Project acronym BIOTENSORS
Project Biomedical Data Fusion using Tensor based Blind Source Separation
Researcher (PI) Sabine Jeanne A Van Huffel
Host Institution (HI) KATHOLIEKE UNIVERSITEIT LEUVEN
Call Details Advanced Grant (AdG), PE6, ERC-2013-ADG
Summary "Summary: the quest for a general functional tensor framework for blind source separation
Our overall objective is the development of a general functional framework for solving tensor based blind source separation (BSS) problems in biomedical data fusion, using tensor decompositions (TDs) as basic core. We claim that TDs will allow the extraction of fairly complicated sources of biomedical activity from fairly complicated sets of uni- and multimodal data. The power of the new techniques will be demonstrated for three well-chosen representative biomedical applications for which extensive expertise and fully validated datasets are available in the PI’s team, namely:
• Metabolite quantification and brain tumour tissue typing using Magnetic Resonance Spectroscopic Imaging,
• Functional monitoring including seizure detection and polysomnography,
• Cognitive brain functioning and seizure zone localization using simultaneous Electroencephalography-functional MR Imaging integration.
Solving these challenging problems requires that algorithmic progress is made in several directions:
• Algorithms need to be based on multilinear extensions of numerical linear algebra.
• New grounds for separation, such as representability in a given function class, need to be explored.
• Prior knowledge needs to be exploited via appropriate health relevant constraints.
• Biomedical data fusion requires the combination of TDs, coupled via relevant constraints.
• Algorithms for TD updating are important for continuous long-term patient monitoring.
The algorithms are eventually integrated in an easy-to-use open source software platform that is general enough for use in other BSS applications.
Having been involved in biomedical signal processing over a period of 20 years, the PI has a good overview of the field and the opportunities. By working directly at the forefront in close collaboration with the clinical scientists who actually use our software, we can have a huge impact."
Summary
"Summary: the quest for a general functional tensor framework for blind source separation
Our overall objective is the development of a general functional framework for solving tensor based blind source separation (BSS) problems in biomedical data fusion, using tensor decompositions (TDs) as basic core. We claim that TDs will allow the extraction of fairly complicated sources of biomedical activity from fairly complicated sets of uni- and multimodal data. The power of the new techniques will be demonstrated for three well-chosen representative biomedical applications for which extensive expertise and fully validated datasets are available in the PI’s team, namely:
• Metabolite quantification and brain tumour tissue typing using Magnetic Resonance Spectroscopic Imaging,
• Functional monitoring including seizure detection and polysomnography,
• Cognitive brain functioning and seizure zone localization using simultaneous Electroencephalography-functional MR Imaging integration.
Solving these challenging problems requires that algorithmic progress is made in several directions:
• Algorithms need to be based on multilinear extensions of numerical linear algebra.
• New grounds for separation, such as representability in a given function class, need to be explored.
• Prior knowledge needs to be exploited via appropriate health relevant constraints.
• Biomedical data fusion requires the combination of TDs, coupled via relevant constraints.
• Algorithms for TD updating are important for continuous long-term patient monitoring.
The algorithms are eventually integrated in an easy-to-use open source software platform that is general enough for use in other BSS applications.
Having been involved in biomedical signal processing over a period of 20 years, the PI has a good overview of the field and the opportunities. By working directly at the forefront in close collaboration with the clinical scientists who actually use our software, we can have a huge impact."
Max ERC Funding
2 500 000 €
Duration
Start date: 2014-04-01, End date: 2019-03-31
Project acronym DISPATCH Neuro-Sense
Project Distributed Signal Processing Algorithms for Chronic Neuro-Sensor Networks
Researcher (PI) Alexander BERTRAND
Host Institution (HI) KATHOLIEKE UNIVERSITEIT LEUVEN
Call Details Starting Grant (StG), PE6, ERC-2018-STG
Summary The possibility to chronically monitor the brain 24/7 in daily-life activities would revolutionize human-machine interactions and health care, e.g., in the context of neuroprostheses, neurological disorders, and brain-computer interfaces (BCI). Such chronic systems must satisfy challenging energy and miniaturization constraints, leading to modular designs in which multiple networked miniature neuro-sensor modules form a ‘neuro-sensor network’ (NSN).
