Artificial Intelligence in Secure PRIvacy-preserving computing coNTinuum

Risultati finali

D4.2 Second release and evaluation of the security tools

Second release of the security tools that offer extended TEE support, including an initial version for patch management and secure boot. This deliverable includes an initial assessment of tools behaviour and performance.

D1.4 Final design of the architecture

This deliverable will describe the final architecture designed in T1.3.

D3.2 First release and evaluation of the monitoring system

This deliverable prototype and accompanying document will present the first version of the monitoring system including the database the monitoring service and the basic probes for the lowlevel monitoring and its evaluation

D7.2 Quality management plan

The document will describe the procedures to be adopted to guarantee adherence of deliverables to scientific and quality standards In particular rules and tools for quality review of deliverables will be detailed

D6.4 Communication, dissemination & stakeholder engagement-2nd report

The 2nd iteration of the report documents impacts achieved, KPIs and plans for the next period vis-à-vis the targeted stakeholders, communications and dissemination of results. It demonstrates progress towards the go-to-market strategy. It will update impacts and activities on clustering and collaborations.

D5.5 Final assessment report, impact analysis, lessons learned & best practice

The final report summarising all the results achieved in the use cases and the expected impact arising from these conclusions. The deliverable will also collect the best practices, learned while implementing the use cases, on the use of AI-SPRINT tools.

D2.2 Initial AI-SPRINT design and runtime tools integration

This document will provide the description of the first implementation of the AISPRINT architecture where all the available components from WP24 work together in an optimized fashion

D7.7 Ethics guidelines

The report will contain the guidelines on Ethics related issues and will present an overview of their implementation in the pilots including the consent form for the patients who will participate in the project

D6.2 Market outlook and forecast report

This deliverable will provide a detailed assessment of the impact potential of the technologies products and services being developed in the project It will highlight the economic potential of the leading vertical markets including growths and trends but also identify important social aspects such as job creation or skills needs

D2.1 First release and evaluation of the AI-SPRINT design tools

This deliverable prototype and accompanying document reports the activities coming from the initial development of the design tools including first releases of the extended version of COMPSs model and of performance models

D3.1 First release and evaluation of the runtime environment

A report and a prototype describing the Continuous Deployment and the Programming Framework Runtime as well as the results of preliminary tests on the technologies to support the design decisions

D3.5 Final release and evaluation of the runtime environment

A report and a prototype describing the final release of the runtime environment.

D3.3 Second release and evaluation of the runtime environment

A report and a prototype describing the first version of the Scheduling for accelerator devices and of the privacy preserving continuous training.

D1.1 State of the art analysis

This deliverable will present the SOTA review performed in T11

D2.3 Second release and evaluation of the AI-SPRINT design tools

This deliverable (prototype and accompanying document) reports on the second implementation of the programming layer, with the initial release of the tools for the AI Models Architecture Design and for the Application Design Space Exploration.

D5.1 Evaluation plan

This deliverable will follow up D12 created in T12 matching the requirements and the AISPRINT technology capabilities with the different use cases and extending the initial set of Business KPIs initially described in Sect 2 for the individual use cases

D1.3 Initial design of the architecture

This deliverable will describe the initial version of the architecture designed in T13

D2.4 Final release and evaluation of the AI-SPRINT design tools

Prototype and accompanying document of the final implementation of all the design tools with their evaluation on representative applications.

D4.3 Final release and evaluation of the security tools

The deliverable will provide the final security tools version and the full evaluation of their capabilities and performance.

D3.4 Final release and evaluation of the monitoring system

A report and a prototype describing the final version of the monitoring system, including high-level application metrics, as well as the probes for the edge devices, and its evaluation.

D4.1 Initial release and evaluation of the security tools

The deliverable will describe the security solutions used in AISPRINT and provide a first release of the Code and Network Security components that provide basic TEE support for security

D2.5 Final AI-SPRINT design and runtime tools integration

This document will provide the description of the implementation of the proposed architecture where all the components from WP2-4 work together in an optimized fashion.

D6.8 Communication, dissemination & stakeholder engagement-final report

This report demonstrates impacts in terms of dissemination and exploitation of results, as well as the critical success factors for AI-SPRINT sustainability, including potential adopters engaged. It will detail all measurable impacts, also for all the outputs and outcomes from clustering and collaborations. It will also report on the graphically designed white paper in T6.2.

