti; Marco Conti; Enzo Mingozzi; Andrea Passarella
Pubblicato in: Computer, Numero Volume: 55, Numero: 9, September 2022, 2022, Pagina/e 54-64, ISSN 1558-0814
Editore: IEEE
DOI: 10.1109/mc.2021.3138690

In-Network Computing With Function as a Service at the Edge

Autori: Claudio Cicconetti; Marco Conti; Andrea Passarella
Pubblicato in: Computer, Numero 30, 2022, Pagina/e Volume: 55, Numero: 9, September 2022, ISSN 1558-0814
Editore: IEEE
DOI: 10.1109/mc.2021.3130659

Design Recommendations for a Collaborative Game of Bird Call Recognition Based on Internet of Sound Practices

Autori: Emmanuel Rovithis; Nikolaos Moustakas; Konstantinos Vogklis; Konstantinos Drossos; Andreas Floros
Pubblicato in: Journal of the Audio Engineering Society, Numero Volume 69 Numero 12 pp. 956-966; December 2021, 2021, Pagina/e 956-966, ISSN 1549-4950
Editore: Audio Engineering Society
DOI: 10.17743/jaes.2021.0043

A One-Shot Framework for Distributed Clustered Learning in Heterogeneous Environments

Autori: Aleksandar Armacki; Dragana Bajović; Dušan Jakovetić; Soummya Kar
Pubblicato in: IEEE Transactions on Signal Processing, Numero Volume: 72, 25 December 2023, 2023, Pagina/e 636-651, ISSN 1053-587X
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tsp.2023.3343561

Feedforward neural networks initialization based on discriminant learning

Autori: Kateryna Chumachenko; Alexandros Iosifidis; Moncef Gabbouj
Pubblicato in: Neural Networks, Numero Volume 146, February 2022, 2022, Pagina/e 220-229, ISSN 0893-6080
Editore: Pergamon Press Ltd.
DOI: 10.1016/j.neunet.2021.11.020

XimSwap: Many-to-Many Face Swapping for TinyML

Autori: Ancilotto, Alberto; Paissan, Francesco; Farella, Elisabetta
Pubblicato in: ACM Transactions on Embedded Computing Systems, Numero May 2023, 2023, Pagina/e 1-16, ISSN 1539-9087
Editore: Association for Computing Machinary, Inc.
DOI: 10.1145/3603173

Sub-mW Keyword Spotting on an MCU: Analog Binary Feature Extraction and Binary Neural Networks

Autori: Gianmarco Cerutti; Lukas Cavigelli; Renzo Andri; Michele Magno; Elisabetta Farella; Luca Benini
Pubblicato in: IEEE Transactions on Circuits and Systems I: Regular Papers, Numero Volume: 69, Numero: 5, May 2022, 2022, Pagina/e 2002-2012, ISSN 1549-8328
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.5281/zenodo.5960706

Dynamic hard pruning of Neural Networks at the edge of the internet

Autori: Lorenzo Valerio, Franco Maria Nardini, Andrea Passarella, Raffaele Perego
Pubblicato in: Journal of Network and Computer Applications, Numero Volume 200, April 2022, 2022, Pagina/e 1-11, ISSN 1084-8045
Editore: Academic Press
DOI: 10.1016/j.jnca.2021.103330

A Survey on Networked Data Streaming With Apache Kafka

Autori: Theofanis P. Raptis, Andrea Passarella
Pubblicato in: IEEE Access, Numero Vol. 11, 2023, Pagina/e 85333 – 85350, ISSN 2169-3536
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/access.2023.3303810

Nonlinear Gradient Mappings and Stochastic Optimization: A General Framework with Applications to Heavy-Tail Noise

Autori: Dus̆an Jakovetić; Dragana Bajović; Anit Kumar Sahu; Soummya Kar; Nemanja Milos̆ević; Dus̆an Stamenković
Pubblicato in: SIAM Journal on Optimization, Numero 15, 2023, Pagina/e Vol. 33, Iss. 2, ISSN 1095-7189
Editore: SIAM
DOI: 10.1137/21m145896x