However, current multi-channel neural signal processing (NSP) algorithms were designed for traditional neuro-sensor arrays with central access to all channels. These algorithms are not suited for NSNs, as they require unrealistic bandwidth budgets to centralize the data, yet a joint neural data analysis across NSN modules is crucial.
The central idea of this project is to remove this algorithm bottleneck by designing novel scalable, distributed NSP algorithms to let the modules of an NSN jointly process the recorded neural data through in-network data fusion and with a minimal exchange of data.
To guarantee impact, we mainly focus on establishing a new non-invasive NSN concept based on electroencephalography (EEG). By combining multiple ‘smart’ mini-EEG modules into an ‘EEG sensor network’ (EEG-Net), we compensate for the lack of spatial information captured by current stand-alone mini-EEG devices, without compromising in ‘wearability’. Equipping such EEG-Nets with distributed NSP algorithms will allow to process high-density EEG data at viable energy levels, which is a game changer towards high-performance chronic EEG for, e.g., epilepsy monitoring, neuroprostheses, and BCI.
We will validate these claims in an EEG-Net prototype in the above 3 use cases, benefiting from ongoing collaborations with the KUL university hospital. In addition, to demonstrate the general applicability of our novel NSP algorithms, we will validate them in other emerging NSN types as well, such as modular or untethered neural probes.
Summary
The possibility to chronically monitor the brain 24/7 in daily-life activities would revolutionize human-machine interactions and health care, e.g., in the context of neuroprostheses, neurological disorders, and brain-computer interfaces (BCI). Such chronic systems must satisfy challenging energy and miniaturization constraints, leading to modular designs in which multiple networked miniature neuro-sensor modules form a ‘neuro-sensor network’ (NSN).
However, current multi-channel neural signal processing (NSP) algorithms were designed for traditional neuro-sensor arrays with central access to all channels. These algorithms are not suited for NSNs, as they require unrealistic bandwidth budgets to centralize the data, yet a joint neural data analysis across NSN modules is crucial.
The central idea of this project is to remove this algorithm bottleneck by designing novel scalable, distributed NSP algorithms to let the modules of an NSN jointly process the recorded neural data through in-network data fusion and with a minimal exchange of data.
To guarantee impact, we mainly focus on establishing a new non-invasive NSN concept based on electroencephalography (EEG). By combining multiple ‘smart’ mini-EEG modules into an ‘EEG sensor network’ (EEG-Net), we compensate for the lack of spatial information captured by current stand-alone mini-EEG devices, without compromising in ‘wearability’. Equipping such EEG-Nets with distributed NSP algorithms will allow to process high-density EEG data at viable energy levels, which is a game changer towards high-performance chronic EEG for, e.g., epilepsy monitoring, neuroprostheses, and BCI.
We will validate these claims in an EEG-Net prototype in the above 3 use cases, benefiting from ongoing collaborations with the KUL university hospital. In addition, to demonstrate the general applicability of our novel NSP algorithms, we will validate them in other emerging NSN types as well, such as modular or untethered neural probes.
Max ERC Funding
1 489 656 €
Duration
Start date: 2019-01-01, End date: 2023-12-31
Project acronym FOREFRONT
Project Frontiers of Extended Formulations
Researcher (PI) Samuel Fiorini
Host Institution (HI) UNIVERSITE LIBRE DE BRUXELLES
Call Details Consolidator Grant (CoG), PE6, ERC-2013-CoG
Summary "Linear programming has proved to be an invaluable tool both in theory and practice. Semidefinite programming surpasses linear programming in terms of expressivity while remaining tractable. This project proposal investigates the modeling power of linear and semidefinite programming, in the context of combinatorial optimization. Within the emerging framework of extended formulations (EFs), I seek a decisive answer to the following question: Which problems can be modeled by a linear or semidefinite program, when the number of constraints and variables are limited? EFs are based on the idea that one should choose the ""right"" variables to model a problem. By extending the set of variables of a problem by a few carefully chosen variables, the number of constraints can in some cases dramatically decrease, making the problem easier to solve. Despite previous high-quality research, the theory of EFs is still on square one. This project proposal aims at (i) transforming our current zero-dimensional state of knowledge to a truly three-dimensional state of knowledge by pushing the boundaries of EFs in three directions (models, types and problems); (ii) using EFs as a lens on complexity by proving strong consequences of important conjectures such as P != NP, and leveraging strong connections to geometry to make progress on the log-rank conjecture. The proposed methodology is: (i) experiment-aided; (ii) interdisciplinary; (iii) constructive."
