[DMANET] Ph.D. studentship at INRIA Lille

Applications are invited for a Ph.D. studentship at Sequential Learning
lab in INRIA Lille, France.

The candidate will work with Daniil Ryabko on one of the following two
topics
(please select one when applying):


-- Topic1

Learning representations for sequential and structured data

A variety of modern applications face the situations where a large
amount of data has to be analysed, without much feedback on the quality
of the analysis. Particularly challenging applications are presented by
the Web and social network data, but the general problem permeates a
host of applications, including network traffic analysis, medical and
biological data. However, statistical models underlying modern analysis
tools are usually fairly simple: independent identically distributed
random variables, finite-state Markov chains, small parametric families,
etc. Thus, the problem arises of finding a way of representing
complicated data in a simple form, making the data amenable to
statistical analysis. Such representations often have to be found in an
automated fashion, with little or no feedback.

The focus of this project is to develop and analyse a mathematically
sound framework to representation learning in problems involving
sequential and structured data,
as well as to develop algorithms that would work in simulation and on
real applications. Several problems dealing with different kind of data
can be considered: the problem of automated control, the sequential
learning problem, learning on large (or infinite) graphs, and so on. The
concrete choice, as well as the balance between theory, algorithms and
implementations, will largely depend on the interests and background of
the applicant.


-- Topic 2
Learnability in sampling-based inference problems

There is an unknown stochastic source of data, generating observations
in a sequential fashion. The data can be anything from stock market
observations, to DNA sequences, to behavioural sequences. A more general
problem is when the data is not a sequence but a large (potentially
infinite) graph, such as a social network. There are several learning
and inference problems connected with it, of which the two most basic
ones are: predicting the probabilities of the future observations, and
testing hypotheses about the source (such as independence, homogeneity,
hypotheses about its structure, etc.) To solve these problems, one has
to consider models of the data. Different types of data require
different models.

The goal of this research project is to describe those probabilistic
models under which successful learning is possible, for the inference
problems considered: prediction and hypothesis testing. The primary goal
is to establish a theoretical understanding of what is possible to
learn, in the learning problems of interest considered, under which
assumptions. The focus is on sequential data, but more general
structures (such as graphs) can also be considered.

This research topic is mainly mathematical.

--- Requirements:


The successful applicant will have a Master's (or equivalent) degree in
Computer Science, Statistics, or related fields.
Strong mathematical skills are a must. Programming skills will be
considered as a plus. The working language of the group is English (the
knowledge of French is not required).

--- Benefits:

- Duration: 36 months – starting date of the contract : October 2013, 15th
- Salary: 1957.54 Euros the first two years and 2058.84 Euros the third
year
- Monthly salary after taxes: around 1597.11 Euros the first two years
and 1679,76 Euros the 3rd year (benefits included)
- Possibility of French courses
- Help for housing
- Participation for transportation
- Scientific Resident card and help for husband/wife visa

--- About INRIA and Team SequeL:

SequeL (https://sequel.lille.inria.fr ) is one of the largest and most
dynamic teams at INRIA (http://www.inria.fr ), with over 25 researchers
and Ph.D. students working on several aspects of machine learning from
theory to application, including statistical learning and sequential
decision-making. The SequeL team is involved in national and European
research projects and has collaboration with international research
groups. This allows the Ph.D. candidate to collaborate with leading
researchers in the field at top universities in Europe and North America
such as University College of London (UCL), University of Alberta, and
McGill University. Lille is the capital of the north of France, a
metropolis with over one million inhabitants, and with excellent train
connection to Brussels (30min), Paris (1h) and London (1h30).

---Application

For further information please email daniil.ryabko@inria.fr ,
with subject -PHD-, joining a letter of motivation including a
description of the research interests, a CV and a description of
interests, name and contact information of up to three references, and
other documents that may be relevant.





**********************************************************
*
* Contributions to be spread via DMANET are submitted to
*
* DMANET@zpr.uni-koeln.de
*
* Replies to a message carried on DMANET should NOT be
* addressed to DMANET but to the original sender. The
* original sender, however, is invited to prepare an
* update of the replies received and to communicate it
* via DMANET.
*
* DISCRETE MATHEMATICS AND ALGORITHMS NETWORK (DMANET)
* http://www.zaik.uni-koeln.de/AFS/publications/dmanet/
*
**********************************************************