Message from EAI President regarding COVID-19
eScripts is currently undergoing an upgrade and heavy maintenance that may cause instability to the service and unsolicited emails going out to users in the upcoming few days. We apologize for any inconvenience that his may cause you and we appreciate your understanding as we continue to improve your experience.
If you encounter significant issues with eScripts, please let us know at publications@eai.eu.

Security and Privacy Issues for the Artificial Intelligence Era Special Issue

Editorial Board
Special Issue Editors:
  • Ximeng Liu (Fuzhou University)
  • Yinbin Miao ()
Aims & Scope
Artificial intelligence, or AI for short, enables machines to adjust their knowledge based on new inputs that were not part of the data used for training these machines, which can be used in the daily application that increases productivity. It is strongly believed that it will make the life of the next generations and created in a way to eliminate modification and disfigurement in a hostile environment. Unfortunately, the AI paradigm brings many security and privacy issues such as information confidentiality, authentication and authorization, and secure communication. To solve the above security and privacy issues, the AI system can use some traditional cryptology methods to guarantee the security of data owner and user, such as using an encryption algorithm to protect the input data of AI model. However, the third-party servers cannot do any operations on the encrypted data as the original meaning of the data is lost after encryption. Some a heavyweight lattice-based cryptosystem can be used to design homomorphic encryption, which supports secure operations on the encrypted data. With nearly unlimited computation and storage resources, the heavyweight homomorphic encryption still cannot be performed on the billions of resource-constrained edge or Internet of things devices. Moreover, besides the outsider attackers, the AI model still has the inherent security problems such as adversarial examples attacks, which intentionally designed to cause the model to make a mistake. It is still hard to ensure that sophisticated AI models that are significantly more intelligent and secure than human beings behave in ways that their designers intended. To address these arising challenges and opportunities different from traditional cloud based architectures, this special issue is interested in inviting and gathering recent advanced security and privacy techniques relevant to the convergence of artificial intelligence.
  • Authentication in artificial intelligence system
  • Lightweight data protection scheme for smart AI devices
  • Secure model protection method for AI
  • Hardware security and privacy issues for AI devices
  • Secure data integrity and validation techniques for smart AI devices
  • Privacy computation and processing protocols on AI platform
  • Secure M2M communication for federated learning and AI
  • Formal security model for cryptographic protocols for AI
  • Future and smart AI-based vulnerability assessment interfaces
  • Against adversarial examples for AI
  • Secure programming models and toolkits for AI
  • Virtualization, security, and management for AI
  • Experiment prototype for secure and trusted AI framework
  • Blockchain and smart contracts for AI system
  • New privacy challenges in AI
Keywords
Manuscript submission deadline:
2020-08-20
Notification of acceptance:
2020-09-20
Submission of final revised paper:
2020-10-22
Publication of special issue (tentative):
2020-11-12

Publication:
Security and Safety