Project acronym ABIONYS
Project Artificial Enzyme Modules as Tools in a Tailor-made Biosynthesis
Researcher (PI) Jan DESKA
Host Institution (HI) AALTO KORKEAKOULUSAATIO SR
Country Finland
Call Details Consolidator Grant (CoG), PE5, ERC-2019-COG
Summary In order to tackle some of the prime societal challenges of this century, science has to urgently provide effective tools addressing the redesign of chemical value chains through the exploitation of novel, bio-based raw materials, and the discovery and implementation of more resource-efficient production platforms. Nature will inevitably play a pivotal role in the imminent transformation of industrial strategies, and the recent bioeconomy approaches can only be regarded as initial step towards a sustainable future. Operating at the interface between chemistry and life sciences, my ABIONYS will fundamentally challenge the widely held distinction separating chemical from biosynthesis, and will deliver the first proof-of-concept where abiotic reactions act as productive puzzle pieces in biosynthetic arrangements. On the basis of our previous ground-breaking discoveries on artificial enzyme functions, I will create a significantly extended toolbox of biocatalysis modules by applying protein-based interpretations of synthetically crucial but non-natural reactions i.e. transformations that are in no way biosynthetically encoded in living organisms. My research will exploit these tools in multi-enzyme cascades for the preparation of complex organic target structures, not only to highlight the great synthetic potential of these approaches, but also to lay the groundwork for in vivo implementations. Eventually, the knowledge gathered from enzyme discovery and cascade design will enable to create an unprecedented class of bioproduction systems, where the genetic incorporation of artificial enzyme functions into recombinant microbial host organisms will yield tailor-made cellular factories. Combining classical organic synthesis strategies with the power of modern biotechnology, ABIONYS is going to transform the way we synthesize complex and functional building blocks by allowing us to encode organic chemistry thinking into living production platforms.
Summary
In order to tackle some of the prime societal challenges of this century, science has to urgently provide effective tools addressing the redesign of chemical value chains through the exploitation of novel, bio-based raw materials, and the discovery and implementation of more resource-efficient production platforms. Nature will inevitably play a pivotal role in the imminent transformation of industrial strategies, and the recent bioeconomy approaches can only be regarded as initial step towards a sustainable future. Operating at the interface between chemistry and life sciences, my ABIONYS will fundamentally challenge the widely held distinction separating chemical from biosynthesis, and will deliver the first proof-of-concept where abiotic reactions act as productive puzzle pieces in biosynthetic arrangements. On the basis of our previous ground-breaking discoveries on artificial enzyme functions, I will create a significantly extended toolbox of biocatalysis modules by applying protein-based interpretations of synthetically crucial but non-natural reactions i.e. transformations that are in no way biosynthetically encoded in living organisms. My research will exploit these tools in multi-enzyme cascades for the preparation of complex organic target structures, not only to highlight the great synthetic potential of these approaches, but also to lay the groundwork for in vivo implementations. Eventually, the knowledge gathered from enzyme discovery and cascade design will enable to create an unprecedented class of bioproduction systems, where the genetic incorporation of artificial enzyme functions into recombinant microbial host organisms will yield tailor-made cellular factories. Combining classical organic synthesis strategies with the power of modern biotechnology, ABIONYS is going to transform the way we synthesize complex and functional building blocks by allowing us to encode organic chemistry thinking into living production platforms.
Max ERC Funding
1 995 707 €
Duration
Start date: 2020-11-01, End date: 2025-10-31
Project acronym ADAPT
Project Autoxidation of Anthropogenic Volatile Organic Compounds (AVOC) as a Source of Urban Air Pollution
Researcher (PI) Matti Rissanen
Host Institution (HI) TAMPEREEN KORKEAKOULUSAATIO SR
Country Finland
Call Details Consolidator Grant (CoG), PE10, ERC-2020-COG
Summary Previous efforts to raise living standards have been based on relentlessly increasing combustion, causing environmental destruction at all scales. In addition to climate-warming CO2, fossil fuel combustion also produces a large number of organic compounds and particulate matter, which deteriorate air quality.
The atmosphere is cleansed from such pollutants by gas-phase oxidation reactions, which are invariably mediated by peroxy radicals (RO2). Oxidation transforms initially volatile and water-insoluble hydrocarbons into water-soluble forms (ultimately CO2), enabling scavenging by liquid droplets. A minor but crucially important alternative oxidation pathway leads to oxidative molecular growth, and formation of atmospheric aerosols. Aerosols impart a huge influence on the atmosphere, from local air quality issues to global climate forcing, yet their formation mechanisms and structures of organic aerosol precursors remains elusive.
