Stojan Trajanovski

Stojan Trajanovski

Auto-generated profile, based on publicly available information

Stojan Trajanovski

Microsoft, London, U.K.

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

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

Representative Publications

View All Publications
17 August 2020 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
SYSong Yang
Song Yang
FLFan Li
Fan Li
STStojan Trajanovski
Stojan Trajanovski
RYRamin Yahyapour
Ramin Yahyapour
XFXiaoming Fu
Xiaoming Fu
Network Function Virtualization (NFV) has been emerging as an appealing solution that transforms complex network functions from dedicated hardware implementations to software instances running in a virtualized environment. In this survey, we provide an overview of recent advances of resource allocation in NFV. We generalize and analyze four representative resource allocation problems, namely, (1) the VNF Placement and Traffic Routing problem, (2) VNF Placement problem, (3) Traffic Routing problem in NFV, and (4) the VNF Redeployment and Consolidation problem. After that, we study the delay calculation models and VNF protection (availability) models in NFV resource allocation, which are two important Quality of Service (QoS) parameters. Subsequently, we classify and summarize the representative work for solving the VPTR problem and the VRC problem by considering various QoS parameters (e.g., cost, delay, reliability and energy) and different scenarios (e.g., edge cloud, online provisioning and distributed provisioning). Finally, we conclude our survey with a short discussion on the state-of-the-art of literatures and emerging topics in the related field, and highlight areas where we expect high potential for future research.

25 September 2020 in IEEE Transactions on Biomedical Engineering
Verified
Journal Article
Closed access
STStojan Trajanovski
Stojan Trajanovski
CSCaifeng Shan
Caifeng Shan
PWPim J. C. Weijtmans
Pim J. C. Weijtmans
SKSusan G. Brouwer de Koning
Susan G. Brouwer de Koning
TRTheo J. M. Ruers
Theo J. M. Ruers
Objective: The utilization of hyperspectral imaging (HSI) in real-time tumor segmentation during a surgery have recently received much attention, but it remains a very challenging task. Methods: In this work, we propose semantic segmentation methods and compare them with other relevant deep learning algorithms for tongue tumor segmentation. To the best of our knowledge, this is the first work using deep learning semantic segmentation for tumor detection in HSI data using channel selection and accounting for more spatial tissue context and global comparison between the prediction map and the annotation per sample. Results and Conclusion: On a clinical data set with tongue squamous cell carcinoma, our best method obtains very strong results of average dice coefficient and area under the ROC-curve of 0.891 +/- 0.053 and 0.924 +/- 0.036, respectively on the original spatial image size. The results show that a very good performance can be achieved even with a limited amount of data. We demonstrate that important information regarding tumor decision is encoded in various channels, but some channel selection and filtering is beneficial over the full spectra. Moreover, we use both visual (VIS) and near-infrared (NIR) spectrum, rather than commonly used only VIS spectrum; although VIS spectrum is generally of higher significance, we demonstrate NIR spectrum is crucial for tumor capturing in some cases. Significance: The HSI technology augmented with accurate deep learning algorithms has a huge potential to be a promising alternative to digital pathology or a doctors' supportive tool in real-time surgeries.

4 October 2013 in Physical Review E
Verified
Journal Article
Closed access
DGDongchao Guo
Dongchao Guo
STStojan Trajanovski
Stojan Trajanovski
RBRuud van de Bovenkamp
Ruud van de Bovenkamp
HWHuijuan Wang
Huijuan Wang
PMPiet Van Mieghem
Piet Van Mieghem
The interplay between disease dynamics on a network and the dynamics of the structure of that network characterizes many real-world systems of contacts. A continuous-time adaptive susceptible-infectious-susceptible (ASIS) model is introduced in order to investigate this interaction, where a susceptible node avoids infections by breaking its links to its infected neighbors while it enhances the connections with other susceptible nodes by creating links to them. When the initial topology of the network is a complete graph, an exact solution to the average metastable-state fraction of infected nodes is derived without resorting to any mean-field approximation. A linear scaling law of the epidemic threshold ${$\tau${}}_{c}$ as a function of the effective link-breaking rate $$\omega${}$ is found. Furthermore, the bifurcation nature of the metastable fraction of infected nodes of the ASIS model is explained. The metastable-state topology shows high connectivity and low modularity in two regions of the $$\tau${},$\omega${}$ plane for any effective infection rate $$\tau${}>{$\tau${}}_{c}$: (i) a ``strongly adaptive'' region with very high $$\omega${}$ and (ii) a ``weakly adaptive'' region with very low $$\omega${}$. These two regions are separated from the other half-open elliptical-like regions of low connectivity and high modularity in a contour-line-like way. Our results indicate that the adaptation of the topology in response to disease dynamics suppresses the infection, while it promotes the network evolution towards a topology that exhibits assortative mixing, modularity, and a binomial-like degree distribution.