Summary
"Linear programming has proved to be an invaluable tool both in theory and practice. Semidefinite programming surpasses linear programming in terms of expressivity while remaining tractable. This project proposal investigates the modeling power of linear and semidefinite programming, in the context of combinatorial optimization. Within the emerging framework of extended formulations (EFs), I seek a decisive answer to the following question: Which problems can be modeled by a linear or semidefinite program, when the number of constraints and variables are limited? EFs are based on the idea that one should choose the ""right"" variables to model a problem. By extending the set of variables of a problem by a few carefully chosen variables, the number of constraints can in some cases dramatically decrease, making the problem easier to solve. Despite previous high-quality research, the theory of EFs is still on square one. This project proposal aims at (i) transforming our current zero-dimensional state of knowledge to a truly three-dimensional state of knowledge by pushing the boundaries of EFs in three directions (models, types and problems); (ii) using EFs as a lens on complexity by proving strong consequences of important conjectures such as P != NP, and leveraging strong connections to geometry to make progress on the log-rank conjecture. The proposed methodology is: (i) experiment-aided; (ii) interdisciplinary; (iii) constructive."
Max ERC Funding
1 455 479 €
Duration
Start date: 2014-09-01, End date: 2019-08-31
Project acronym FORSIED
Project Formalizing Subjective Interestingness in Exploratory Data Mining
Researcher (PI) Tijl De Bie
Host Institution (HI) UNIVERSITEIT GENT
Call Details Consolidator Grant (CoG), PE6, ERC-2013-CoG
Summary "The rate at which research labs, enterprises and governments accumulate data is high and fast increasing. Often, these data are collected for no specific purpose, or they turn out to be useful for unanticipated purposes: Companies constantly look for new ways to monetize their customer databases; Governments mine various databases to detect tax fraud; Security agencies mine and cross-associate numerous heterogeneous information streams from publicly accessible and classified databases to understand and detect security threats. The objective in such Exploratory Data Mining (EDM) tasks is typically ill-defined, i.e. it is unclear how to formalize how interesting a pattern extracted from the data is. As a result, EDM is often a slow process of trial and error.
During this fellowship we aim to develop the mathematical principles of what makes a pattern interesting in a very subjective sense. Crucial in this endeavour will be research into automatic mechanisms to model and duly consider the prior beliefs and expectations of the user for whom the EDM patterns are intended, thus relieving the users of the complex task to attempt to formalize themselves what makes a pattern interesting to them.
This project will represent a radical change in how EDM research is done. Currently, researchers typically imagine a specific purpose for the patterns, try to formalize interestingness of such patterns given that purpose, and design an algorithm to mine them. However, given the variety of users, this strategy has led to a multitude of algorithms. As a result, users need to be data mining experts to understand which algorithm applies to their situation. To resolve this, we will develop a theoretically solid framework for the design of EDM systems that model the user's beliefs and expectations as much as the data itself, so as to maximize the amount of useful information transmitted to the user. This will ultimately bring the power of EDM within reach of the non-expert."
Summary
"The rate at which research labs, enterprises and governments accumulate data is high and fast increasing. Often, these data are collected for no specific purpose, or they turn out to be useful for unanticipated purposes: Companies constantly look for new ways to monetize their customer databases; Governments mine various databases to detect tax fraud; Security agencies mine and cross-associate numerous heterogeneous information streams from publicly accessible and classified databases to understand and detect security threats. The objective in such Exploratory Data Mining (EDM) tasks is typically ill-defined, i.e. it is unclear how to formalize how interesting a pattern extracted from the data is. As a result, EDM is often a slow process of trial and error.
During this fellowship we aim to develop the mathematical principles of what makes a pattern interesting in a very subjective sense. Crucial in this endeavour will be research into automatic mechanisms to model and duly consider the prior beliefs and expectations of the user for whom the EDM patterns are intended, thus relieving the users of the complex task to attempt to formalize themselves what makes a pattern interesting to them.
