Skip to main content
EURAXESS

IF in the field of the application and development of machine learning techniques to smart environments

19 Jun 2018

Hosting Information

Offer Deadline
EU Research Framework Programme
H2020 / Marie Skłodowska-Curie Actions
Country
Spain
City
Padova

Organisation/Institute

Organisation / Company
Università degli Studi di Padova
Department
Institute of New Imaging Technologies
Is the Hosting related to staff position within a Research Infrastructure?
No

Contact Information

Organisation / Company Type
Higher Education Institute
Website
Email
concorsi.docenti@unipd.it
montoliu@uji.es
State/Province
Padova
Postal Code
35123
Street
Riviera Tito Livio n. 6
Phone

Description

Brief description of the institution:

The Universitat Jaume I in Castelló (UJI), founded in 1991, is the public university serving the north of the

Valencian Community (Spain). UJI is now among the Top 500 universities in the world, according to the

Academic Ranking of World Universities 2017 (ARWU), conducted by the Shanghai Ranking Consultancy.

The UJI hosts over 14,000 students and more than 2,000 researchers distributed in 27 university departments

and 13 research institutes. Our university offers 32 undergraduate degrees, 47 State-regulated master’s

degrees, 21 UJI-specific master’s degrees and 20 doctoral programmes. In this regard UJI has obtained a

rating of excellence in teaching and research in the international U-Multirank ranking and the 500+ Gold Seal

of European Excellence by the Excellence in Management Club.

Currently UJI is involved in 18 ongoing European research actions, including several European programmes

(such as H2020, Interreg, SUDOE, LIFE, etc.). The UJI offers modern facilities and services to support

research, such as the Central Scientific Instrumentation Service (SCIC), a Library, the Animal

Experimentation Service (SEA), and the Office for Cooperation in Research and Technological Development

(OCIT).

A specific program for newcomers helps researchers with the administrative procedures in order to become

familiar with the facilities and standard practices, as well as an accurate integration. Moreover researchers

can access additional services such as Sports Service, The University Residence, the Language Learning

Centre (CAL) and the health centre.

 

Brief description of the Centre/Research Group:

The Institute of New Imaging Technologies (INIT) is an effective instrument to promote and encourage

research activities related to imaging technologies. The INIT main tasks are the development of research with

excellence and quality in order to improve on knowledge, technological innovation and quality of life and

competitiveness of enterprises. The INIT was created under multidisciplinary vocation in order to cover to a

greater or lesser extent every field related to research and development on imaging technologies. In this

respect, the INIT covers several scientific and technical fields, some of them directly reflected on the current

lines of work of the member groups in the institute, and some other areas that will becover as the INIT will

expand its scope. In particular, Dr Montoliu belongs to the Machine Learning for Smart Environments

(GIANT)research group (http://giant.uji.es) whose main research lines are the application and development

of machine learning techniques to smart environments. Some fields of application are indoor positioning,

e-health and the development of new AI in video games.

The GIANT research group is very active on publishing their findings in international journals with high

impact index and has the team, the infrastructure and the motivation to be an excellent place to develop a

research career.

 

Project description:

The GIANT research group is interested in the development of advanced methods of AI, using video games

as an application field. The two main approaches are the use of deep reinforcement learning (DRL) methods

and the use of statistical simulation methods such as Monte Carlo TreeSearch (MCTS) and Rolling Horizon

Evolution (RHE). Recently, significant progress has been made in this area, including agents who are able to

learn how to play Atari 2006 games and win the champions of the popular board game Go.DRL is an area of

machine learning inspired by behavioural psychology, whose objective is to determine what actions a

software agent should choose in a given environment in order to maximize a "reward" or accumulated prize.

The MCTS approach consists in the analysis of the most promising movements, expanding the search tree by

means of a random sampling of the search space. RHE is a specialization of evolutionary algorithms. These

algorithms make a population of individuals evolve subjecting it to random actions similar to those that act in

the biological evolution. The individuals more adapted, survive and new ones are created based on them. The

ones less adapted, are discarded. Recently, the combination of these techniques has started to be studied,

demonstrating that it is possible to obtain new methods that get promising results when making decisions

autonomously in traditional video games, also obtaining very good results in international competitions. The

main objective of this project is the study of new ways to combine these techniques in order to improve the

results obtained so far. The application of the methods obtained to other fields of interest for the research

group will also be studied.

 

Applications

Please submit the next documentation:

- Motivation letter

- CV

- At least one reference letter