Distributed Recursive Estimation under Heavy-Tail Communication Noise

Autori: Dusan Jakovetic; Manojlo Vukovic; Dragana Bajovic; Anit Kumar Sahu; Soummya Kar
Pubblicato in: SIAM, Numero 41, 2023, Pagina/e Vol. 61 (3), ISSN 1095-7138
Editore: SIAM
DOI: 10.1137/22m1477015

Efficient topic partitioning of Apache Kafka for high-reliability real-time data streaming applications

Autori: Theofanis P. Raptis, Claudio Cicconetti, Andrea Passarella
Pubblicato in: Future Generation Computer Systems, Numero Volume 154, May 2024, 2024, Pagina/e 173-188, ISSN 0167-739X
Editore: Elsevier BV
DOI: 10.1016/j.future.2023.12.028

PhiNets: a scalable backbone for low-power AI at the edge

Autori: Francesco Paissan; Alberto Ancilotto; Elisabetta Farella
Pubblicato in: ACM Transactions on Embedded Computing Systems, 2022, Pagina/e 1-19, ISSN 1539-9087
Editore: Association for Computing Machinary, Inc.
DOI: 10.1145/3510832

Automatic social distance estimation for photographic studies: Performance evaluation, test benchmark, and algorithm

Autori: Mert Seker; Anssi Männistö; Alexandros Iosifidis; Jenni Raitoharju
Pubblicato in: Machine Learning with Applications, Numero vol. 10, number 100427, 2022, ISSN 2666-8270
Editore: Elsevier
DOI: 10.1016/j.mlwa.2022.100427

Engineering Resource-Efficient Data Management for Smart Cities with Apache Kafka

Autori: Theofanis P. Raptis; Claudio Cicconetti; Manolis Falelakis; Grigorios Kalogiannis; Tassos Kanellos; Tomás Pariente Lobo
Pubblicato in: Future Internet, Numero Volume 15, Numero 2, 2023, Pagina/e 1-22, ISSN 1999-5903
Editore: MDPI
DOI: 10.3390/fi15020043

Large Deviations for Products of Non-Identically Distributed Network Matrices With Applications to Communication-Efficient Distributed Learning and Inference

Autori: Petrovic, Nemanja; Bajovic, Dragana; Kar, Soummya; Jakovetic, Dušan; Sahu, Anit Kumar
Pubblicato in: IEEE Transactions on Signal Processing, Numero Vol. 71, 2023, 2023, Pagina/e 1319 - 1333, ISSN 1941-0476
Editore: IEEE
DOI: 10.1109/tsp.2023.3263254

Inaccuracy rates for distributed inference over random networks with applications to social learning

Autori: Dragana Bajovic
Pubblicato in: IEEE Transactions on Information Theory, Numero 28, 2023, Pagina/e Volume: 70, Numero: 1, ISSN 1557-9654
Editore: IEEE
DOI: 10.1109/tit.2023.3324866

Conference proceedings (42)

Analysis of the Effect of Low-Overhead Lossy Image Compression on the Performance of Visual Crowd Counting for Smart City Applications

Autori: Bakhtiarnia, Arian; Leporowski, Błażej; Esterle, Lukas; Iosifidis, Alexandros
Pubblicato in: Bakhtiarnia , A , Leporowski , B , Esterle , L & Iosifidis , A 2022 , Analysis of the Effect of Low-Overhead Lossy Image Compression on the Performance of Visual Crowd Counting for Smart City Applications . in 2022 IEEE International Smart Cities Conference (ISC2) . IEEE , 8th IEEE International Smart Cities Conference, ISC2 2022 , Pafos , Cyprus , 26/09/2022 . https://doi.org/10.1109/ISC255366.20, Numero 26-29 September 2022, 2022, Pagina/e 1-5, ISSN 2687-8860
Editore: IEEE
DOI: 10.1109/isc255366.2022.9921939

ConfMix: Unsupervised Domain Adaptation for Object Detection via Confidence-based Mixing

Autori: Giulio Mattolin; Luca Zanella; Elisa Ricci; Yiming Wang
Pubblicato in: IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023, Numero 03-07/01/2023, 2023, Pagina/e 423-433
Editore: IEEE
DOI: 10.1109/wacv56688.2023.00050

Cluster-level pseudo-labelling for source-free cross-domain facial expression recognition

