software framework for runtime-Adaptive and secure deep Learning On Heterogeneous Architectures

Projektinformationen

ALOHA

ID Finanzhilfevereinbarung: 780788

DOI

10.3030/780788

Projekt abgeschlossen

EK-Unterschriftsdatum 20 November 2017

Startdatum 1 Januar 2018

Enddatum 30 Juni 2021

Finanziert unter

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

Gesamtkosten

€ 5 976 415,00

EU-Beitrag

€ 5 976 415,00

Koordiniert durch

STMICROELECTRONICS SRL
Italy

CORDIS bietet Links zu öffentlichen Ergebnissen und Veröffentlichungen von HORIZONT-Projekten.

Links zu Ergebnissen und Veröffentlichungen von RP7-Projekten sowie Links zu einigen Typen spezifischer Ergebnisse wie Datensätzen und Software werden dynamisch von OpenAIRE abgerufen.

Leistungen

Open Research Data Pilot (3)

Open Data Management Plan – First Update

ALOHA takes part in the Pilot on Open Research Data. This report provides the data management plan, updated at the second year period.

Open Data Management Plan

ALOHA takes part in the Pilot on Open Research Data. This report provides the data management plan, provided at the beginning of the project.

Open Data Management Plan - Final Update

ALOHA takes part in the Pilot on Open Research Data. This report provides the data management plan, updated at the third and final year period.

Documents, reports (25)

Kick-off progress report

Reports on the management activity and status of the project in the first six months. This deliverable will collect also the relevant feedback coming from the scientific and administrative meetings already organized

Report on automated application partitioning and mapping

Report on the research and development activities concerning the utilities converging in the ALOHA automated tool for application partitioning and mapping. This report will serve also as a lightweight form of documentation for D3.3 and D3.4

Exploitation plan - Final Update

This final version of the report will describe the final plan for the exploitation of foreground.

Report on automated algorithm configuration

Report on the research and development activities concerning the utilities converging in the ALOHA automated algorithm configuration tool. This report will serve also as a lightweight form of documentation for D2.3 and D2.4.

Plan for the dissemination and communication - First Update

Updated Plan for the communication and dissemination activities during and after the project lifetime

Exploitation plan – First Update

Updated plan for the exploitation activities during and after the project lifetime. This midterm release is an update of the ALOHA preliminary plan.

First progress and management report

Reports on the management activity and status of the project in the first year. This deliverable will collect also the relevant feedback coming from the scientific and administrative meetings already organized.

Report on tool flow integration - Update

Updated report on the activities related with the ALOHA tool flow integration.

Final progress and management report

Reports on the management activity and status of the project in the third and final year. This deliverable will collect also the relevant feedback coming from the scientific and administrative meetings already organized.

Report on the exploitation activities

Report on the exploitation activities during the first part of the project.

Report on general specifications and interface definition

Report on the specification of the project tool flow, expected features of independent tools and methods under development, as well as a design of the interfaces between them.

Report on automated algorithm configuration – Update

Updated report on the research and development activities concerning the utilities converging in the ALOHA automated algorithm configuration tool. This report will serve also as a lightweight form of documentation for D2.3 and D2.4.

Exploitation plan

Plan for the exploitation activities during and after the project lifetime. This first release, at M4, will define ALOHA preliminary plan.

Project quality handbook

This handbook constitutes a set of project templates, on the project management process, review process, quality checks, meetings organisation, etc communicated to all partners

Report on automated application partitioning and mapping – Update

Updated report on the research and development activities concerning the utilities converging in the ALOHA automated tool for application partitioning and mapping. This report will serve also as a lightweight form of documentation for D3.3 and D3.4.

Report on tool flow integration

Report on the activities related with the ALOHA tool flow integration.

Report on hardware abstraction layer techniques

Report on the research and development activities concerning support for hardware abstraction layer and runtime environment for ALOHA reference computing platforms. This report will serve also as a lightweight form of documentation for D4.3 and D4.4.

Plan for the dissemination and communication - Final Update

The final version of the plan will describe the final plan after the project lifetime.

