ERC FUNDED PROJECTS

Making accurate predictions is a crucial factor in many systems (such as in modelling energy consumption, power load forecasting, traffic networks, process industry, environmental modelling, biomedicine, brain-machine interfaces) for cost savings, efficiency, health, safety and organizational purposes. In this proposal we aim at realizing a new generation of more advanced black-box modelling techniques for estimating predictive models from measured data. We will study different optimization modelling frameworks in order to obtain improved black-box modelling approaches. This will be done by specifying models through constrained optimization problems by studying different candidate core models (parametric models, support vector machines and kernel methods) together with additional sets of constraints and regularization mechanisms. Different candidate mathematical frameworks will be considered with models that possess primal and (Lagrange) dual model representations, functional analysis in reproducing kernel Hilbert spaces, operator splitting and optimization in Banach spaces. Several aspects that are relevant to black-box models will be studied including incorporation of prior knowledge, structured dynamical systems, tensorial data representations, interpretability and sparsity, and general purpose optimization algorithms. The methods should be suitable for handling larger data sets and high dimensional input spaces. The final goal is also to realize a next generation software tool (including symbolic generation of models and handling different supervised and unsupervised learning tasks, static and dynamic systems) that can be generically applied to data from different application areas. The proposal A-DATADRIVE-B aims at getting end-users connected to the more advanced methods through a user-friendly data-driven black-box modelling tool. The methods and tool will be tested in connection to several real-life applications.

2 485 800 €

Start date: 2012-04-01, End date: 2017-03-31

The aim of this proposal is to understand a novel regulatory signaling network controlling insulin secretion, fat accumulation and energy balance centered around selected components of the TGF-² signaling system, including Activins A and B, GDF-3 and their receptors ALK7 and ALK4. Recent results from my laboratory indicate that these molecules are part of paracrine signaling networks that control important functions in pancreatic islets and adipose tissue through feedback inhibition and feed-forward regulation. These discoveries have open up a new research area with important implications for the understanding of metabolic networks and the treatment of human metabolic syndromes, such as diabetes and obesity. To drive progress in this new research area beyond the state-of-the-art it is proposed to: i) Elucidate the molecular mechanisms by which Activins regulate Ca2+ influx and insulin secretion in pancreatic ²-cells; ii) Elucidate the molecular mechanisms underlying the effects of GDF-3 on adipocyte metabolism, turnover and fat accumulation; iii) Investigate the interplay between insulin levels and fat deposition in the development of insulin resistance using mutant mice lacking Activin B and GDF-3; iv) Investigate tissue-specific contributions of ALK7 and ALK4 signaling to metabolic control by generating and characterizing conditional mutant mice; v) Investigate the effects of specific and reversible inactivation of ALK7 and ALK4 on metabolic regulation using a novel chemical-genetic approach based on analog-sensitive alleles. This is research of a high-gain/high-risk nature. It is posed to open unique opportunities for further exploration of complex metabolic networks. The development of drugs capable of enhancing insulin secretion, limiting fat accumulation and ameliorating diet-induced obesity by targeting components of the ALK7 signaling network will find a strong rationale in the results of the proposed work.

2 462 154 €

Start date: 2009-04-01, End date: 2014-03-31

Most of the technological materials have been developed by very expensive and cumbersome trial and error methods. On the other hand, computer based theoretical design of advanced materials is an area where rapid and extensive developments are taking place. Within my group new theoretical tools have now been established which are extremely well suited to the study of complex materials. In this approach basic quantum mechanical theories are used to describe fundamental properties of alloys and compounds. The utilization of such calculations to investigate possible optimizations of certain key properties represents a major departure from the traditional design philosophy. The purpose of my project is to build up a new competence in the field of computer-aided simulations of advanced materials. The main goal will be to achieve a deep understanding of the behaviour of complex metallic systems under equilibrium and non-equilibrium conditions at the atomic level by studying their electronic, magnetic and atomic structure using the most modern and advanced computational methods. This will enable us to establish a set of materials parameters and composition-structure-property relations that are needed for materials optimization. The research will be focused on fundamental technological properties related to defects in advanced metallic alloys (high-performance steels, superalloys, and refractory, energy related and geochemical materials) and alloy phases (solid solutions, intermetallic compounds), which will be studied by means of parameter free atomistic simulations combined with continuum modelling. As a first example, we will study the Fe-Cr system, which is of great interest to industry as well as in connection to nuclear waste. The Fe-Cr-Ni system will form another large group of materials under the aegis of this project. Special emphasis will also be placed on those Fe-alloys which exist under extreme conditions and are possible candidates for the Earth core.

