StatScale Research Associates (University of Cambridge)
Applications are invited for three Research Associate positions available in the Department of Pure Mathematics and Mathematical Statistics to work in the area of Statistics.
Duties include developing and conducting individual and collaborative research objectives, proposals and projects. The role holders will be expected to plan and manage their own research and administration, with guidance if required. They must be able to communicate material of a technical nature and be able to build internal and external contacts. They may be asked to assist in the supervision of student projects, the development of student research skills, provide instruction or plan/deliver seminars relating to the research area. The successful candidates will be working on one of three projects, as per below.
Post A (StatScale, with Dr Rajen Shah)
StatScale is an exciting collaboration between researchers at the University of Cambridge and Lancaster University. The programme will tackle the important inference challenges arising from large-scale data and data streams. This multi-million pound research initiative is funded by the Engineering and Physical Sciences Research Council (EPSRC), the two participating institutions and a number of committed project partners. The programme is led by Idris Eckley, Paul Fearnhead (Lancaster) and Rajen Shah, Richard Samworth (Cambridge). You should have, or be close to completing, a PhD in Statistics or a related discipline. You will be experienced in one or more of the following areas: high-dimensional statistics, nonparametric statistics and streaming data. Demonstrable ability to produce academic writing of the highest publishable quality is essential. This position would be supervised by Dr Rajen Shah at the University of Cambridge.
Post B (StatScale, with Professor Richard Samworth)
Project as above. This position would be supervised by Professor Richard Samworth at the University of Cambridge.
Post C (EPSRC and StatScale, with Professor Richard Samworth)
The project is primarily funded by Professor Samworth's EPSRC fellowship 'Statistical Methodology and Theory in the Big Data Era'. You should have, or be close to completing, a PhD in Statistics or a related discipline; you will also have demonstrated a capacity to produce world-class research. Priority will be given to applicants with expertise in relevant fields such as high-dimensional and nonparametric statistical inference.
Fixed-term: The funds for these posts are available for 3 years in the first instance.
Informal inquiries can be made by contacting or apply
StatScale PhD Studentship (Lancaster University)
Streaming Identification of Dependency Networks in Complex Systems
Technological advances mean that we can measure complex systems at increasing resolution and frequency. A popular way to visualise, model, and interpret activity in such complex systems is to associate variables with a probabilistic graphical model and underlying dependency network, for instance, the behaviour of one subset of variables may depend only on a few other regions. Unfortunately, this dependency network is not directly observable and needs to be inferred, furthermore, we must often consider that the network may change structure over time, and thus the data-generating process may be non-homogeneous. This PhD project will consider the challenging statistical and computational task of how to specify, and then estimate such dependency networks (and their evolution) when data is ingested in a streaming, i.e. sequential manner. Estimation in this online setting requires urgent research in order to understand the multiple trade-offs between computational cost, model-complexity, and both optimisation (computational) and statistical estimation error.
Depending on the applicants preference this project may touch on more applied or theoretical aspects of the dynamic graph estimation problem. Methodologically, this may encompass topics such as: locally-stationary wavelet models, regularisation in the context of changepoints and/or model-selection, development of new scalable optimisation algorithms, and analysis of their solutions in the presence of statistical error. On the application side, there are many directions which can be explored, however, neuroscience and finance represent two key domains where these methods are rapidly gaining influence. For instance, we may attempt to understand how brain connectivity is associated with different tasks, or in finance, how dependency structures between asset classes can impact portfolio construction.
A full studentship is available for 3.5 years, covering fees plus maintenance grant of approximately £16,000 pa tax free for those who count as Home/EU students for the purposes of fees. Unfortunately, non-UK/EU applicants would need to provide funding to cover the difference between home and overseas fee rates.
A first-class degree, or masters degree (or equivalent) in an appropriate subject.
Deadline for Applications:
Applications will be reviewed as they are received with an initial closing date of 14th January, or until positions are filled. The studentship will start at a mutually agreed date, but ideally no later than October 2019.
Please send a cover letter, CV and undergraduate/masters transcript(s), to firstname.lastname@example.org. The cover letter should clearly explain your motivation for applying for this specific studentship.