Dr Adam M Sykulski
|Themes||Multivariate and High Dimensional Data, Stochastic Modelling and Time Series|
I am a statistician specialising in time series, stochastic process and spatiotemporal data. I am particularly interested in processes that are nonstationary, locally stationary, anisotropic and multivariate/high-dimensional. I have a particular application focus in modelling oceanographic time series obtained from floats and drifters. Please do get in touch if you have overlapping research interests.
I am currently an EU-funded Marie Curie Research Fellow. I am based at NorthWest Research Associates in Seattle, USA - but am collaborating closely with the UCL Department of Statistical Science, where I return for the final year of my fellowship in April 2016.
My research focuses primarily on developing semi-parametric modelling and estimation techniques for multivariate data. I have a particular interest in frequency-domain modelling, and estimation of parameters via the Whittle Likelihood (see recent papers below).
I have applied my research to oceanographic data. In particular, I am implementing our methods on data from the Global Drifter Program: a large database of worldwide satellite-tracked freely-drifting instruments known as 'drifters'. Our new techniques allow us to make insightful new findings which improves climate modelling and our ability to respond to environmental threats such as oil spills.
I also have an active research interest (from my PhD) in decision theory problems related to the multi-armed bandit problem. In particular, I am interested in on-line tuning of the exploration-exploitation tradeoff and also extensions to multi-player problems which bring in ideas from game theory.
- JSM 2014, "Removing Bias from Whittle Estimators", Tuesday 5th August 2014, 9.35am, Session 299 (Nonlinear Models and Other Optimization Problems), Boston Convention Center, USA.
- RSS 2014, "Semi-parametric inference of time series using the Whittle Likelihood", Tuesday 2nd September 2014, 4.30pm, Session 4.2 (Time Series Session), Sheffield City Hall, UK.
Time Series Analysis, Spatiotemporal Processes, Nonstationarity, Anisotropy, Applications in Oceanography, Decision Theory, Game Theory, Tennis.
Selected publications and preprints
- Sykulski, AM, Olhede, SC, Lilly, JM and Danioux, E (2013) Lagrangian Time Series Models for Ocean Surface Drifter Trajectories. http://arxiv.org/abs/1312.2923
- Sykulski, AM, Olhede, SC, Lilly, JM and Early, JJ (2013) The Whittle Likelihood for Complex-Valued Time Series. http://arxiv.org/abs/1306.5993
- Sykulski, AM, Adams, NM and Jennings, NR (2010) On-line adaptation of exploration in the one-armed bandit with covariates problem. In: 9th International Conference on Machine Learning and Applications (ICMLA 2010), 12th-14th Dec 2010, Washington DC, USA. pp. 459-464. http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5708871&tag=1
- Munoz de Cote, E., Sykulski, A.M., Chapman, A.C. and Jennings, N.R. (2010) Automated planning in repeated adversarial games. In: 26th Conference on Uncertainty in Artificial Intelligence (UAI 2010), 8th-11th July 2010, Catalina Island, California, USA. pp. 376-383. http://arxiv.org/abs/1203.3498
- Olhede, SC, Sykulski, AM and Pavliotis, GA (2009) Frequency domain estimation of integrated volatility for Ito processes in the presence of market-microstructure noise. SIAM - Multiscale Modeling and Simulation, 8 (2). pp. 393-427. http://epubs.siam.org/doi/abs/10.1137/090756363
Page last modified on 26 jun 14 11:35