A5011451698- Auction Theory and Applications
- Metaheuristic Optimization Algorithms Research
- Evolutionary Algorithms and Applications
- Music Technology and Sound Studies
- Data Visualization and Analytics
- Music and Audio Processing
- Neural Networks and Applications
- Advanced Multi-Objective Optimization Algorithms
- Complex Systems and Time Series Analysis
- Neural Networks and Reservoir Computing
- Cellular Automata and Applications
- Nonlinear Dynamics and Pattern Formation
- Consumer Market Behavior and Pricing
- Modular Robots and Swarm Intelligence
- Innovative Human-Technology Interaction
- Bone fractures and treatments
- Reinforcement Learning in Robotics
- Game Theory and Applications
- Advanced Software Engineering Methodologies
- Slime Mold and Myxomycetes Research
- Anatomy and Medical Technology
- Game Theory and Voting Systems
- Hand Gesture Recognition Systems
- Retinal and Optic Conditions
- Supply Chain and Inventory Management
St. George's University
2023
University of Toledo
2022
Princess Royal Hospital
2020
Studix
2015-2016
University of Oslo
2011-2015
University of Birmingham
2004-2010
Novel computing systems are increasingly being composed of large numbers heterogeneous components, each with potentially different goals or local perspectives, and connected in networks which change over time. Management such quickly becomes infeasible for humans. As such, future should be able to achieve advanced levels autonomous behaviour. In this context, the system's ability self-aware self-express important. This paper surveys definitions current understanding self-awareness...
We propose a novel framework for efficient parallelization of deep reinforcement learning algorithms, enabling these algorithms to learn from multiple actors on single machine. The is algorithm agnostic and can be applied on-policy, off-policy, value based policy gradient algorithms. Given its inherent parallelism, the efficiently implemented GPU, allowing usage powerful models while significantly reducing training time. demonstrate effectiveness our by implementing an advantage actor-critic...
Work on human self-awareness is the basis for a framework to develop computational systems that can adaptively manage complex dynamic tradeoffs at runtime. An architectural case study in cloud computing illustrates framework's potential benefits.
Modern compute systems continue to evolve towards increasingly complex, heterogeneous and distributed architectures. At the same time, functionality performance are no longer only aspects when developing applications for such systems, additional concerns as flexibility, power efficiency, resource usage, reliability cost becoming important. This does not raise question of how efficiently develop but also cope with dynamic changes in application behaviour or system environment. The EPiCS...
To solve multi-objective problems, multiple reward signals are often scalarized into a single value and further processed using established single-objective problem solving techniques. While the field of optimization has made many advances in applying scalarization techniques to obtain good solution trade-offs, utility these multi-agent learning domain not yet been thoroughly investigated. Agents learn their decisions by linearly scalarizing at local level, while acceptable system wide...
We study heterogeneity among nodes in self-organizing smart camera networks, which use strategies based on social and economic knowledge to target communication activity efficiently. compare homogeneous configurations, when cameras the same strategy, with heterogeneous different strategies. Our first contribution is establish that static leads new outcomes are more efficient than those possible homogeneity. Next, two forms of dynamic investigated: nonadaptive mixed adaptive strategies, learn...
In this paper we study the self-organising behaviour of smart camera networks which use market-based handover object tracking responsibilities to achieve an efficient allocation objects cameras. Specifically, compare previously known homogeneous configurations, when all cameras same marketing strategy, with heterogeneous each makes its own, possibly different strategy. Our first contribution is establish that such heterogeneity strategies can lead system wide outcomes are Pareto superior...
To assess the proportion of patients with distal radius fractures (DRFs) who were managed nonoperatively during COVID-19 pandemic in accordance British Orthopaedic Association BOAST guidelines, would have otherwise been considered for an operative intervention.We retrospectively reviewed radiographs and clinical notes all DRFs nonoperatively, following publication guidelines on management urgent trauma between 26 March 18 May 2020. Radiological parameters including radial height,...
In this paper, we introduce a novel gesture recognition algorithm named the ant learning (ALA), which aims at eliminating some of limitations with current leading algorithms, especially Hidden Markov Models. It requires minimal training instances and greatly reduces computational overhead required by both classification. ALA takes advantage pheromone mechanism from colony optimization. uses tables to represent gestures, scales well complexity. Our experimental results show that can achieve...
A system for decentralized synchronization of musical agents is presented, inspired by Mirollo and Strogatz' pulse-coupled oscillator model the synchronous flashing certain species firefly. While most previous work on oscillators assume fixed (close to) equal frequencies, presented tackles challenge different starting frequencies. Open source implementations in Puredata, Max, Matlab are provided. Test results setups six nodes show that reach a state harmonic synchrony, where fire events...
Non-arteritic ischemic optic neuropathy (NAION) is a common cause of acute, painless monocular vision loss in adults older than 50. NAION diagnosis exclusion established once arteritic disease and other etiologies acute have been ruled out. Clinicians need to distinguish from (AION) since failing appropriately treat patients presenting with AION results an inferior prognosis. often associated risk factors like obstructive sleep apnea, atherosclerosis, diabetes mellitus, hypertension,...