Jon Crowcroft

Bio: Jon Crowcroft is an academic researcher from University of Cambridge. The author has contributed to research in topics: The Internet & Multicast. The author has an hindex of 87, co-authored 672 publications receiving 38848 citations. Previous affiliations of Jon Crowcroft include Memorial University of Newfoundland & Information Technology University.

Federated Principal Component Analysis

Abstract: We present a federated, asynchronous, and $(\varepsilon, \delta)$-differentially private algorithm for PCA in the memory-limited setting. Our algorithm incrementally computes local model updates using a streaming procedure and adaptively estimates its $r$ leading principal components when only $\mathcal{O}(dr)$ memory is available with $d$ being the dimensionality of the data. We guarantee differential privacy via an input-perturbation scheme in which the covariance matrix of a dataset $\mathbf{X} \in \mathbb{R}^{d \times n}$ is perturbed with a non-symmetric random Gaussian matrix with variance in $\mathcal{O}\left(\left(\frac{d}{n}\right)^2 \log d \right)$, thus improving upon the state-of-the-art. Furthermore, contrary to previous federated or distributed algorithms for PCA, our algorithm is also invariant to permutations in the incoming data, which provides robustness against straggler or failed nodes. Numerical simulations show that, while using limited-memory, our algorithm exhibits performance that closely matches or outperforms traditional non-federated algorithms, and in the absence of communication latency, it exhibits attractive horizontal scalability.

Dynamic SLA-based QoS control for third generation wireless networks: the CADENUS extension

Abstract: With the evolution of the QoS-capable third generation wireless networks, the wireless community has been increasingly looking for a framework that can provide an effective, network independent, end-to-end QoS control. In this paper, we first construct such a framework and then describe how dynamic SLA-based control can be used to achieve end-to-end QoS in a wired and wireless (UMTS) environment. The proposed framework, which is an extension of the IST CADENUS project, offers an effective wired-wireless QoS translation, an efficient QoS control and management, and a dynamic SLA policy-based QoS provisioning.

End to end reliable multicast transport protocol requirements for collaborative multimedia systems

Abstract: Multi-party collaborative multimedia applications require data to be transmitted reliably and efficiently in order to provide a guaranteed quality of service (QoS). The multimedia applications can vary from distributed games and shared whiteboards to interactive video conferencing. These applications often involve a large number of participants and are interactive in nature, with participants dynamically joining and leaving the applications. In order to provide many-to-many interaction when the number of participants is large, IP multicasting is a very good option for communication. IP multicasting provides scalability and efficient routing but does not provide the reliability that these multimedia applications may require. Though a lot of research has been done on reliable multicast transport protocols, it really seems that the only way of doing a reliable multicast is to build it for a given purpose like conference control in multimedia conferencing. This paper compares some of the available multicast transport protocols and analyses the most suitable features and functionalities provided by these protocols for a facet of conference control: floor control. The goal is to find or design a reliable multicast transport protocol which would scale to tens or hundreds of participants scattered across the Internet and which would deliver the control messages reliably.

Privacy trading in the surveillance capitalism age viewpoints on 'privacy-preserving' societal value creation

Abstract: In the modern era of the mobile apps (part of the era of surveillance capitalism, a famously coined term by Shoshana Zuboff), huge quantities of data about individuals and their activities offer a wave of opportunities for economic and societal value creation. However, the current personal data ecosystem is mostly de-regulated, fragmented, and inefficient. On one hand, end-users are often not able to control access (either technologically, by policy, or psychologically) to their personal data which results in issues related to privacy, personal data ownership, transparency, and value distribution. On the other hand, this puts the burden of managing and protecting user data on profit-driven apps and ad-driven entities (e.g., an ad-network) at a cost of trust and regulatory accountability. Data holders (e.g., apps) may hence take commercial advantage of the individuals' inability to fully anticipate the potential uses of their private information, with detrimental effects for social welfare. As steps to improve social welfare, we comment on the the existence and design of efficient consumer-data releasing ecosystems aimed at achieving a maximum social welfare state amongst competing data holders. In view of (a) the behavioral assumption that humans are 'compromising' beings, (b) privacy not being a well-boundaried good, and (c) the practical inevitability of inappropriate data leakage by data holders upstream in the supply-chain, we showcase the idea of a regulated and radical privacy trading mechanism that preserves the heterogeneous privacy preservation constraints (at an aggregate consumer, i.e., app, level) upto certain compromise levels, and at the same time satisfying commercial requirements of agencies (e.g., advertising organizations) that collect and trade client data for the purpose of behavioral advertising. More specifically, our idea merges supply function economics, introduced by Klemperer and Meyer, with differential privacy, that, together with their powerful theoretical properties, leads to a stable and efficient, i.e., a maximum social welfare, state, and that too in an algorithmically scalable manner. As part of future research, we also discuss interesting additional techno-economic challenges related to realizing effective privacy trading ecosystems.

I(FIB)F: Iterated bloom filters for routing in named data networks

Abstract: Named Data Networks provide a clean-slate redesign of the Future Internet for efficient content distribution. Because Internet of Things are expected to compose a significant part of Future Internet, most content will be managed by constrained devices. Such devices are often equipped with limited CPU, memory, bandwidth, and energy supply. However, the current Named Data Networks design neglects the specific requirements of Internet of Things scenarios and many data structures need to be further optimized. The purpose of this research is to provide an efficient strategy to route in Named Data Networks by constructing a Forwarding Information Base using Iterated Bloom Filters defined as I(FIB)F. We propose the use of content names based on iterative hashes. This strategy leads to reduce the overhead of packets. Moreover, the memory and the complexity required in the forwarding strategy are lower than in current solutions. We compare our proposal with solutions based on hierarchical names and Standard Bloom Filters. We show how to further optimize I(FIB)F by exploiting the structure information contained in hierarchical content names. Finally, two strategies may be followed to reduce: (i) the overall memory for routing or (ii) the probability of false positives.

I and i

Abstract: There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality. Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

Community detection in graphs

Abstract: The modern science of networks has brought significant advances to our understanding of complex systems. One of the most relevant features of graphs representing real systems is community structure, or clustering, i. e. the organization of vertices in clusters, with many edges joining vertices of the same cluster and comparatively few edges joining vertices of different clusters. Such clusters, or communities, can be considered as fairly independent compartments of a graph, playing a similar role like, e. g., the tissues or the organs in the human body. Detecting communities is of great importance in sociology, biology and computer science, disciplines where systems are often represented as graphs. This problem is very hard and not yet satisfactorily solved, despite the huge effort of a large interdisciplinary community of scientists working on it over the past few years. We will attempt a thorough exposition of the topic, from the definition of the main elements of the problem, to the presentation of most methods developed, with a special focus on techniques designed by statistical physicists, from the discussion of crucial issues like the significance of clustering and how methods should be tested and compared against each other, to the description of applications to real networks.