26 March 2013 in Journal of Complex Networks
Verified
Journal Article
Open access
PDF file
STStojan Trajanovski
Stojan Trajanovski
JMJavier Martín-Hernández
Javier Martín-Hernández
WWWynand Winterbach
Wynand Winterbach
PMPiet Van Mieghem
Piet Van Mieghem
We study the robustness of networks under node removal, considering random node failure, as well as targeted node attacks based on network centrality measures. Whilst both of these have been studied in the literature, existing approaches tend to study random failure in terms of average-case behavior, giving no idea of how badly network performance can degrade purely by chance. Instead of considering average network performance under random failure, we compute approximate network performance probability density functions as functions of the fraction of nodes removed. We find that targeted attacks based on centrality measures give a good indication of the worst-case behavior of a network. We show that many centrality measures produce similar targeted attacks and that a combination of degree centrality and eigenvector centrality may be enough to evaluate worst-case behavior of networks. Finally, we study the robustness envelope and targeted attack responses of networks that are rewired to have high- and low-degree assortativities, discovering that moderate assortativity increases confer more robustness against targeted attacks whilst moderate decreases confer more robustness against random uniform attacks.

30 January 2023 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
NHNan He
Nan He
SYSong Yang
Song Yang
FLFan Li
Fan Li
STStojan Trajanovski
Stojan Trajanovski
Liehuang ZhuLiehuang Zhu
Liehuang Zhu
XFXiaoming Fu
Xiaoming Fu
The efficacy of Network Function Virtualization (NFV) depends critically on (1) where the virtual network functions (VNFs) are placed and (2) how the traffic is routed. Unfortunately, these aspects are not easily optimized, especially under time-varying network states with different QoS requirements. Given the importance of NFV, many approaches have been proposed to solve the VNF placement and Service Function Chaining (SFC) routing problem. However, those prior approaches mainly assume that the network state is static and known, disregarding dynamic network variations. To bridge that gap, we leverage Markov Decision Process (MDP) to model the dynamic network state transitions. To jointly minimize the delay and cost of NFV providers and maximize the revenue, we first devise a customized Deep Reinforcement Learning (DRL) algorithm for the VNF placement problem. The algorithm uses the attention mechanism to ascertain smooth network behavior within the general framework of network utility maximization (NUM). We then propose attention mechanism-based DRL algorithm for the SFC routing problem, which is to find the path to deliver traffic for the VNFs placed on different nodes. The simulation results show that our proposed algorithms outperform the state-of-the-art algorithms in terms of network utility, delay, cost, and acceptance ratio.

6 June 2012 in Physical Review E
Verified
Journal Article
Closed access
STS. Trajanovski
S. Trajanovski
SSS. Scellato
S. Scellato
ILI. Leontiadis
I. Leontiadis
The study of real-world communication systems via complex network models has greatly expanded our understanding on how information flows, even in completely decentralized architectures such as mobile wireless networks. Nonetheless, static network models cannot capture the time-varying aspects and, therefore, various temporal metrics have been introduced. In this paper, we investigate the robustness of time-varying networks under various failures and intelligent attacks. We adopt a methodology to evaluate the impact of such events on the network connectivity by employing temporal metrics in order to select and remove nodes based on how critical they are considered for the network. We also define the temporal robustness range, a new metric that quantifies the disruption caused by an attack strategy to a given temporal network. Our results show that in real-world networks, where some nodes are more dominant than others, temporal connectivity is significantly more affected by intelligent attacks than by random failures. Moreover, different intelligent attack strategies have a similar effect on the robustness: even small subsets of highly connected nodes act as a bottleneck in the temporal information flow, becoming critical weak points of the entire system. Additionally, the same nodes are the most important across a range of different importance metrics, expressing the correlation between highly connected nodes and those that trigger most of the changes in the optimal information spreading. Contrarily, we show that in randomly generated networks, where all the nodes have similar properties, random errors and intelligent attacks exhibit similar behavior. These conclusions may help us in design of more robust systems and fault-tolerant network architectures.