Scholarships and funding

Deep Convolutional Neural Networks (ConvNets) have demonstrated state-of-the-art performance in many machine learning problems involving image classification and speech recognition. Over the last few years several advances in the design of ConvNets have not only led to a further boost in achieved accuracy on image recognition tasks but also played a crucial role as a feature extractor for other tasks such as object detection, localization, semantic segmentation and image retrieval. However, the complexity and size of ConvNets have limited their use in mobile applications and embedded systems.

This PhD studentship will investigate ways to optimize these deep neural networks using model-architecture co-design and enable mass deployment of deep-learning based applications in consumer products.

The PhD student will design architectures for hardware convolution engines for scenarios with limited hardware resources and tight power and latency constraints. The student will also investigate automated tools to solve the difficult problem of designing neural networks under complexity constraints and will describe a design-space-exploration tool that automatically discovers good neural network models with efficient hardware implementations.

Contact: Professor Klaus McDonald-Maier (kdm@essex.ac.uk), Dr Xiaojun Zhai (xzhai@essex.ac.uk), and Dr Shoaib Ehsan (sehsan@essex.ac.uk)

Closing date: 11 January 2019.

For the next level of robot intelligence, maps need to extend beyond geometry and appearance — they need to contain semantics. The inclusion of rich semantic information within a dense map enables a much greater range of functionality than geometry alone.

The PhD student will be investigating how knowledge from visual SLAM and machine-learned labelling can be brought together to enable powerful semantic and object-aware mapping in extreme environments. Recent works on dense semantically annotated 3D maps of indoor scenes hint at the improvements possible with significantly longer trajectories, such as those of an autonomous robot in a nuclear facility, making direct use of the semantically annotated 3D map.More specifically, the PhD student will tackle this problem using visual SLAM approach as it is a key enabler for both ground-based and aerial platforms (such as drones).

The PhD student will employ event cameras, as they have enormous potential for fast and low power vision algorithms for robots, and will develop new algorithms for solving the challenging problem of visual SLAM. The PhD student will perform joint estimation of 3D scene structure, 6-DoF camera motion and up to scale scene intensity from a single event camera moving in an unstructured extreme environment of which it has no prior knowledge.

Contact: Professor Klaus McDonald-Maier (kdm@essex.ac.uk), and Dr Shoaib Ehsan (sehsan@essex.ac.uk)

Closing date: 11 January 2019.

The expansion of the Internet has brought about an exciting research area full of unexplored, well-motivated research directions. Its use as a primary computing platform has changed the way computation is now performed. In many contexts, reaching an outcome or a collective decision depends on information that is privately controlled by several agents, who have an interest in the actual outcome that will be realised. As a result, any individual agent might benefit, to some extent, by falsely reporting his private information. This, in turn, can greatly affect the performance of the overall system.

The goal of this project is to explore settings of strategic interaction in the intersection of computer science and economics: To what extent does the behaviour of the individual agents affect the overall outcome? Can the agents be motivated towards a more desirable behaviour? Algorithmic Game Theory is a perfect tool to model and analyse such competitive situations. It can be applied to examine the incentives of users in several different contexts like scheduling, block-chain applications, or finance, among others.

Supervisors:

Dr Maria Kyropoulou (maria.kyropoulou@essex.ac.uk)

Dr Carmine Ventre (c.ventre@essex.ac.uk)

Closing date - Friday 11 January 2019.

Interviews for shortlisted candidates are expected to take place during week commencing Monday 21 January 2019.