D1.2 Requirements analysis

This report will describe the requirements for the use cases and AISPRINT tools identified in T12 D12 will be the base for the requirement repository which will be continuously updated until M18

D6.7 Market outlook and forecast report update

This deliverable will provide an update to the preliminary assessment detailed in D6.2 of the impact potential of the technologies, products and services being developed in the AI-SPRINT framework. It will also summarize the discussions, results and lessons learned of the two webinars focused on market outlook and forecasts (D6.5 and D6.6).

D6.1 Communication, dissemination & stakeholder engagement-1st report

The report will be the first of three iterations providing all the necessary details on activities to be performed visvis the targeted stakeholders communications and dissemination of results based on measurable indicators and monitoring It will include activities on clustering and collaborations

D6.6 AI-SPRINT Market outlook and forecasts update webinar

Following the first webinar (D6.5), this second webinar will provide updated market outlook and forecasts.

D6.5 AI-SPRINT Market outlook and forecasts webinar

This webinar will explore the market outlook for the technologies that make up the AI-SPRINT offering, including growth and differentiation across vertical sectors.

D7.3 Data management plan

The plan will explain how project data will be collected processed managed and stored the level of confidentiality of data and the approval process to publish project data and results following EC guidelines

Pubblicazioni

A Randomized Greedy Method for AI Applications Component Placement and Resource Selection in Computing

Autori: Hamta Sedghani, Federica Filippini, Danilo Ardagna
Pubblicato in: Proceedings of the 2021 IEEE International Conference on Joint Cloud Computing (JCC), Numero Yearly - 12th edition of JCC, 2021
Editore: IEEE

Formal Foundations for Intel SGX Data Center Attestation Primitives

Autori: Muhammad Usama Sardar, Rasha Faqeh, Christof Fetzer
Pubblicato in: International Conference on Formal Engineering Methods (ICFEM), Numero 22, 2021, Pagina/e 16, ISBN 978-3-030-63405-6
Editore: Springer
DOI: 10.13140/rg.2.2.36760.21768

Discriminative Adversarial Privacy: Balancing Accuracy and Membership Privacy in Neural Networks

Autori: Eugenio Lomurno and Alberto Archetti and Francesca Ausonio and Matteo Matteucci
Pubblicato in: 2023
Editore: arXiv
DOI: 10.48550/arxiv.2306.03054

Perun: Confidential Multi-stakeholder Machine Learning Framework with Hardware Acceleration Support

Autori: Wojciech Ozga, Do Le Quoc, Christof Fetzer
Pubblicato in: IFIP Annual Conference on Data and Applications Security and Privacy, Numero Yearly - 35th edition of DBSEC, 2021, Pagina/e 20
Editore: Springer
DOI: 10.1007/978-3-030-81242-3_11

Pareto-Optimal Progressive Neural Architecture Search

Autori: E. Lomurno, S. Samele, M. Matteucci, D. Ardagna. POLIMI
Pubblicato in: NEvo@Work ACM Workshop on NeuroEvolution@Work, 2021
Editore: ACM

Bridging the Gap: Enhancing the Utility of Synthetic Data via Post-Processing Techniques

Autori: Andrea Lampis, Eugenio Lomurno, Matteo Matteucci
Pubblicato in: 2023
Editore: arXiv
DOI: 10.48550/arxiv.2305.10118

Understanding Trust Assumptions for Attestation in Confidential Computing

Autori: Muhammad Usama Sardar
Pubblicato in: 52nd Annual IEEE/IFIP International Conference on Dependable Systems and Networks, Numero Supplemental Volume (DSN-S), 2022, Pagina/e pp. 49-50,, ISSN 1803-7232
Editore: IEEE

A Last-Level Defense for Application Integrity and Confidentiality

Autori: Gabriel P. Fernandez, Andrey Brito, Ardhi Putra Pratama Hartono, Muhammad Usama Sardar, Christof Fetzer
Pubblicato in: UCC '23: IEEE/ACM. 16th International Conference on Utility and Cloud Computing (UCC '23), 2023, Pagina/e 10 pages
Editore: ACM
DOI: 10.48550/arxiv.2311.06154

Hierarchical Management of Extreme-Scale Task-Based Applications

Autori: Lordan, F., Puigdemunt, G., Vergés, P., Conejero, J., Ejarque, J., Badia, R.M.
Pubblicato in: European Conference on Parallel Processing, 2023, Pagina/e pp 111–124, ISBN 978-3-031-39697-7
Editore: Springer Nature

CHIMA: a Framework for Network Services Deployment and Performance Assurance

Autori: Battiston, E.; Moro, D.; Verticale, G. Capone, A.
Pubblicato in: 2022 IEEE 8th International Conference on Network Softwarization (NetSoft), 2022
Editore: Institute of Electrical and Electronics Engineers (IEEE)

