Gabriele Costa

Gabriele Costa

Auto-generated profile, based on publicly available information

Gabriele Costa

IMT School for Advanced Studies, Lucca, Italy

Research Focus & Metrics

Academic Career Length

18Years (2009-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 August 2020 in Computers & Security
Verified
Journal Article
Closed access
FSFederico Sinigaglia
Federico Sinigaglia
RCRoberto Carbone
Roberto Carbone
GCGabriele Costa
Gabriele Costa
Nicola ZannoneNicola Zannone
Nicola Zannone

1 August 2020 in Computers & Security
Verified
Journal Article
Closed access
EREnrico Russo
Enrico Russo
GCGabriele Costa
Gabriele Costa
AAAlessandro Armando
Alessandro Armando

18 March 2013
Verified
Conference Paper
Open access
PDF file
AAAlessandro Armando
Alessandro Armando
GCGabriele Costa
Gabriele Costa
AMAlessio Merlo
Alessio Merlo

18 June 2014 in Computer
Verified
Journal Article
Closed access
AAAlessandro Armando
Alessandro Armando
GCGabriele Costa
Gabriele Costa
LVLuca Verderame
Luca Verderame
AMAlessio Merlo
Alessio Merlo
The current mobile application distribution model cannot cope with the complex security requirements of the emerging "bring your own device" (BYOD) paradigm. A secure metamarket architecture supports the definition and enforcement of BYOD policies and offers a promising prototype implementation tested under realistic conditions.

14 December 2022 in IEEE Access
Verified
Journal Article
Open access
GCGabriele Costa
Gabriele Costa
FPFabio Pinelli
Fabio Pinelli
SSSimone Soderi
Simone Soderi
GTGabriele Tolomei
Gabriele Tolomei
Federated learning (FL) goes beyond traditional, centralized machine learning by distributing model training among a large collection of edge clients. These clients cooperatively train a global, e.g., cloud-hosted, model without disclosing their local, private training data. The global model is then shared among all the participants which use it for local predictions. This paper proves that FL systems can be turned into covert channels to implement a stealth communication infrastructure. The main intuition is that, during federated training, a malicious sender can poison the global model by submitting purposely crafted examples. Although the effect of the model poisoning is negligible to other participants and does not alter the overall model performance, it can be observed by a malicious receiver and used to transmit a sequence of bits. We mounted our attack on an FL system to verify its feasibility. Experimental evidence shows that this covert channel is reliable, efficient, and extremely hard to counter. These results highlight that our new attacker model threatens FL infrastructures.

23 July 2014
Verified
Conference Paper
Open access
PDF file
AAAlessandro Armando
Alessandro Armando
GCGabriele Costa
Gabriele Costa
AMAlessio Merlo
Alessio Merlo
LVLuca Verderame
Luca Verderame