Autori: Alessandro Conti; Paolo Rota; Yiming Wang; Elisa Ricci
Pubblicato in: The 33rd British Machine Vision Conference (BMVC 2022), Numero 21 – 24/11/2022, 2022, Pagina/e 1-13
Editore: BMVC
DOI: 10.5281/zenodo.7296310

Scalable neural architectures for end-to-end environmental sound classification

Autori: Francesco Paissan; Alberto Ancilotto; Alessio Brutti; Elisabetta Farella
Pubblicato in: ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing, 2022, Pagina/e 641-645, ISSN 1520-6149
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/icassp43922.2022.9746093

MARVEL: Multimodal Extreme Scale Data Analytics for Smart Cities Environments

Autori: Dragana Bajovic, Arian Bakhtiarnia, George Bravos, Alessio Brutti, Felix Burkhardt, Daniel Cauchi, Antony Chazapis, Claire Cianco, Nicola Dall'Asen, Vlado Delic, Christos Dimou, Djordje Djokic, Antonio Escobar-Molero, Lukas Esterle, Florian Eyben, Elisabetta Farella, Thomas Festi, Artemios Geromitsos, Giannis Giakoumakis, George Hatzivasilis, Sotiris Ioannidis, Alexandros Iosifidis, Theodora Kalli
Pubblicato in: BalkanCom 2021 - Fourth International Balkan Conference on Communications and Networking, 2022
Editore: IEEE
DOI: 10.5281/zenodo.5547217

Low-complexity acoustic scene classification in DCASE 2022 Challenge

Autori: Irene Martín Morató; Francesco Paissan; Alberto Ancilotto; Toni Heittola; Annamaria Mesaros; Elisabetta Farella; Alessio Brutti; Tuomas Oskari Virtanen
Pubblicato in: 2022 Challenge. 7th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE 2022), Numero 3 November 2022, 2022, Pagina/e 1-5
Editore: DCASE
DOI: 10.48550/arxiv.2206.03835

Learning distinctive features helps, provably

Autori: Firas Laakom, Jenni Raitoharju, Alexandros Iosifidis and Moncef Gabbouj
Pubblicato in: European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2023), Numero September 18–22, 2023, 2023, Pagina/e 206-222
Editore: Springer
DOI: 10.1007/978-3-031-43415-0_13

Dynamic Split Computing for Efficient Deep Edge Intelligence

Autori: Arian Bakhtiarnia; Nemanja Milosevic; Qi Zhang; Dragana Bajovic; Alexandros Iosifidis
Pubblicato in: ICML 2022 - The Thirty-ninth International Conference on Machine Learning - Workshop on Dynamic Neural Networks, 2022
Editore: PMLR
DOI: 10.5281/zenodo.6737212

Personalized Federated Learning via Convex Clustering

Autori: Aleksandar Armacki; Dragana Bajovic; Dusan Jakovetic; Soummya Kar
Pubblicato in: 8th IEEE International Smart Cities Conference 2022, MARV workshop, Numero 26-29 September 2022, 2022, Pagina/e 1-7, ISBN 978-1-6654-8561-6
Editore: IEEE
DOI: 10.1109/isc255366.2022.9921863

Dynamic Split Computing for Efficient Deep EDGE Intelligence

Autori: Arian Bakhtiarnia; Nemanja Milošević; Qi Zhang; Dragana Bajović; Alexandros Iosifidis
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, Numero 04-10 June 2023, 2023, Pagina/e 1-5, ISSN 2379-190X
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10096914

Enabling energy efficient machine learning on a Ultra-Low-Power vision sensor for IoT

Autori: Paissan, Francesco; Gottardi, Massimo; Farella, Elisabetta
Pubblicato in: SLOHA 2021 – DATE Friday Workshop on System-level Design Methods for Deep Learning on Heterogeneous Architectures, Numero SLOHA/2021/01, 2021
Editore: SLOHA 2021
DOI: 10.48550/arxiv.2102.01340

Optimizing PhiNet architectures for the detection of urban sounds on low-end devices

Autori: Alessio Brutti; Francesco Paissan; Alberto Ancilotto; Elisabetta Farella
Pubblicato in: EUSIPCO 2022 - 30th European Signal Processing Conference, 2022
Editore: EURASIP
DOI: 10.5281/zenodo.6641845