Scalable Random Forest with DataParallel Computing

Autori: Vázquez-Novoa, F., Conejero, J., Tatu, C., Badia, R.M.
Pubblicato in: European Conference on Parallel Processing, 2023, Pagina/e pp 397–410, ISBN 978-3-031-39697-7
Editore: Springer Nature

Securing the Execution of ML Workflows across the Compute Continua

Autori: Francesc-Josep Lordan Gomis, André Martin, Daniele Lezzi
Pubblicato in: ICPE '23 Companion: Companion of the 2023 ACM/SPEC International Conference on Performance Engineering, 2023, ISBN 9798400700729
Editore: Association for Computing Machinery
DOI: 10.1145/3578245

ADAM-CS - Advanced Asynchronous Monotonic Counter Service

Autori: André Martin, Cong Lian, Franz Gregor, Robert Krahn, Valerio Schiavoni, Pascal Felber, Christof Fetzer
Pubblicato in: DSN 2021, 2021, Pagina/e 12
Editore: IEEE
DOI: 10.1109/dsn48987.2021.00053

A Datalog Hammer for Supervisor Verification Conditions Modulo Simple Linear Arithmetic

Autori: Martin Bromberger, Irina Dragoste, Rasha Faqeh (TUD), Christof Fetzer (TUD), Markus Krötzsch (TUD), Christoph Weidenbach (TUD)
Pubblicato in: Frontiers of Combining Systems - 13th International Symposium, FroCoS 2021, Birmingham, UK, September 8-10, 2021, Proceedings, Numero Yearly, Lecture Notes in Computer Science, 2021
Editore: Springer

Formal Foundations for SCONE Attestation

Autori: M. U. Sardar and C. Fetzer
Pubblicato in: 2022 52nd Annual IEEE/IFIP International Conference on Dependable Systems and Networks - Supplemental Volume (DSN-S), 2022, Pagina/e pp. 31-32, ISBN 978-1-6654-0260-6
Editore: IEEE
DOI: 10.1109/dsn-s54099.2022.00020

Federated Survival Forests

Autori: A. Archetti, M. Matteucci
Pubblicato in: International Joint Conference on Neural Networks, 2023
Editore: Institute of Electrical Electronics Engineer
DOI: 10.48550/arxiv.2302.02807

Advancing Design and Runtime Management of AI Applications with AI-SPRINT. Applications, Challenges & Concerns). 2021.

Autori: H. Sedghani, D. Ardagna, M. Matteucci, G. A. Fontana, G. Verticale, F. Amarilli* R. Badia, D. Lezzi++ I. Blanquer** A. Martin*** K. Wawruch.+++ *POLIMI, ++BSC, **UPV, *** TUD, +++ 7BULLS
Pubblicato in: AIML 2021 Workshop Proceedings (Advances in Artificial Intelligence & Machine Learning, 2021
Editore: IEEE

Anticipate, Ensemble and Prune: Improving Convolutional Neural Networks via Aggregated Early Exits

Autori: Simone Sarti, Eugenio Lomurno, Matteo Matteucci
Pubblicato in: Procedia Computer Science, Numero Volume 222, 2023, Pages 519-528, 2023, ISSN 1877-0509
Editore: Elsevier
DOI: 10.1016/j.procs.2023.08.190

OSCAR-P and aMLLibrary: Performance Profiling and Prediction of Computing Continua Applications

Autori: Enrico Galimberti, Bruno Guindani, Federica Filippini, Hamta Sedghani, Danilo Ardagna, Germán Moltó, Miguel Caballer
Pubblicato in: ICPE '23 Companion: Companion of the 2023 ACM/SPEC International Conference on Performance Engineering, 2023, ISBN 9798400700729
Editore: Association for Computing Machinery
DOI: 10.1145/3578245.3584941

Enhancing Once-For-All: A Study on Parallel Blocks, Skip Connections and Early Exits

Autori: S. Sarti, E. Lomurno, A. Falanti, M. Matteucci
Pubblicato in: International Joint Conference on Neural Networks, 2023
Editore: Institute of Electrical and Electronic Engineers
DOI: 10.48550/arxiv.2302.01888

Colony: Parallel Functions as a Service on the Cloud-Edge Continuum

Autori: Francesc Lordan, Daniele Lezzi & Rosa M. Badia
Pubblicato in: 2021, ISBN 978-3-030-85665-6
Editore: Springer

Secure execution of ML workflows on the Computing Continuum

Autori: F. J. Lordan Gomis, A. Martin, D. Lezzi
Pubblicato in: ICPE '23 Companion: Companion of the 2023 ACM/SPEC International Conference on Performance Engineering, 2023
Editore: Association for Computing Machinery
DOI: 10.1145/3578245.3584727

