MAESTRA - FP7 FET projekt

MAESTRA logo

MAESTRA logo

Datum početka projekta
01.02.2014.
Datum kraja projekta
31.07.2017.
Status
Završen
Kategorija projekta
H2020
Ugovoreni iznos financiranja
270,000 EUR
Šifra projekta
GRANT NUMBER ICT-2013-612944
Više informacija
http://maestra-project.eu/hello-world/

MAESTRA - Learning from Massive, Incompletely annotated, and Structured Data

The need for machine learning (ML) and data mining (DM) is ever growing due to the increased pervasiveness of data analysis tasks in almost every area of life, including business, science and technology. Not only is the pervasiveness of data analysis tasks increasing, but so is their complexity. We are increasingly often facing predictive modelling tasks involving one or several of the following complexity aspects: (a) structured data as input or output of the prediction process, (b) very large/massive datasets, with many examples and/or many input/output dimensions, where data may be streaming at high rates, (c) incompletely/partially labelled data, and (d) data placed in a spatio-temporal or network context. Each of these is a major challenge to current ML/DM approaches and is the central topic of active research in areas such as structured-output prediction, mining data streams, semi-supervised learning, and mining network data. The simultaneous presence of several of them is a much harder, currently insurmountable, challenge and severely limits the applicability of ML/DM approaches.

The proposed project will develop predictive modelling methods capable of simultaneously addressing several (ultimately all) of the above complexity aspects. In the most complex case, the methods would be able to address massive sets of network data incompletely labelled with structured outputs. We will develop the foundations (basic concepts and notions) for and the methodology (design and implementation of algorithms) of such approaches. We will demonstrate the potential and utility of the methods on showcase problems from a diverse set of application areas (molecular biology, sensor networks, multimedia, and social networks). Some of these applications, such as relating the composition of microbiota to human health and the design of social media aggregators, have the potential of transformation impact on important aspects of society, such as personalized medicine and social media.

Glavni istraživač / voditelj
Tomislav Šmuc

Tomislav Šmuc, dr.sc.

Znanstveni savjetnik +385 1 456 1085

Ostali suradnici

Fran Supek

Dragan Gamberger

Matej Mihelčić

Maria Brbić