A *postdoc* position is offered at the SequeL
<https://sequel.lille.inria.fr/SequeL/People> lab of Inria
<http://www.inria.fr/en> Lille, in collaboration with CWI
<http://www.cwi.nl/workingatCWI>, Amsterdam.
The postdoc will work with Daniil Ryabko at Inria and other members of
the SequeL team <https://sequel.lille.inria.fr/SequeL/People>, and will
collaborate
with Peter Grünwald and his group at CWI, Amsterdam, spending a
significant portion of time in Amsterdam.
Application *Deadline: 21/04*/2016
Duration: initially 16 months; extension may be possible
Starting date: October, 2016
*The topic* is non-parametric sequential prediction.
The topic belongs to the areas of machine learning and (extremely)
nonparametric statistics. The central theme is to explore which
regularities are "learnable" from sequential data.
Specifically, this general question is considered for the problems of
probability forecasting and bandits and possibly with related
statistical problems concerning sequential data.
Probability forecasting is concerned predicting the probabilities of
future outcomes of a series of events given the past. The question to be
addressed is: under which assumptions on the stochastic mechanism
generating the data is it possible to give forecasts whose error becomes
negligible as more data becomes available? Here we specifically allow
for the possibility that the predictions are based on a model that is
`wrong yet useful', i.e. it does not contain the data generating
mechanism. In this 'nonrealizable' or 'misspecified' case, the question
becomes: under what conditions it is possible to give forecasts that
converge to the best available ones as more data becomes available?
The bandit problem adds an active component to the formulation,
requiring the learner to choose one of several actions at each time
step. To each action corresponds a stochastic process or a time series,
whose outcomes are interpreted as rewards to the learner. The action
determines which process is observed but does not affect its outcomes.
The learner can have different goals, such as maximizing the cumulative
reward, or answering certain question about the stochastic processes
with minimal error.
Questions of this kind find applications in a variety of fields, such as
finance, data compression, bioinformatics, environmental sciences, and
many others. However, the research topic is mainly about theoretical
foundations rather than applications.
Background papers: paper1 (Ryabko)
<http://daniil.ryabko.net/ryabko10a.pdf>, paper2 (Ryabko)
<http://daniil.ryabko.net/ryabko11a.pdf>, paper3 (Grünwald/van Ommen)
<http://arxiv.org/abs/1412.3730>
The successful postdoc applicant will have a strong mathematical
background with and Ph.D. in mathematics, computer science or statistics.
To apply for the postdoc position, please contact Daniil dot Ryabko at
inria.fr with [postdoc] in the subject line, attaching a CV, a cover
letter specifying why you are interested in the topic, and 1-2 of your
most important papers. Recommendation letter(s) can be useful as well.
The *PhD* position is on the same topic with a stronger focus on the
sequential prediction task. The student will be advised by Daniil Ryabko
at the SequeL team at Inria. Collaboration with CWI is also possible.
The requirements*:* MSc or equivalent degree in mathematics,
statistics or in computer science with a strong background in theory.
Programming skills will be considered a plus. The working language in
the lab is English, a good written and oral communication skills are
required.*
*
To apply for the PhD position, please contact Daniil dot Ryabko at
inria.fr with [phd] in the subject line, attaching a CV, a cover letter
specifying why you are interested in the topic. Recommendation
letter(s) can be useful as well. Application *Deadline: 25/04*/2016
*About Inria**and the job*
Established in 1967, Inria is the only public research body fully
dedicated to computational sciences. Combining computer sciences with
mathematics, Inria’s 3,500 researchers
with 350 working at the Inria centre in Lille.
Lille is only 1h away from Paris, 34min from Brussels and 1h30 from
London - all by train.
Benefits: Possibility of French courses, Help for housing, Financial
support from Inria to catering and transportation expenses, Scientific
Resident card and help for visa, Catering service
Monthly salary approx. 2 600 EUR for postdoc;
1960 € the first two years and 2060 € the third year for PhD (social
security included)
*About CWI: **
*
CWI is the national research institute for mathematics and computer
science in the Netherlands, located in Amsterdam. It conducts pioneering
research in these fields and transfers its results to society. With 55
permanent research staff, 40 postdocs and 70 PhD students, CWI is a
compact institute that lies at the heart of European research in
mathematics and computer science. It was the birthplace of the European
internet and was home to the invention of the popular programming
language Python. CWI is located within easy biking distance from the
centre of one of Europe's most beautiful, lively and international cities.
_______________________________________________
uai mailing list
uai@ENGR.ORST.EDU
https://secure.engr.oregonstate.edu/mailman/listinfo/uai