PhiNet-GAN: Bringing real-time face swapping to embedded devices

Autori: Ancilotto, Alberto; Paissan, Francesco; Farella, Elisabetta
Pubblicato in: 2023 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops), Numero 13-17 March 2023, 2023, Pagina/e 677-682
Editore: IEEE
DOI: 10.1109/percomworkshops56833.2023.10150292

On Efficiently Partitioning a Topic in Apache Kafka

Autori: Theofanis P. Raptis; Andrea Passarella
Pubblicato in: CITS20222 - 2022 International Conference on Computer, Information, and Telecommunication Systems, 2022
Editore: IEEE
DOI: 10.5281/zenodo.6659879

Continuous Security Assurance of Modern Supply- Chain Ecosystems with Application in Autonomous Driving

Autori: George Hatzivasilis, Sotiris Ioannidis, Grigoris Kalogiannis, Manolis Chatzimpyrros, George Spanoudakis, Guillermo Jiménez Prieto, Araceli Rojas Morgan, Miguel Juaniz Lopez, Cataldo Basile, Jose Francisco Ruiz
Pubblicato in: IEEE CSR Workshop on Cyber Resilience and Economics (CRE2023), Numero 31 July 2023 - 02 August 2023, 2023, Pagina/e 464-469, ISBN 979-8-3503-1170-9
Editore: IEEE
DOI: 10.1109/csr57506.2023.10224971

Design Guidelines for Apache Kafka Driven Data Management and Distribution in Smart Cities

Autori: Theofanis P. Raptis; Claudio Cicconetti; Manolis Falelakis; Tassos Kanellos; Tomas Pariente Lobo
Pubblicato in: 2022 IEEE International Smart Cities Conference (ISC2), Numero 26-29 September 2022, 2022, Pagina/e 1-7
Editore: IEEE
DOI: 10.1109/isc255366.2022.9922546

Crowd Counting on Heavily Compressed Images with Curriculum Pre-Training

Autori: Arian Bakhtiarnia; Qi Zhang; Alexandros Iosifidis
Pubblicato in: 2023 IEEE Symposium Series on Computational Intelligence (SSCI 2023), Numero 05-08 December 2023, 2023, Pagina/e 559-564
Editore: IEEE
DOI: 10.1109/ssci52147.2023.10371805

Masking Speech Contents by Random Splicing: is Emotional Expression Preserved?

Autori: Felix Burkhardt; Anna Derington; Matthias Kahlau; Klaus Scherer; Florian Eyben; Björn Schuller
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing, Numero 04-10 June 2023, 2023, Pagina/e 1-5
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10097094

Accurate Gigapixel Crowd Counting by Iterative Zooming and Refinement

Autori: Bakhtiarnia, Arian; Zhang, Qi; Iosifidis, Alexandros
Pubblicato in: IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2024), Numero 14-19/04/2024, 2024
Editore: IEEE
DOI: 10.48550/arxiv.2305.09271

Gradient Based Clustering

Autori: Aleksandar Armacki; Dragana Bajovic; Dusan Jakovetic; Soummya Kar
Pubblicato in: ICML 2022 - 2022 International Conference on Machine Learning, 2022
Editore: PMLR
DOI: 10.5281/zenodo.7007266

Speech-based Age and Gender Prediction with Transformers

Autori: Burkhardt Felix, Wagner Johannes, Wierstorf Hagen, Eyben Florian, Schuller Björn
Pubblicato in: Speech Communication; 15th ITG Conference, Numero 20-22 September 2023, 2023, ISBN 978-3-8007-6164-7
Editore: VDE
DOI: 10.30420/456164008

XiNet: Efficient Neural Networks for tinyML

Autori: Ancilotto, Alberto; Francesco, Paissan; Farella, Elisabetta
Pubblicato in: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, Numero Oct 2, 2023 – Oct 6, 2023, 2023, Pagina/e 16968-16977
Editore: CVF
DOI: 10.5281/zenodo.10629941

Are We Missing Confidence in Pseudo-LiDAR Methods for Monocular 3D Object Detection?