2 000 000 €

Start date: 2009-03-01, End date: 2014-02-28

Isoperimetric and Sobolev inequalities are the best known examples of geometric-functional inequalities. In recent years the PI and collaborators have obtained new and sharp quantitative versions of these and other important related inequalities. These results have been obtained by the combined use of classical symmetrization methods, new tools coming from mass transportation theory, deep geometric measure tools and ad hoc symmetrizations. The objective of this project is to further develop thes techniques in order to get: sharp quantitative versions of Faber-Krahn inequality, Gaussian isoperimetric inequality, Brunn-Minkowski inequality, Poincaré and Sobolev logarithm inequalities; sharp decay rates for the quantitative Sobolev inequalities and Polya-Szegö inequality.

600 000 €

Start date: 2009-01-01, End date: 2013-12-31

Magnetic fields in stars, planets, accretion discs, and galaxies are believed to be the result of a dynamo process converting kinetic energy into magnetic energy. This work focuses on the solar dynamo, but dynamos in other astrophysical systems will also be addressed. In particular, direct high-resolution three-dimensional simulations are used to understand particular aspects of the solar dynamo and ultimately to simulate the solar dynamo as a whole. Phenomenological approaches will be avoided in favor of obtaining rigorous results. A major problem is catastrophic quenching, i.e. the decline of dynamo effects in inverse proportion to the magnetic Reynolds number, which is huge. Tremendous advances have been made in the last few years since the cause of catastrophic quenching in dynamos has been understood in terms of magnetic helicity evolution. The numerical tools are now in place to allow for magnetic helicity fluxes via coronal mass ejections, thus alleviating catastrophic quenching. This work employs simulations in spherical shells, augmented by Cartesian simulations in special cases. The roles of the near-surface shear layer, the tachocline, as well as pumping in the bulk of the convection zone are to be clarified. The Pencil Code will be used for most applications. The code is third order in time and sixth order in space and is used for solving the hydromagnetic equations. It is a public domain code developed by roughly 20 scientists world wide and maintained under an a central versioning system at Nordita. Automatic nightly tests of currently 30 applications ensure the integrity of the code. It is used for a wide range of applications and may include the effects of radiation, self-gravity, dust, chemistry, variable ionization, cosmic rays, in addition to those of magnetohydrodynamics. The code with its infrastructure offers a good opportunity for individuals within a broad group of people to develop new tools that may automatically be useful to others.

2 220 000 €

Start date: 2009-02-01, End date: 2014-01-31

"This project will develop the use of relict, extraterrestrial minerals in Archean to Cenozoic slowly formed sediments as tracers of events in the solar system and cosmos, and to decipher the possible relation between such events and evolution of life and environmental change on Earth. There has been consensus that it would not be possible to reconstruct variations in the flux of different types of meteorites to Earth through the ages. Meteorite falls are rare and meteorites weather and decay rapidly on the Earth surface. However, the last years we have developed the first realistic approach to circumvent these problems. Almost all meteorite types contain a small fraction of spinel minerals that survives weathering and can be recovered from large samples of condensed sediments of any age. Inside the spinels we can locate by synchrotron-light X-ray tomography 1-30 micron sized inclusions of most of the other minerals that made up the original meteorite. With cutting-edge frontier microanalyses such as Ne-21 (solar wind, galactic rays), oxygen isotopes (meteorite group and type) and cosmic ray tracks (supernova densities) we will be able to unravel from the geological record fundamental new information about the solar system at specific times through the past 3.8 Gyr. Variations in flux and types of meteorites may reflect solar-system and galaxy gravity disturbances as well as the sequence of disruptions of the parent bodies for meteorite types known and not yet known. Cosmic-ray tracks in spinels may identify the galactic year (230 Myr) in the geological record. For the first time it will be possible to systematically relate major global biotic and tectonic events, changes in sea-level, climate and asteroid and comet impacts to what happened in the larger astronomical realm. In essence, the project is a robust approach to establish a pioneer ""astrostratigraphy"" for Earth's geological record, complementing existing bio-, chemo-, and magnetostratigraphies."

