Who We Are
Nils Lid Hjort is professor of mathematical statistics at the Department of Mathematics at the University of Oslo. His research interests span a broad range of topics, from foundations and methodology to applications. His books include Highly Structured Stochastic Systems (with P.J. Green and S. Richardson, 2003), Model Selection and Model Averaging (with G. Claeskens, 2008), Bayesian Nonparametrics (with C. Holmes, P. Müller and S.G. Walker, 2010), and Confidence, Likelihood, Probability: Inference With Confidence Distributions (with T. Schweder, 2016). Lists of publications, talks and other contributions may be found via the Google Scholar and Cristin profiles. Hjort was the first recipient of the Ludwig von Drake award and is a winner of the Sverdrup Prize. He is an elected member of the Norwegian Academy of Science and Letters and of the Royal Norwegian Society for Sciences and Letters. His work has contributed to changing the Olympics. Hjort has edited two volumes with Don Rosa (one and two), and sings with Unikor.
Gudmund Horn Hermansen: He will work on topics within Bayesian nonparametrics, along with themes related to focused inference and model selection for time series processes and change-point problems. Gudmund completed his PhD in statistics from the University of Oslo in September 2014, with Nils LH and Arnoldo Frigessi as supervisors. During his PostDoc period with us, from October 2014 to April 2017, he also maintained a part-time position and then a fuller-time research position at the Norwegian Computing Centre (NR), working with applications in statistical geology, machine learning, and yet further application domains. He has been instrumental in building up the Oslo Data Science Meetup. Since August 2017, he has taken up a first amanuensis position with the BI Norwegian Business School in Oslo, at the Department of Economics, and since October 2017 he also has an amanuensis-II position (one day a week) with the Department of Mathematics. Gudmund is a high-level juggler.
Kristoffer Herland Hellton is our second PostDoc (and our second cellist), starting out with us in May 2015. His PhD work involved multivariate analysis for data characterised by both high dimensions and measurement errors, where one of the goals is to exploit statistical genetics for improving cancer treatments. Since October 2016 he has taken up a position as a senior research statistician with the Norwegian Computing Centre, while maintaining a 20% position with us at FocuStat. He plays the cello in various chamber music ensembles.
(The unruly crowd of PostDocs, PhDs, Master students occasionally manages to disturb the professorial ongoings.)
- Céline Marie Løken Cunen (from August 2014), with Nils LH and Bo Lindqvist as supervisors: She has worked on methodological aspects concerning the use of confidence distributions for combination of information from different sources. The project at one stage also involved an application in evolutionary biology, concerning the estimation of the diversification rate in groups of species over evolutionary time, but later on turned its attention to wars and whales. Céline has a bachelor's degree in biology and a master's degree in applied statistics, both from the Norwegian University of Life Sciences. Her PhD dissertation Wars and Whales: Extension and Applications of Confidence Curves and Focused Model Selection was completed in June 2019, and she very successfully defended her work in a public disputation December 2019. During her active PhD years she has given talks on three different continents, and her work on the Wars of the Roses has been picked up in a dozen languages. She plays the cello in the University Symphony Orchestra.
- Sam-Erik Walker (from October 2014), with Nils LH as main supervisor: He will work on using robust and computationally convenient M-estimators for focussed statistical inference in various contexts. The project will also involve an application in ensemble based probabilistic forecasting of air pollution for Oslo. Sam-Erik has a bachelor's degree in computer science and a master's degree in statistics, the latter on probabilistic modelling of air pollution from road traffic, both from the University of Oslo. He has more than 25 years of work experience from the Norwegian Institute for Air Research (NILU).
- Emil Aas Stoltenberg (from August 2015), with Sven Ove Samuelsen and Nils LH as superviors. His Master project with Nils LH was completed in June 2015, with thesis title "The c-Loss Function: Balancing Total and Individual Risk in the Simultaneous Estimation of Poisson Means". It concerns simultaneous estimation of an ensemble of Poisson parameters, using different loss functions, where certain estimation strategies uniformly dominate the maximum likelihood procedure. He also works with empirical Bayes methods, yielding estimators that nonparametrically smooth towards e.g. given Poisson regression models. Emil holds a previous Master's degree in political science, with thesis title "Bayesian Forecasting of Election Results in Multiparty Systems". His political forecasting work put him in the news, and in several television programmes he was educating the Norwegian population on the ins and outs of the US presidential election processes when Trump won. Emil plays basketball at a high national level, and has contributed to Nordberg Basket's phenomenal rise through the divisional system from 2011 to 2016. Emil is since August 2015 a PhD student with the PharmaTox research programme. The project relates to possible long-term effects of pharmaceuticals on neurodevelopment. Also other research directions will be worked with within his PhD project, including estimation in multi-parameter Poisson models, modelling and inference for on-off data, and inference for stochastic volatility.