In a paradigm change, RO2 was recently found to undergo autoxidation, enabling rapid aerosol precursor formation even at sub-second time-scales – in stark contrast to the long processing times (days - weeks) previously assumed to be necessary. We have shown how abundant biogenic hydrocarbons (BVOC) autoxidize, but due to key structural differences, the same pathways are not available for anthropogenic hydrocarbons (AVOC), and thus they were not expected to autoxidize. My preliminary experiments reveal that AVOCs do autoxidize, but the mechanism enabling this remain unknown. Crucially, the co-reactants shown to inhibit BVOC seem to enforce AVOC autoxidation – potentially explaining the recent mysterious discovery of new-particle formation in polluted megacities. In ADAPT, I will use a combination of novel mass spectrometric detection methods fortified by theoretical calculations, to solve the mechanism of AVOC autoxidation. This will directly assist both air quality management, and the design of cleaner fuels and engines.
Summary
Previous efforts to raise living standards have been based on relentlessly increasing combustion, causing environmental destruction at all scales. In addition to climate-warming CO2, fossil fuel combustion also produces a large number of organic compounds and particulate matter, which deteriorate air quality.
The atmosphere is cleansed from such pollutants by gas-phase oxidation reactions, which are invariably mediated by peroxy radicals (RO2). Oxidation transforms initially volatile and water-insoluble hydrocarbons into water-soluble forms (ultimately CO2), enabling scavenging by liquid droplets. A minor but crucially important alternative oxidation pathway leads to oxidative molecular growth, and formation of atmospheric aerosols. Aerosols impart a huge influence on the atmosphere, from local air quality issues to global climate forcing, yet their formation mechanisms and structures of organic aerosol precursors remains elusive.
In a paradigm change, RO2 was recently found to undergo autoxidation, enabling rapid aerosol precursor formation even at sub-second time-scales – in stark contrast to the long processing times (days - weeks) previously assumed to be necessary. We have shown how abundant biogenic hydrocarbons (BVOC) autoxidize, but due to key structural differences, the same pathways are not available for anthropogenic hydrocarbons (AVOC), and thus they were not expected to autoxidize. My preliminary experiments reveal that AVOCs do autoxidize, but the mechanism enabling this remain unknown. Crucially, the co-reactants shown to inhibit BVOC seem to enforce AVOC autoxidation – potentially explaining the recent mysterious discovery of new-particle formation in polluted megacities. In ADAPT, I will use a combination of novel mass spectrometric detection methods fortified by theoretical calculations, to solve the mechanism of AVOC autoxidation. This will directly assist both air quality management, and the design of cleaner fuels and engines.
Max ERC Funding
2 689 147 €
Duration
Start date: 2021-02-01, End date: 2026-01-31
Project acronym ANTILEAK
Project Development of antagonists of vascular leakage
Researcher (PI) Pipsa SAHARINEN
Host Institution (HI) HELSINGIN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), LS4, ERC-2017-COG
Summary Dysregulation of capillary permeability is a severe problem in critically ill patients, but the mechanisms involved are poorly understood. Further, there are no targeted therapies to stabilize leaky vessels in various common, potentially fatal diseases, such as systemic inflammation and sepsis, which affect millions of people annually. Although a multitude of signals that stimulate opening of endothelial cell-cell junctions leading to permeability have been characterized using cellular and in vivo models, approaches to reverse the harmful process of capillary leakage in disease conditions are yet to be identified. I propose to explore a novel autocrine endothelial permeability regulatory system as a potentially universal mechanism that antagonizes vascular stabilizing ques and sustains vascular leakage in inflammation. My group has identified inflammation-induced mechanisms that switch vascular stabilizing factors into molecules that destabilize vascular barriers, and identified tools to prevent the barrier disruption. Building on these discoveries, my group will use mouse genetics, structural biology and innovative, systematic antibody development coupled with gene editing and gene silencing technology, in order to elucidate mechanisms of vascular barrier breakdown and repair in systemic inflammation. The expected outcomes include insights into endothelial cell signaling and permeability regulation, and preclinical proof-of-concept antibodies to control endothelial activation and vascular leakage in systemic inflammation and sepsis models. Ultimately, the new knowledge and preclinical tools developed in this project may facilitate future development of targeted approaches against vascular leakage.
