Visible to the public Biblio

Filters: Author is Svetinovic, Davor  [Clear All Filters]
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z 
Habib ur Rehman, Muhammad, Mukhtar Dirir, Ahmed, Salah, Khaled, Svetinovic, Davor.  2020.  FairFed: Cross-Device Fair Federated Learning. 2020 IEEE Applied Imagery Pattern Recognition Workshop (AIPR). :1–7.
Federated learning (FL) is the rapidly developing machine learning technique that is used to perform collaborative model training over decentralized datasets. FL enables privacy-preserving model development whereby the datasets are scattered over a large set of data producers (i.e., devices and/or systems). These data producers train the learning models, encapsulate the model updates with differential privacy techniques, and share them to centralized systems for global aggregation. However, these centralized models are always prone to adversarial attacks (such as data-poisoning and model poisoning attacks) due to a large number of data producers. Hence, FL methods need to ensure fairness and high-quality model availability across all the participants in the underlying AI systems. In this paper, we propose a novel FL framework, called FairFed, to meet fairness and high-quality data requirements. The FairFed provides a fairness mechanism to detect adversaries across the devices and datasets in the FL network and reject their model updates. We use a Python-simulated FL framework to enable large-scale training over MNIST dataset. We simulate a cross-device model training settings to detect adversaries in the training network. We used TensorFlow Federated and Python to implement the fairness protocol, the deep neural network, and the outlier detection algorithm. We thoroughly test the proposed FairFed framework with random and uniform data distributions across the training network and compare our initial results with the baseline fairness scheme. Our proposed work shows promising results in terms of model accuracy and loss.
Svetinovic, Davor.  2017.  Blockchain Engineering for the Internet of Things: Systems Security Perspective. Proceedings of the 3rd ACM International Workshop on IoT Privacy, Trust, and Security. :1–1.
The Internet of Things (IoT) technology has a potential to bring the benefits of intelligently interconnecting not just computers and humans, but most of everyday things. IoT has a promise of opening significant business process improvement opportunities leading to economic growth and cost reductions. However, there are many challenges facing IoT, including significant scalability and security challenges due to the integration of potentially huge number of things into the network. Many of scalability and security issues stem from a centralized, primarily client/server, architecture of IoT systems and frameworks. Blockchain technology, as a relativelly new approach to decentralized computation and assets management and transfer, has a potential to help solve a number of scalability and security issues that IoT is facing, primarilly through the removal of centralized points of failure for such systems. As such, blockchain technology and IoT integration provides a promising direction and it has recently generated significant research interest, e.g., [4]. In this talk, we present our experiences based on our recent project in enhancing security and privacy in decentralized energy trading in smart grids using blockchain, multi-signatures and anonymous messaging streams [1], that has built upon our previous work on Bitcoin-based decentralized carbon emissions trading infrastructure model [2]. In particular, we present the blockchain systems security issues within the context of IoT security and privacy requirements [3]. This is done with the intention of producing an early integrated security model for blockchain-powered IoT systems [5]. The presentation is constrained to the discussion of the architecture-level requirements [6]. Finally, we will present the main opportunity loss if the integration ignores the full realization of the real-world asset transaction paradigm.
Yousuf, Soha, Svetinovic, Davor.  2019.  Blockchain Technology in Supply Chain Management: Preliminary Study. 2019 Sixth International Conference on Internet of Things: Systems, Management and Security (IOTSMS). :537–538.

Despite significant research, the supply chain management challenges still have a long way to go with respect to solving the issues such as management of product supply information, product lifecycle, transport history, etc. Given the recent rise of blockchain technology in various industrial sectors, our work explores the issues prevalent in each stage of the supply chain and checks their candidacy for the implementation using blockchain technology. The analysis is performed in terms of the characteristics of trust and decentralization with respect to forming a generalized framework. The main contribution of this work is to create a conceptual overview of the areas where blockchain integrates with supply chain management in order to benefit further research and development.