The project involves applying a conceptual modeling method called R2M, Role and Request Modeling. The innovative method, and its supporting software (R2M]ST), was developed at the University of British Columbia. R2M assists business analysts to accurately capture organizational roles, interactions, services and IT resources as described by business process owners and abstract the information that can be used by decision makers. Although there has been significant industry interest, the method has not been tested in more complex small to medium sized enterprises (SME).
This work can serve as the foundation of relative value strategies and portfolio sensitivity analysis. This project research on building a strategy for analyzing the sensitivity of mortgage backed securities (MBS) pool pricing to changes in financial market and underlying security characteristics, such as the property value of the underlying real estate. There is a large database including real estate property location, real estate price indices by region and property type, as well as mortgage security pool information.
Weyes Eyes, Inc. is developing a product that uses intelligent video analytics software to alert end users to pre-configured events as they occur in video streams from cameras set up in the home or office. The user will be able to specify which types of events to capture (doors opening, people walking by, etc.) and the software will detect these events using sparse feature-tracking and objectrecognition methods. Many state-of-the-art feature detectors and descriptors may be suitable for this task, and Weyes Eyes wishes to know which ones work best for their specific application.
Integration is a core IT operation, and is aided by a number of available best-practice techniques for integration (integration patterns). However the application of these patterns has little to no automated support. They are applied by consultants on a per-customer basis, making it an expensive and time- consuming task. A computer science student from the University of Toronto will work with IBM at their Toronto Center for Advanced Studies on building a set of heuristics to aid the understanding and the application of integration patterns.
The goal of the project is to improve game load times on an electronic gaming machine (EGM). Currently the bottleneck in loading a new game is the processing of game data read from memory when a new game is selected by the user. The current software based method of processing bogs down the EGM. The student will develop a processing technique that uses specialized computer hardware (FPGA)technology and will not have a detrimental effect on the EGM. The design will make use of existing computer hardware resources in the EGM and therefore must be designed with hardware resource usage in mind.
The project shall evaluate and extend the state-of-the-art in document clustering according to semantics extracted from natural language documents. This will require testing of current methods to identify their limitations and proposal of new methods based on empirical observations. Two complementary techniques shall be evaluated: methods for incremental taxonomy growth and for calculating semantic similarity among documents.
WiMAX (Worldwide Interoperability for Microwave Access) is an innovative communication technology for broadband wireless access to the Internet. Because of high bandwidth and large coverage, WiMAX has experienced exponential growth over the past years. So far, many telecommunication companies around the world have started to offer WiMAX services. Despite the popularity of WiMAX, the reliability of WiMAX-based transmission in the Internet (end-to-end data transmission involving WiMAX links) has not been thoroughly studied.
The classification of acoustic signals whose factors of variation are due to different atmospheric and sound propagation effect is a challenging problem. The internship will explore new learning algorithms for this application, which have the potential to capture some of the complex structure in the data.
Fast structural variation (deletion, insertion and inversion) detection between genome of different individuals is the main goal of this project. The internship team is planning in extending new algorithms to reduce the number of false positive calls (especially for deletions) and to parallelize it using Graphics Processing Unit (GPU). The standard approach implementation of the algorithms, as a result of high computational needs, is not fast enough for every day use by health science centers (such as hospitals).