Potential Research Areas
For the academic year 2013-2014 our department is recruiting bright students interested particularly in the following research areas (listed in alphabetical order):
Autonomous Underwater Vehicles: The Memorial Explorer AUV is a large and capable underwater vehicle that is used for a variety of scientific missions. We are investigating ways of increasing the vehicle's autonomy so that it can remain underwater for longer periods and process imagery data from the seabed. Research topics include autonomous navigation and underwater imaging.
Bioinformatics: Develop and apply computational approaches to better understand biological and biomedical processes using 'omics data (such as genomics, proteomics, and metabolomics). Research in this area aims to develop and improve computational approaches to analyze and integrate complex data sets, and to handle data generated by new experimental biotechnology.
Bio-inspired computing: It's main premise is that life, essentially, is information processing and that man-made systems used for information processing can benefit from concepts and recipes nature developed over eons. The techniques learned through bio-inspiration can in turn be applied to the examination of very complex scientific and technical systems. Our research topics on methods include "learning and adaptation", "evolution and optimization", "intelligence and decision support", "complex (adaptive) systems", collective behavior, genetic regulation and their application in pattern recognition, data mining, computational finance, bioinformatics, and Science and Engineering in general.
Cheminformatics: Computational chemists generate an enormous amount of chemical structure information on a daily basis, and robust methods for search and retrieval of this information is essential. Research in this area can range from development of metrics to uniquely identify molecules, to development of algorithms to find "similar" structures or substructures, to visualization methods for 3D structures.
Cloud computing: To gain widespread enterprise level acceptance, cloud providers need to address several challenging demands: safeguard data confidentiality, provide concurrent and reliable access to data via transactions (as with traditional databases) and guarantee user anonymity. Our current work aims to integrate the above objectives providing transactional guarantees over encrypted outsourced data - more specifically, designing concurrency control protocols for the different cloud sharing configurations, with different confidentiality requirements.
Cognitive Science: Computational complexity analysis is used to assist in both formulating robust psychological theories of and evaluating algorithmic options for AI-system implementations of cognitive human abilities. Potential topics include such analyses relative to analogy-based cognitive processing, language acquisition, and problem solving.
Computational Combinatorics: The area often entails such things as the enumeration and analysis of combinatorial objects such as graph colourings, combinatorial designs, etc. It also includes algorithmic aspects of combinatorics, ranging from the design of algorithmic construction techniques to establishing results about computational complexity.
Computational Modelling in Physiology: Modelling has become, and will continue to be, an important tool to investigate physiological phenomena in the health and life sciences. Current topics of interest include: (i) the generation of thrombin during blood coagulation, particularly in children; and (ii) intracellular calcium dynamics in cardiac Purkinje cells. Research in both of these areas is necessarily interdisciplinary and is conducted in close collaboration with researchers in the related fields.
Computer Graphics: Study how to generate and manipulate images and videos using computers. Possible topics include image and video editing, geometry modeling, image-based rendering, programming on graphics hardwares, etc.
Computer Vision: Investigate novel techniques for processing, analyzing, and understanding images and videos captured from the real world. Research topics include, but are not limited to, stereo vision, motion estimation, and foreground segmentation.
Data Integration: Integration of multiple data sources is required in many different application areas, such as e-medicine or e-healthcare, e-government, business intelligence and data analytics, enterprise application integration and others. Data integration can include structured information, such as that stored in databases or ontologies, as well unstructured information, e.g. collected from web pages. Data integration encompasses the entire process of identifying data sources, preparing and transforming data and data sources, matching or aligning data schemata, and implementing logical and physical integration mechanisms.
Diagnostic Imaging: Diagnostic or Medical Imaging (DI) comprises the set of related technologies and programs that are used to interrogate human anatomy and physiology by generating a projection of a region of interest. The objective of the DI procedure is to allow the physician to render a diagnosis or to monitor a therapy. In a perfect world these procedures would be perfectly sensitive and perfectly specific. Often, this is not the case. Disease may be missed or misdiagnosed. Minimizing these errors remains the objective of DI research.
Hardware and Software Verification: How can descriptions of hardware and software systems be proved to be free of errors? Hardware and software verification seeks to apply logic and mathematics to answering this question. We are particularly focused on methods that apply to parallel programs, where software testing is least effective.
Modeling and performance evaluation of concurrent systems: Timed Petri nets are used to model the behavior of concurrent systems with temporal attributes reprtesenting the durations of system's activities. Different techniques are then used to obtain performance indices of timed models. For some models, structural analysis is sufficient for performance analysis. If a model is bounded and its state space is reasonably small, enhaustive reachability analysis can be used. Performance measures can also be estimated using discrete-event simulation of timed models. A wide variety of systems can be studied in this way, including computer and communication networks, multiprocessor systems, transportation networks, manufacturing systems and so on.
Privacy-Preserving Data Analysis for Health Research: Provide a common framework for the secure computation of health data including secure building blocks and privacy-preserving protocols for the set of common and widely used statistical analysis and data mining methods in health research. Using this framework data users are able to run the queries on the encrypted data without having access to the original raw data, which could be kept off the network after the encrypted data is generated in the setup phase. The framework will also be integrated with statistical software, like SAS, which is widely used for statistical analyses on data in many research centres and health units.
Robotics: We use computational complexity analysis to evaluate various types of algorithms for designing, modifying, and verifying groups composed of one or more robots. Potential topics include such analyses relative to purely reactive or hybrid architecture robots (see also Swarm Robotics below).
Services Computing: Services can be described in an implementation-independent, semantic fashion. They are published in generally accessible web repositories, can be composed into new (more complex) services and executed by referring back to the service providers. "Black box" services are gradually replaced by services whose executions can be monitored and controlled (in a limited way) by consumers. Our research aims at incorporating transactional properties for such services to enable flexible and dynamic compositions.
Swarm Robotics: Study how a swarm of simple robots are able to collectively complete a task. Interesting tasks range from aggregation and sorting to collective construction. Research in swarm robotics is inspired by the self-organizing abilities of social insects such as ants and bees, but also involves practical knowledge of robotic systems.
Transactional Memory: Memory operations are grouped into atomic units. Demarcation into atomic units is intended to free the programmers from explicit synchronization of different units using locks, critical sections, etc. Nesting of the atomic units occurs naturally, with procedure calls. Currently we are working on characterizing correct concurrent executions of atomic units and on compositional properties. The ultimate goal is to simplify concurrent programming and to increase the speed of computation in multi-core machines.
Web Service Repositories: They serve as a registry for service specifications and allow formal reasoning about these services. We investigate questions arising during the design and engineering of such repositories. Prototype repositories are implemented and benchmark testing performed. Currently we are particularly interested in web repository engineering processes, quality of service measurements, and tools which aid developers during repository design, configuration, testing, implementation and management.
Wireless networking and mobile computing: A mobile ad hoc network consists of solely mobile hosts and dispenses with infrastructure. With limited transmission range and continuous mobility of the hosts, many computer networking issues become remarkably more interesting, such as routing, data forwarding, topology management, link adaptation, etc. Furthermore, network-wide end-to-end connectivity is never achievable; and yet communication among devices is still desired. This can be facilitated by exploring techniques of delay-tolerant networking. Innovative niche applications in environment monitoring, social networking and deep-space communications provide essential thrust to motivate our research.