- Vinnie Ko (from August 2016), with Nils LH and Ingrid Hobæk Haff as supervisors: His work has involved pair-copula constructions for multiple dependence, including aspects of both estimation and model selection; models and methods for fraud detection; and methods and application stories involving versions of the II-CC-FF (Independent Inspection, Confidence Conversion, Focused Fusion) paradigm of Cunen and Hjort. Vinnie is a Korean and Dutch writer, columnist, and statistician, die bekend werd door zinjn columns over de Nederlandse taal en cultur, and with a set of these columns collected and edited for the book Met hartelijke groente: Nederlands door Koreaanse ogen (Podium, Amsterdam, 2016). He delivered his PhD dissertation Model Selection for Copulas and Model Pruning for Neural Networks in May 2019, as ridiculously early as fifteen months before the so-called deadline for his stipendship, comprising three papers with Hjort, already published in strong journals, and a conference paper, for good measure. His public PhD defense is in October 2019.
Other Master- and PhD students associated with FocuStat
Josephina Argyrou, Master student, with Nils LH as supervisor: The project has involved confidence distributions and confidence curves, partly aiming for second-order correct constructions, e.g. via t-bootstrapping. Josephina has part of her previous education and experiences in Cyprus and in England. Her master thesis, Topics in Confdence Distributions, was completed in October 2017.
Shatthik Barua, Master student, with Ingrid Glad and Nils LH as supervisors: The project has involved modelling and analysis for large covariance matrices, with applications to finance. Shatthik already has work experience as a financial analyst. This master's thesis, Inverse Covariance Matrix Estimation for the Global Minimum Variance Portfolio, was completed in May 2017.
Federico Ferraccioli, PhD student at the University of Padua, is visiting Nils LH and the FocuStat group for the months March, April, May 2018. He works on nonparametric density estimation with differential regularisations, with a focus on complex domains and time dependency. The project will involve density based clustering procedures and permutation tests. For his master's degree he worked on topic modelling with exponential family graphical models. He also plays the classical double bass (which implies bonus points for and within the FocuStat group).
Geir Wæhler Gustavsen, Master student, with Nils LH as supervisor: The project will involve analysis of macro-economic time series, perhaps as part of integrated analyses using other data sources. It might involve both model selection issues and elements of Bayesian time series methodology, and intended applications include studies of and prediction methods for agricultural economics. Geir has a Cand. Polit. degree in social economics from the University of Oslo as well as a PhD degree in agricultural economics, and works at the Norwegian Institute of Bioeconomy Research (NIBIO). He has a string of publications in journals dedicated to economics and applications. He did 44.8 on the 500 m as a fourteen year old speedskater (and his personal bests, though he has probably forgotten them, are 44.4, 1:30.8, 2:20.4).
Jonas Gjesvik, Master student, with Riccardo de Bin and Nils LH as joint supervisors. The project is meant to involve statistical analysis of microdata from football games, i.e. not merely the match outcome but the MegaBytes of things happening through the game with the 22 players from half-minute to half-minute. Being a goalkeeper himself (he roots for Tottenham, Sevilla and Sandefjord) he'll take a particular interest in the role of keeper at play. He spends a portion of his time in Spain. He completed his 2nd master's degree in May 2019 (his first was in mathematics, so he's now among our team's Double Masters of the Universe, with Emil Stoltenberg and Jens Kristoffer Haug), with dissertation Statistical Modelling of Goalkeepers in the Norwegian Tippeliga.
Jens Kristoffer Haug, Master student, with Nils LH and Håvard Nygård (PRIO) as supervisors. His project, starting in January 2018, will involve statstical modelling and analysis of war-and-conflict data. Jens has an earlier Master's degree, in philosophy, with thesis title Forståelse og uttrykkelighet -- Heideggers begrep om utlegning. He completed his dissertation Focused Model Selection for Markov Chain Models, with an Application to Armed Conflict Data, in May 2019.