Report on the exploitation activities - Update

Report on the exploitation activities during the final part of the project.

Final report

It will report on the overall activities of the ALOHA project. It will contain also a chapter on awareness and wider societal implications.

Second progress and management report

Reports on the management activity and status of the project in the second year. This deliverable will collect also the relevant feedback coming from the scientific and administrative meetings already organized.

Report on hardware abstraction layer techniques – Update

Updated report on the research and development activities concerning support for hardware abstraction layer and runtime environment for ALOHA reference computing platforms. This report will serve also as a lightweight form of documentation for D4.3 and D4.4.

Dissemination and communication report - Update

Report on the dissemination and communication during the second part of the project. This deliverable will assess the impact of communication and dissemination (social media diffusion, website analytics, press releases, events participation and publications). It will also report on the actions taken towards the community of users and their feedback.

Plan for the dissemination and communication

ALOHA preliminary plan for the communication and dissemination activities during and after the project lifetime.

Dissemination and communication report

Report on the dissemination and communication during the first part of the project. This deliverable will assess the impact of communication and dissemination (social media diffusion, website analytics, press releases, events participation and publications). It will also report on the actions taken towards the community of users and their feedback.

Demonstrators, pilots, prototypes (8)

Project digital presence

This deliverable provides (and reports on) the release of the official project’s website for dissemination and exchange of data outside and inside the consortium, and on the project’s presence on dedicated social media channels or groups (in particular on Twitter and LinkedIn).

Final release of the automated algorithm configuration tool

This deliverable includes a final release of the automated algorithm configuration tool. It will consist of a complete version, ready to be integrated in the overall ALOHA tool flow

Final release of the hardware abstraction layer utilities

This deliverable includes a final release of the utilities and libraries implementing hardware abstraction layer and runtime environment for ALOHA reference computing platforms. It will consist of a complete version, implementing also runtime platform management, ready to be integrated in the overall tool flow.

First release of the hardware abstraction layer utilities

This deliverable includes a first release of the utilities and libraries implementing hardware abstraction layer and runtime environment for ALOHA reference computing platforms. It will contain simple and static implementation of DL algorithms on each considered platform, making use of the techniques and libraries under development for hardware abstraction.

First release of the automated algorithm configuration tool

This deliverable includes a first release of the automated algorithm configuration tool. It will contain standalone implementation of the single utilities, to be used as a reference for integration, use-case refinement and demonstrator preparation.

First release of the automated tool for application partitioning and mapping

This deliverable includes a first release of the automated tool for application partitioning and mapping. It will contain standalone implementation of the single utilities, to be used as a reference for integration, use-case refinement and demonstrator preparation

Final prototype of the integrated tool flow

The deliverable involves a final release of the ALOHA tool flow

Final release of the automated tool for application partitioning and mapping

This deliverable includes a final release of the automated tool for application partitioning and mapping. It will consist of a complete version, ready to be integrated in the overall ALOHA tool flow.

Other (1)

ALOHA Final Workshop

Final workshop to present the project results

Veröffentlichungen

Conference proceedings (7)

Optimization and deployment of CNNs at the edge - the ALOHA experience

Autoren: Paolo Meloni, Luca Benini, Maura Pintor, Battista Biggio, Bernhard Moser, Natalia Shepeleva, Nikos Fragoulis, Ilias Theodorakopoulos, Michael Masin, Francesca Palumbo, Daniela Loi, Paola Busia, Gianfranco Deriu, Andy D. Pimentel, Dolly Sapra, Todor Stefanov, Svetlana Minakova, Francesco Conti
Veröffentlicht in: Proceedings of the 16th ACM International Conference on Computing Frontiers - CF '19, 2019, Seite(n) 326-332, ISBN 9781-450366854
Herausgeber: ACM Press
DOI: 10.1145/3310273.3323435

Architecture-aware design and implementation of CNN algorithms for embedded inference: the ALOHA project