Autori: Andrea Simonelli, Samuel Rota Bulò, Lorenzo Porzi, Peter Kontschieder, Elisa Ricci
Pubblicato in: ICCV2021 - International Conference on Computer Vision 2021, 2021
Editore: IEEE/CVF
DOI: 10.5281/zenodo.5795266

Improving latency performance trade-off in keyword spotting applications at the edge

Autori: Paissan, Francesco; Sahabdeen, Anisha Mohamed; Alberto, Ancilotto; Farella, Elisabetta
Pubblicato in: 2023 9th International Workshop on Advances in Sensors and Interfaces (IWASI), Numero 08-09 June 2023, 2023, Pagina/e 299-304, ISSN 2836-7936
Editore: IEEE
DOI: 10.1109/iwasi58316.2023.10164328

Continual Learning for Automated Audio Captioning Using The Learning Without Forgetting Approach

Autori: Jan Berg; Konstantinos Drossos
Pubblicato in: DCASE2021 - Detection and Classification of Acoustic Scenes and Events 2021 Workshop, 2021, Pagina/e 140-144, ISBN 978-84-09-36072-7
Editore: UPF
DOI: 10.5281/zenodo.5723053

The Best of Many Worlds: Scheduling Machine Learning Inference on CPU-GPU Integrated Architectures

Autori: Rafail Tsirbas; Giorgos Vasiliadis; Sotiris Ioannidis
Pubblicato in: IPDPSW2022 - 2022 IEEE International Parallel and Distributed Processing Symposium Workshops, 2022, Pagina/e 55-64, ISBN 978-1-6654-9747-3
Editore: IEEE
DOI: 10.1109/ipdpsw55747.2022.00017

Fairness and underspecification in acoustic scene classification: The case for disaggregated evaluations

Autori: Andreas Triantafyllopoulos; Manuel Milling; Konstantinos Drossos; Björn W. Schuller
Pubblicato in: DCASE 2021 - Detection and Classification of Acoustic Scenes and Events 2021, 2021, Pagina/e 70-74, ISBN 978-84-09-36072-7
Editore: UPF
DOI: 10.5281/zenodo.5552489

WaveTransformer: An Architecture for Audio Captioning Based on Learning Temporal and Time-Frequency Information

Autori: An Tran; Konstantinos Drossos; Tuomas Virtanen
Pubblicato in: EUSIPCO 2021 - 29th European Signal Processing Conference, Numero 08 December 2021, 2021, Pagina/e 576-580, ISBN 978-1-6654-0900-1
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.23919/eusipco54536.2021.9616340

On the Role of Smart Vision Sensors in Energy-Efficient Computer Vision at the Edge

Autori: Alberto Ancilotto; Francesco Paissan; Elisabetta Farella
Pubblicato in: PerCom Workshops2022 - 2022 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events, 2022, ISBN 978-1-6654-1647-4
Editore: IEEE
DOI: 10.1109/percomworkshops53856.2022.9767380

Multi-Exit Vision Transformer for Dynamic Inference

Autori: Bakhtiarnia, Arian; Zhang, Qi; Iosifidis, Alexandros
Pubblicato in: BMVC 2021 - 32nd British Machine Vision Conference, 2021, Pagina/e 1-1
Editore: BMVC
DOI: 10.5281/zenodo.5604479

Responsible AI at the edge: towards privacy-preserving smart cities

Autori: Luca Zanella, Yiming Wang, Nicola Dall’Asen, Alberto Ancilotto, Francesco Paissan, Elisa Ricci, Elisabetta Farella, Alessio Brutti and Marco Pistore
Pubblicato in: Ital-IA 2022 Convegno del Laboratorio nazionale CINI-AIIS, 2021
Editore: N/A
DOI: 10.5281/zenodo.5960431

Stateless or Stateful FaaS? I'll Take Both!

Autori: Carlo Puliafito; Claudio Cicconetti; Marco Conti; Enzo Mingozzi; Andrea Passarella
Pubblicato in: 2022 IEEE International Conference on Smart Computing (SMARTCOMP), Numero 20-24 June 2022, 2022, Pagina/e 62-69, ISSN 2693-8340
Editore: IEEE
DOI: 10.1109/smartcomp55677.2022.00024

Communication Efficient Model-Aware Federated Learning for Visual Crowd Counting and Density Estimation in Smart Cities

