 |
 |
Applications are invited for 2 PhD scholarships in the Statistics Department of Trinity College Dublin in the area of Bayesian data analysis. These scholarships begin on 1st September 2013. They are awarded for one year initially, and are extended for a further 3 years upon satisfactory performance in the first year. The scholarship will pay for all European Union fees and give a tax-free stipend to cover living expenses. Expertise Required: Bachelor's degree or M.Sc. with a significant component of probability, statistics or machine learning. Some experience of theory and/or application of Bayesian methods, and programming (particularly for multi-core or multi-thread processing), is an advantage. Focus: One scholarship is available to work on each of the following projects: 1. Unstructured data analysis at scale. One component of the explosion in data volumes comes from unstructured data, the classic example being computer log files, that consist of unstructured text and numbers, and other forms of “peer to peer” data. In that context, such data contain information about the condition of a computer network, may identify security weaknesses that are being exploited, or allow prediction of network use and reliability. In this project, the question of how to map and summarise the data efficiently for analysis, and apply it to the various questions listed above, will be explored. 2. Bayesian methods for streaming data This project will look to apply and extend the methods that have already been developed for sequential Bayesian inference to streaming data at scale. The key requirements will be to propose and explore the properties of the necessary approximations that allow fast implementation of the sequential inference, tailored to the specific model that is chosen. A secondary goal will be to explore its implementation on multi-core processors. The position is part of the newly created Insight Research Centre. This is a multi-institutional research project funded by Science Foundation Ireland. Interested candidates can apply by submitting a CV to Simon Wilson at [log in to unmask] The deadline for submissions is 31st March 2013. Informal enquiries can be made to the same address. You may leave the list at any time by sending the command SIGNOFF allstat to [log in to unmask], leaving the subject line blank.