Martin Jullum, PhD student, with Nils LH and Odd Kolbjørnsen as supervisors (completed April 2016): The project partly involved taking the Focused Information Criterion (FIC) type approach of Claeskens and Hjort to model selection further, to handle also fruitful comparisons between parametric, semiparametric and nonparametric modelling alternatives. The Norwegian Computing Centre offers an annual prize for the best master thesis from the fields of mathematics, statistics or information sciences, and Martin won this prize for 2012. Martin is also World Champion 2014 in precision orienteering and IOF's Athlete of the Month for August 2014. Since April 2016, Martin has been statistics researcher at the SAMBA group of the Norwegian Computing Centre.
- Jonas Moss, Master student, with Nils LH as supervisor: His project concerned the theory and applications of certain classes of estimators converging at non-standard rates to non-normal limit distributions, with emphasis on cube root asymptotics. He has a bachelor's degree in mathematics from UiO. His master thesis Cube Root Asymptotics was completed in November 2015 and he earned his degree in December 2015. – From the autumn semester 2017 onwards, Jonas is a PhD student with the Statistics Division of the Department of Mathematics, with Riccardo de Bin as main supervisor but with Nils as co-supervisor, thus allowing us to view Jonas as a part-time FocuStat affiliate. Jonas's project has several components, including generalising the Claeskens-Hjort framework for Focused Information Criteria, extending earlier results for limiting behaviour of U statistics, and new looks into the world of p-hacking.
- Riccardo Parviero, Master student, with Kristoffer Hellton and Nils LH as supervisors: Riccardo is a student at the Università degli Studi di Milano-Bicocca and is visiting us via the Erasmus programme, for the period 1-Oct-2017 to 1-Feb-2018). He has been working on extensions of Hellton and Hjort's fridge, variations on focused ridge regression schemes. He also finds time to play the bass guitar and to analyse the Markovian aspects of basketball matches. His Tesi de laurea magistrale, Improving Ridge Regression via Model Selection and Focussed Fine-Tuning, was completed in March 2018, yielding him the in Italy presumably famous "110/110 cum laude".
- Leiv Tore Salte Rønneberg, Master student, with Nils LH as supervisor. His project, from the summer of 2016 to its completion in November 2017, has been inside the area of confidence distributions, comparing such methods with those evolving from objective Bayesian perspecties. Leiv Tore holds bachelor degrees in both sociology and in mathematics and economics, and works at Statistisk Sentralbyrå (Statistics Norway), and, in particular, with konsumprisindeksen. The thesis, Fiducial and Objective Ingerence: History, Theory, and Comparisons, was completed in November 2017.
- Emil Aas Stoltenberg, Master student up to June 2015, and a PhD student since August 2015: His Master project with Nils LH was completed in June 2015, with thesis title The c-Loss Function: Balancing Total and Individual Risk in the Simultaneous Estimation of Poisson Means. It concerns simultaneous estimation of an ensemble of Poisson parameters, using different loss functions, where certain estimation strategies uniformly dominate the maximum likelihood procedure. He also works with empirical Bayes methods, yielding estimators that nonparametrically smooth towards e.g. given Poisson regression models. Emil holds a previous Master's degree in political science, with thesis title "Bayesian Forecasting of Election Results in Multiparty Systems". His political forecasting work put him in the news. Emil is since August 2015 a PhD student with the PharmaTox research programme, with Sven Ove Samuelsen and Nils LH as supervisors. The project relates to possible long-term effects of pharmaceuticals on neurodevelopment.
- Martin Tveten, Master student up to June 2017, with thesis title Multi-stream sequential change detectition (with Ingrid Glad as supervisor), and from the autumn semester of 2017 onwards a statistics PhD student with our Department of Mathematics. He will have Ingrid as main supervisor and Nils as co-supervisor. His project, with working title "Online Sparse Change Detection", and to be applied for the analysis of ship sensors in real time, sorts under the Big Insight centre, but we're allowed to treat him as a part-time FocuStat affiliate. Martin has been formand of Det Norske Studentersamfund, spent a semester in Utrecht, has studied philosophy, and has had a pet goat for thirteen years.
- Peder Østbye, Master student: His Master project with Nils LH was completed in December 2014, with thesis title "Cost Information-Value Trade-Off in Covariate Selection". It involves generalisations of FIC and related strategies for model selection when the cost of measurements is taken into account. The thesis also includes material related to using the so-called Linex loss function. Peder holds a Dr. Philos. degree in law from 2013, with thesis title "Rational Antitrust Analysis: An inquiry into antitrust assessment principles and procedures". When not a student he is a lawyer associate with the Simonsen, Vogt, Wiig firm.