On the Acceleration of FaaS Using Remote GPU Virtualization

Autori: Naranjo Delgado, D. M., Contreras, M., Moltó, G., Risco, S., Blanquer, I., Prades, J., & Silla, F
Pubblicato in: Companion of the 2023 ACM/SPEC International Conference on Performance Engineering, 2023, ISBN 9798400700729
Editore: Association for Computing Machinery
DOI: 10.1145/3578245.3584933

Challenges Towards Modeling and Generating Infrastructure-as-Code

Autori: G. Nedeltcheva, B. Xiang, L. Niculut, D. Benedetto
Pubblicato in: FastContinuum 2023, Numero ICPE '23 Companion: Companion of the 2023 ACM/SPEC International Conference on Performance Engineering, 2023
Editore: Association for Computing Machinery
DOI: 10.1145/3578245.3584937

Heterogeneous Datasets for Federated Survival Analysis Simulation

Autori: A.Archetti, E. Lomurno, F. Lattari, A. Martin, M. Matteucci
Pubblicato in: FastContinuum workshop (Coimbra, Portugal, 15/16 April 2023), 2023
Editore: Association for Computing Machinery
DOI: 10.1145/3578245.3584935

A Random Greedy based Design Time Tool for AI Applications Component Placement and Resource Selection in Computing Continua

Autori: Hamta Sedghani, Federica Filippini, Danilo Ardagna
Pubblicato in: IEEE International Conference on Edge Computing (EDGE) 2021 (pp. 1-9), Numero Yearly - 5th edition of IEEE EDGE, 2021
Editore: IEEE

Scheduling Deep Learning Jobs Training in the Cloud: Comparing Multiple Approaches

Autori: M. Precuzzi, F. Filippini, D. Ardagna
Pubblicato in: SPARK-2022 proceedings (in CEUR-WS Workshop Proceedings), 2022
Editore: RWTH Aachen University

A Sorted Datalog Hammer for Supervisor Verification Conditions Modulo Simple Linear Arithmetic

Autori: Martin Bromberger, Irina Dragoste, Rasha Faqeh, Christof Fetzer, Larry González, Markus Krötzsch, Maximilian Marx, Harish K Murali, Christoph Weidenbach
Pubblicato in: Lecture Notes in Computer Science book series (LNCS,volume 13243), 2022
Editore: Springer
DOI: 10.1007/978-3-030-99524-9_27

SGDE: Secure Generative Data Exchange for Cross-Silo Federated Learning

Autori: E. Lomurno, A. Archetti, L. Cazzella, S. Samele, L. Di Perna, M. Matteucci
Pubblicato in: ACM International Conference on Artificial Intelligence and Pattern Recognition, 2022
Editore: Association for Computing Machinery
DOI: 10.1145/3573942.3573974

CHORS: hardening high-assurance security systems with trusted computing

Autori: Wojciech Ozga, Rasha Faqeh, Do Le Quoc, Franz Gregor, Silvio Dragone, Christof Fetzer
Pubblicato in: ACM/SIGAPP Symposium on Applied Computing, 2022
Editore: ACM
DOI: 10.1145/3477314.3506961

Trustworthy confidential virtual machines for the masses

Autori: Anna Galanou, Khushboo Bindlish, Luca Preibsch, Yvonne-Anne Pignolet, Christof Fetzer, Rüdiger Kapitza
Pubblicato in: Middleware '23: 24th International Middleware Conference, 2023, Pagina/e Pages 316–328, ISBN 9798400701771
Editore: Association for Computing Machinery
DOI: 10.1145/3590140.3629124

SinClave: Hardware-assisted Singletons for TEEs

Autori: Franz Gregor, Robert Krahn, Do Le Quoc, Christof Fetzer
Pubblicato in: Middleware '23: 24th International Middleware Conference, 2023
Editore: Association for Computing Machinery
DOI: 10.1145/3590140.3629107

Serverless Workflows for Containerised Applications in the Cloud Continuum

Autori: Risco, Sebastián; Moltó, Germán; Naranjo, Diana M.; Blanquer, Ignacio
Pubblicato in: Journal of Grid Computing, Numero 2, 2021, ISSN 1570-7873
Editore: Kluwer Academic Publishers
DOI: 10.1007/s10723-021-09570-2

POPNASv3: A pareto-optimal neural architecture search solution for image and time series classification