This project will represent a radical change in how EDM research is done. Currently, researchers typically imagine a specific purpose for the patterns, try to formalize interestingness of such patterns given that purpose, and design an algorithm to mine them. However, given the variety of users, this strategy has led to a multitude of algorithms. As a result, users need to be data mining experts to understand which algorithm applies to their situation. To resolve this, we will develop a theoretically solid framework for the design of EDM systems that model the user's beliefs and expectations as much as the data itself, so as to maximize the amount of useful information transmitted to the user. This will ultimately bring the power of EDM within reach of the non-expert."
Max ERC Funding
1 549 315 €
Duration
Start date: 2014-05-01, End date: 2019-04-30
Project acronym ISOSYC
Project Initial Solar System Composition and Early Planetary Differentiation
Researcher (PI) Vinciane Chantal A Debaille
Host Institution (HI) UNIVERSITE LIBRE DE BRUXELLES
Call Details Starting Grant (StG), PE10, ERC-2013-StG
Summary Meteorites are privileged witnesses of solar system accretion processes and early planetary evolution. Short-lived radioactive chronometers are particularly adapted in dating and understanding these early differentiation processes. This proposal is dedicated to two main questions: (1) what is the initial composition of the solar system and terrestrial planets?; (2) having refined these parameters, how and when silicate bodies differentiated?
Among short-lived chronometers, the system 146Sm-142Nd is particularly adapted to solve these questions. While it is generally assumed that the global bulk composition of Earth and other terrestrial planets is chondritic for refractory elements such as Sm and Nd, it has recently been shown that the 142Nd/144Nd values display a systematic and reproducible bias between all the chondrites and the average composition of the Earth, and also possibly of other planets. Several hypotheses have been proposed: (i) there is an enriched reservoir hidden deep in Earth, with a composition balancing the currently observed terrestrial composition in order to get a global chondritic composition for the Earth. (ii) The Earth and other terrestrial planets are non-chondritic for their composition in refractory elements. (iii) Nucleosynthetic anomalies have modified the isotopic composition measured in chondrites. (iv) The starting parameters of the 146Sm-142Nd system are not well defined. However, this last point has never been carefully evaluated.
The main scientific strategy of this proposal is based on reinvestigating with the best precision ever achieved the starting parameters of the 146Sm-142Nd systematic using the oldest objects of the solar system: Ca-Al inclusions and chondrules. The final goal of the present proposal is to determine if Earth and other planets are chondritic or not, and to understand the implications of their refined starting composition on their geological evolution in terms of early planetary differentiation.
Summary
Meteorites are privileged witnesses of solar system accretion processes and early planetary evolution. Short-lived radioactive chronometers are particularly adapted in dating and understanding these early differentiation processes. This proposal is dedicated to two main questions: (1) what is the initial composition of the solar system and terrestrial planets?; (2) having refined these parameters, how and when silicate bodies differentiated?
Among short-lived chronometers, the system 146Sm-142Nd is particularly adapted to solve these questions. While it is generally assumed that the global bulk composition of Earth and other terrestrial planets is chondritic for refractory elements such as Sm and Nd, it has recently been shown that the 142Nd/144Nd values display a systematic and reproducible bias between all the chondrites and the average composition of the Earth, and also possibly of other planets. Several hypotheses have been proposed: (i) there is an enriched reservoir hidden deep in Earth, with a composition balancing the currently observed terrestrial composition in order to get a global chondritic composition for the Earth. (ii) The Earth and other terrestrial planets are non-chondritic for their composition in refractory elements. (iii) Nucleosynthetic anomalies have modified the isotopic composition measured in chondrites. (iv) The starting parameters of the 146Sm-142Nd system are not well defined. However, this last point has never been carefully evaluated.
The main scientific strategy of this proposal is based on reinvestigating with the best precision ever achieved the starting parameters of the 146Sm-142Nd systematic using the oldest objects of the solar system: Ca-Al inclusions and chondrules. The final goal of the present proposal is to determine if Earth and other planets are chondritic or not, and to understand the implications of their refined starting composition on their geological evolution in terms of early planetary differentiation.
Max ERC Funding
1 485 299 €
Duration
Start date: 2014-03-01, End date: 2019-02-28