Autoren: Paolo Meloni, Daniela Loi, Gianfranco Deriu, Andy D. Pimentel, Dolly Saprat, Maura Pintort, Battista Biggio, Oscar Ripolles, David Solans, Francesco Conti, Luca Benini, Todor Stefanov, Svetlana Minakova, Bernhard Moser, Natalia Shepeleva, Michael Masin, Francesca Palumbo, Nikos Fragoulis, Ilias Theodorakopoulos
Veröffentlicht in: 2018 30th International Conference on Microelectronics (ICM), 2018, Seite(n) 52-55, ISBN 978-1-5386-8167-1
Herausgeber: IEEE
DOI: 10.1109/ICM.2018.8704093

ALOHA - an architectural-aware framework for deep learning at the edge

Autoren: P. Meloni, O. Ripolles, D. Solans, M. Pintor, B. Biggio, T. Stefanov, S. Minakova, N. Fragoulis, I. Theodorakopoulos, M. Masin, F. Palumbo, D. Loi, G. Deriu, A. D. Pimentel, D. Sapra, B. Moser, N. Shepeleva, F. Conti, L. Benini
Veröffentlicht in: Proceedings of the Workshop on INTelligent Embedded Systems Architectures and Applications - INTESA '18, 2018, Seite(n) 19-26, ISBN 9781-450365987
Herausgeber: ACM Press
DOI: 10.1145/3285017.3285019

A runtime-adaptive cognitive IoT node for healthcare monitoring

Autoren: Matteo Antonio Scrugli, Daniela Loi, Luigi Raffo, Paolo Meloni
Veröffentlicht in: Proceedings of the 16th ACM International Conference on Computing Frontiers - CF '19, 2019, Seite(n) 350-357, ISBN 9781-450366854
Herausgeber: ACM Press
DOI: 10.1145/3310273.3323160

Manipulating Machine Learning: Poisoning Attacks and Countermeasures for Regression Learning

Autoren: Matthew Jagielski, Alina Oprea, Battista Biggio, Chang Liu, Cristina Nita-Rotaru, Bo Li
Veröffentlicht in: 2018 IEEE Symposium on Security and Privacy (SP), 2018, Seite(n) 19-35, ISBN 978-1-5386-4353-2
Herausgeber: IEEE
DOI: 10.1109/sp.2018.00057

Explaining Black-box Android Malware Detection

Autoren: Marco Melis, Davide Maiorca, Battista Biggio, Giorgio Giacinto, Fabio Roli
Veröffentlicht in: 2018 26th European Signal Processing Conference (EUSIPCO), 2018, Seite(n) 524-528, ISBN 978-9-0827-9701-5
Herausgeber: IEEE
DOI: 10.23919/eusipco.2018.8553598

Adversarial Malware Binaries: Evading Deep Learning for Malware Detection in Executables

Autoren: Bojan Kolosnjaji, Ambra Demontis, Battista Biggio, Davide Maiorca, Giorgio Giacinto, Claudia Eckert, Fabio Roli
Veröffentlicht in: 2018 26th European Signal Processing Conference (EUSIPCO), 2018, Seite(n) 533-537, ISBN 978-9-0827-9701-5
Herausgeber: IEEE
DOI: 10.23919/eusipco.2018.8553214

Peer reviewed articles (2)

An integrated hardware/software design methodology for signal processing systems

Autoren: Lin Li, Carlo Sau, Tiziana Fanni, Jingui Li, Timo Viitanen, François Christophe, Francesca Palumbo, Luigi Raffo, Heikki Huttunen, Jarmo Takala, Shuvra S. Bhattacharyya
Veröffentlicht in: Journal of Systems Architecture, Ausgabe 93, 2019, Seite(n) 1-19, ISSN 1383-7621
Herausgeber: Elsevier BV
DOI: 10.1016/j.sysarc.2018.12.010

Wild patterns: Ten years after the rise of adversarial machine learning

Autoren: Battista Biggio, Fabio Roli
Veröffentlicht in: Pattern Recognition, Ausgabe 84, 2018, Seite(n) 317-331, ISSN 0031-3203
Herausgeber: Pergamon Press
DOI: 10.1016/j.patcog.2018.07.023