Autori: Falanti, Andrea; Lomurno, Eugenio; Ardagna, Danilo; Matteucci, Matteo
Pubblicato in: Applied Soft Computing, Numero Volume 145, September 2023, 2023, ISSN 1872-9681
Editore: Science-DIrect
DOI: 10.1016/j.asoc.2023.110555

A Stochastic Approach for Scheduling AI Training Jobs in GPU-based Systems

Autori: Federica Filippini, Jonatha Anselmi, Danilo Ardagna, Bruno Gaujal
Pubblicato in: IEEE Transactions on Cloud Computing, Numero vol. , no. 01, 2023, Pagina/e pp. 1-17,, ISSN 2168-7161
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/tcc.2023.3336540

Perun: Confidential Multi-stakeholder Machine Learning Framework with Hardware Acceleration Support

Autori: Wojciech Ozga, Do Le Quoc, Christof Fetzer (TUD)
Pubblicato in: Springer - DBSec 2021: Data and Applications Security and Privacy XXXV, 2021, ISSN 1611-3349
Editore: Spinger
DOI: 10.1007/978-3-030-81242-3_11

Demystifying Attestation in Intel Trust Domain Extensions via Formal Verification

Autori: Muhammad Usama Sardar, Saidgani Musaev, Christof Fetzer
Pubblicato in: IEEE ACCESS, 2021, ISSN 2169-3536
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/access.2021.3087421

A Path Relinking Method for the Joint Online Scheduling and Capacity Allocation of DL Training Workloads in GPU as a Service Systems

Autori: Federica Filippini, Marco Lattuada, Michele Ciavotta, Arezoo Jahani, Danilo Ardagna, Edoardo Amaldi
Pubblicato in: IEEE TRANSACTIONS ON SERVICES COMPUTING, Numero VOL. 16, NO. 3, MAY/JUNE 2023, 2023, ISSN 1939-1374
Editore: Institute of Electrical and Electronics Engineers

Rescheduling Serverless Workloads Across the Cloudto-Edge Continuum

Autori: Sebastián Risco, Caterina Alarcón, Sergio Langarita, Miguel Caballer, Germán Moltó
Pubblicato in: Future Generation Computer Systems, Numero Volume 153, April 2024, Pages 457-466, 2023, ISSN 0020-0255
Editore: Elsevier BV
DOI: 10.1016/j.future.2023.12.015

Capacity Planning for Dependable Services

Autori: R. Faqeh, A. Martin, V. Schiavoni, P. Bhatotia, P. Felber, C. Fetzer
Pubblicato in: Theoretical Computer Science, Numero Volume 976, 17 October 2023, 114126, 2023, ISSN 0346-251X
Editore: Pergamon Press Ltd.
DOI: 10.1016/j.tcs.2023.114126

Performance Prediction of Deep Learning Applications Training in GPU as a Service Systems

Autori: Marco Lattuada, Eugenio Gianniti, Danilo Ardagna (POLIMI), Li Zhang
Pubblicato in: Springer Computing, 2021, ISSN 2193-1801
Editore: Springer Science and Business Media Deutschland GmbH
DOI: 10.1007/s10586-021-03428-8

A Serverless Gateway for Event-driven Machine Learning Inference in Multiple Clouds

Autori: Diana M. Naranjo, Sebastián Risco, Germán Moltó, Ignacio Blanquer (UPV)
Pubblicato in: Concurrency and Computation. Practice and Experience, Numero Ahead of Print, 2022, ISSN 1532-0634
Editore: John Wiley & Sons Ltd
DOI: 10.1002/cpe.6728

TaScaaS: A Multi-Tenant Serverless Task Scheduler and Load Balancer as a Service

Autori: Vicent Gimenez-Alventosa; German Molto; J. Damian Segrelles
Pubblicato in: IEEE Access, Vol 9, Pp 125215-125228 (2021), Numero 3, 2021, ISSN 2169-3536
Editore: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/access.2021.3109972

Network Function Decomposition and Offloading on Heterogeneous Networks With Programmable Data Planes

Autori: Daniele Moro, Giacomo Verticale, Antonio Capone
Pubblicato in: IEEE Open Journal of the Communications Society, 2021, ISSN 2644-125X
Editore: IEEE
DOI: 10.1109/ojcoms.2021.3101366

SPACE4AI-R: a Runtime Management Tool for AI Applications Component Placement and Resource Scaling in Computing Continua

Autori: Federica Filippini, Hamta Sedghani, Danilo Ardagna
Pubblicato in: 2023
Editore: Politecnico di Milano

Risultati finali

Pubblicazioni

Condividi questa pagina Condividi questa pagina sui social network

Scarica Scarica il contenuto della pagina