Autori: Armacki, Aleksandar; Milosevic, Nemanja; Bajovic, Dragana; Kar, Soummya; Jakovetic, Dusan; Bakhtiarnia, Arian; Esterle, Lukas; Muscat, Adrian; Festi, Thomas
Pubblicato in: 2023 31st European Signal Processing Conference (EUSIPCO), Numero 04-08 September 2023, 2023, Pagina/e 875-879
Editore: IEEE
DOI: 10.23919/eusipco58844.2023.10289938

Improving the Accuracy of Early Exits in Multi-Exit Architectures via Curriculum Learning

Autori: Arian Bakhtiarnia; Qi Zhang; Alexandros Iosifidis
Pubblicato in: IJCNN 2021 - International Joint Conference on Neural Networks 2021, 2021, Pagina/e 1-8, ISBN 978-1-6654-4597-9
Editore: IEEE
DOI: 10.1109/ijcnn52387.2021.9533875

PromptMix: Text-to-image diffusion models enhance the performance of lightweight networks

Autori: Arian Bakhtiarnia; Qi Zhang; Alexandros Iosifidis
Pubblicato in: 2023 International Joint Conference on Neural Networks (IJCNN), Numero 18-23 June 2023, 2023, Pagina/e 1-8, ISBN 978-1-6654-8867-9
Editore: IEEE
DOI: 10.1109/ijcnn54540.2023.10191962

Weighted Edit Distance for Country Code Recognition in License Plates

Autori: Kateryna Chumachenko; Alexandros Iosifidis; Moncef Gabbouj
Pubblicato in: EUSIPCO 2022 - 30th European Signal Processing Conference, 2022
Editore: EURASIP
DOI: 10.5281/zenodo.6736842

Automatic analysis of the emotional content of speech in daylong child-centered recordings from a neonatal intensive care unit

Autori: Einari Vaaras; Sari Ahlqvist-Björkroth; Konstantinos Drossos; Okko Räsänen
Pubblicato in: INTERSPEECH 2021, 2021
Editore: ISCA
DOI: 10.21437/interspeech.2021-303

Large deviations rates for stochastic gradient descent with strongly convex functions

Autori: Dragana Bajovic; Dusan Jakovetic; Soummya Kar
Pubblicato in: Proceedings of The 26th International Conference on Artificial Intelligence and Statistics, Numero 25-27 Apr, 2023, 2023, Pagina/e 10095-10111
Editore: PMLR
DOI: 10.5281/zenodo.7537625

Within-layer Diversity Reduces Generalization Gap

Autori: Firas Laakom, Jenni Raitoharju, Alexandros Iosifidis and Moncef Gabbouj
Pubblicato in: ICML2021 - Workshop on Information-Theoretic Methods for Rigorous, Responsible, and Reliable Machine Learning, 2021
Editore: N/A
DOI: 10.5281/zenodo.5772760

Assessment of Self-Attention on Learned Features For Sound Event Localization and Detection

Autori: Parthasaarathy Sudarsanam; Archontis Politis; Konstantinos Drossos
Pubblicato in: DCASE2021 - Detection and Classification of Acoustic Scenes and Events 2021 Workshop, 2021, Pagina/e 100-104, ISBN 978-84-09-36072-7
Editore: UPF
DOI: 10.5281/zenodo.5723115

Stateless or stateful FaaS? I'll take both!

Autori: Carlo Puliafito; Claudio Cicconetti; Marco Conti; Enzo Mingozzi; Andrea Passarella
Pubblicato in: SMARTCOMP 2022 - International Conference on Smart Computing 2022, 2022
Editore: IEEE
DOI: 10.5281/zenodo.6839548

Evaluating Short-Term Forecasting of Multiple Time Series in IoT Environments

Autori: Christos Tzagkarakis; Pavlos Charalampidis; Stylianos Roubakis; Alexandros Fragkiadakis; Sotiris Ioannidis
Pubblicato in: EUSIPCO 2022 - 30th European Signal Processing Conference, 2022
Editore: EURASIP
DOI: 10.5281/zenodo.6656593

Other (1)

Automatic Social Distance Estimation From Images: Performance Evaluation, Test Benchmark, and Algorithm

Autori: Mert Seker; Ansi Mannisto; Alexandros Iosifidis; Jenni Raitoharju
Pubblicato in: Technical Report, 2022
Editore: N/A
DOI: 10.5281/zenodo.6737052