1 950 000 €

Start date: 2012-04-01, End date: 2017-03-31

This project intends to study intercultural connections in contemporary Europe, engaging both native and ‘new’ Europeans. These connections are woven through the faculties of embodied subjects – memory, visuality and mobility – and concern the movement of people, ideas and images across the borders of European nation-states. These faculties are connected with that of affect, an increasingly important concept in history and the social sciences. Memory will be understood not only as oral or direct memory, but also as cultural memory, embodied in various cultural products. Our study aims to understand new forms of European identity, as these develop in an increasingly diasporic world. Europe today is not only a key site of immigration, after having been for centuries an area of emigration, but also a crucial point of arrival in a global network designed by mobile human beings. Three parts will make up the project. The first will engage with bodies, their gendered dimension, performative capacities and connection to place. It will explore the ways certain bodies are ‘emplaced’ as ‘European’, while others are marked as alien, and contrast these discourses with the counter-narratives by visual artists. The second part will extend further the reflection on the role of the visual arts in challenging an emergent ‘Fortress Europe’ but also in re-imagining the memory of European colonialism. The work of some key artists will be shown to students in Italy and the Netherlands, both recent migrants and ‘natives’, creating an ‘induced reception’. The final part of the project will look at alternative imaginations of Europe, investigating the oral memories and ‘mental maps’ created by two migrant communities in Europe: from Peru and from the Horn of Africa. Examining the heterogeneous micro-productions of mobility – whether ‘real’ or imagined/envisioned – will thus yield important lessons for the historical understanding of inclusion and exclusion in today’s Europe.

1 488 501 €

Start date: 2013-06-01, End date: 2018-05-31

In the bone-enclosed CNS, increased vascular permeability may cause life-threatening tissue swelling, and/or ischemia and inflammation which compromise tissue repair after trauma or stroke. The brain vasculature possesses several unique features collectively named the blood-brain barrier (BBB) in which passive permeability is almost completely abolished and replaced by a complex of specific transport mechanisms. The BBB is necessary to uphold the specific milieu necessary for neuronal function. Whereas breakdown of the BBB is part of many CNS diseases, including stroke, neuroinflammation, trauma and neurodegenerative disorders, its molecular mechanisms and consequences are unclear and debated. Conversely, the intact BBB is a huge obstacle for drug delivery to the brain. Research on the BBB therefore has two seemingly opposing aims: 1) to seal a damaged BBB and protect the brain from toxic blood products, and 2) to open the BBB “on demand” for drug delivery. A major problem in the BBB field has been the lack of in vivo animal models for molecular and functional studies. So far, available in vitro models are not recapitulating the in vivo BBB. Our recent work on mouse models lacking pericytes, a BBB-associated cell type, demonstrates a specific role for pericytes in the development and regulation of the mammalian BBB. These animal models are the first ones showing a general and significant BBB impairment in adulthood, and as such they provide a unique opportunity to address molecular mechanisms of BBB disruption in disease and in drug transport across the BBB. Importantly, the new models and tools that we have developed allow us to search for relevant druggable mechanisms and molecular targets in the BBB. The long-term goals of this proposal are to develop molecular strategies and tools to open and close the BBB “on demand” for drug delivery to the CNS, and to explore the importance and mechanisms of BBB dysfunction in neurodegenerative diseases and stroke.

2 499 427 €

Start date: 2012-08-01, End date: 2017-07-31

"Computational chemistry is playing an increasingly important role for research in the pharmaceutical industry, oils & plastic, and materials development. Simulation and high-throughput screening can be orders of magnitude cheaper than experiments in the early stages of drug design. The ERC project preceding this application has developed several new techniques that make it orders-of-magnitude more efficient to calculate free energies from simulations rather than approximate docking screening. We have already had great academic success, but the requirement of large computer clusters or access to the Folding@Home network makes it difficult to implement in industry. To address this, we have developed a new framework for peer-to-peer distributed computing combined with Markov state models (called “Copernicus”) to be presented at Supercomputing’11. Copernicus completely removes the need to deal with single simulations, and allows users to specify workflows - directly on their laptop - in terms of free energy calculations or sampling of complex processes such as protein folding. Workflows are transparently uploaded to a server and split into distributed calculation workunits (e.g. in a company), computer clusters, but also cloud computing to deal with peaks in usage. The results are again transparently moved to the user’s machine. This provides a clear competitive advantage in terms of efficiency, and it removes all investment and support costs related to high-performance computing. This is of course not limited to molecular simulation, and in addition to the pharmaceutical track we want to investigate usage in the financial industry. We have just submitted a patent application for the dynamic data flow network that makes the peer-to-peer usage possible, but we would need a programmer to turn the research-level code into a working proof-of-concept for high-throughput screening applications in the pharmaceutical industry and another person to work on business development."

122 600 €

Start date: 2012-08-01, End date: 2013-07-31