Summary
Dysregulation of capillary permeability is a severe problem in critically ill patients, but the mechanisms involved are poorly understood. Further, there are no targeted therapies to stabilize leaky vessels in various common, potentially fatal diseases, such as systemic inflammation and sepsis, which affect millions of people annually. Although a multitude of signals that stimulate opening of endothelial cell-cell junctions leading to permeability have been characterized using cellular and in vivo models, approaches to reverse the harmful process of capillary leakage in disease conditions are yet to be identified. I propose to explore a novel autocrine endothelial permeability regulatory system as a potentially universal mechanism that antagonizes vascular stabilizing ques and sustains vascular leakage in inflammation. My group has identified inflammation-induced mechanisms that switch vascular stabilizing factors into molecules that destabilize vascular barriers, and identified tools to prevent the barrier disruption. Building on these discoveries, my group will use mouse genetics, structural biology and innovative, systematic antibody development coupled with gene editing and gene silencing technology, in order to elucidate mechanisms of vascular barrier breakdown and repair in systemic inflammation. The expected outcomes include insights into endothelial cell signaling and permeability regulation, and preclinical proof-of-concept antibodies to control endothelial activation and vascular leakage in systemic inflammation and sepsis models. Ultimately, the new knowledge and preclinical tools developed in this project may facilitate future development of targeted approaches against vascular leakage.
Max ERC Funding
1 999 770 €
Duration
Start date: 2018-05-01, End date: 2023-04-30
Project acronym CATCH
Project Cross-dimensional Activation of Two-Dimensional Semiconductors for Photocatalytic Heterojunctions
Researcher (PI) Wei CAO
Host Institution (HI) OULUN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), PE8, ERC-2020-COG
Summary Spacetime defines existence and evolution of materials. A key path to human’s sustainability through materials innovation can hardly circumvent materials dimensionalities. Despite numerous studies in electrically distinct 2D semiconductors, the route to engage them in high-performance photocatalysts remains elusive. Herein, CATCH proposes a cross-dimensional activation strategy of 2D semiconductors to implement practical photocatalysis. It operates electronic structures of dimensionally paradoxical 2D semiconductors and spatially limited nD (n=0-2) guests, directs charge migration processes, mass-produces advanced catalysts and elucidates time-evolved catalysis. Synergic impacts crossing 2D-nD will lead to > 95%/hour rates for pollutant removal and >20% quantum efficiencies for H2 evolution under visible light. CATCH enumerates chemical coordination and writes reaction equations with sub-nanosecond precision.
CATCH employs density functional theory optimization and data mining prediction to select most probable heterojunctional peers from hetero/homo- dimensions. Through facile but efficient wet and dry synthesis, nanostructures will be bonded to basal planes or brinks of 2D slabs. CATCH benefits in-house techniques for product characterizations and refinements and emphasizes on cutting-edge in situ studies to unveil photocatalysis at advanced photon sources. Assisted with theoretical modelling, ambient and time-evolved experiments will illustrate photocatalytic dynamics and kinetics in mixed spacetime.
CATCH unites low-dimensional materials designs by counting physical and electronic merits from spacetime confinements. It metrologically elaborates photocatalysis in an elevated 2D+nD+t, alters passages of materials combinations crossing dimensions, and directs future photocatalyst designs. Standing on cross-dimensional materials innovation and photocatalysis study, CATCH breaks the deadlock of practical photocatalysis that eventually leads to sustainability.
Summary
Spacetime defines existence and evolution of materials. A key path to human’s sustainability through materials innovation can hardly circumvent materials dimensionalities. Despite numerous studies in electrically distinct 2D semiconductors, the route to engage them in high-performance photocatalysts remains elusive. Herein, CATCH proposes a cross-dimensional activation strategy of 2D semiconductors to implement practical photocatalysis. It operates electronic structures of dimensionally paradoxical 2D semiconductors and spatially limited nD (n=0-2) guests, directs charge migration processes, mass-produces advanced catalysts and elucidates time-evolved catalysis. Synergic impacts crossing 2D-nD will lead to > 95%/hour rates for pollutant removal and >20% quantum efficiencies for H2 evolution under visible light. CATCH enumerates chemical coordination and writes reaction equations with sub-nanosecond precision.
CATCH employs density functional theory optimization and data mining prediction to select most probable heterojunctional peers from hetero/homo- dimensions. Through facile but efficient wet and dry synthesis, nanostructures will be bonded to basal planes or brinks of 2D slabs. CATCH benefits in-house techniques for product characterizations and refinements and emphasizes on cutting-edge in situ studies to unveil photocatalysis at advanced photon sources. Assisted with theoretical modelling, ambient and time-evolved experiments will illustrate photocatalytic dynamics and kinetics in mixed spacetime.
CATCH unites low-dimensional materials designs by counting physical and electronic merits from spacetime confinements. It metrologically elaborates photocatalysis in an elevated 2D+nD+t, alters passages of materials combinations crossing dimensions, and directs future photocatalyst designs. Standing on cross-dimensional materials innovation and photocatalysis study, CATCH breaks the deadlock of practical photocatalysis that eventually leads to sustainability.
Max ERC Funding
1 999 946 €
Duration
Start date: 2021-05-01, End date: 2026-04-30
Project acronym CAVITYQPD
Project Cavity quantum phonon dynamics
Researcher (PI) Mika Antero Sillanpaeae
Host Institution (HI) AALTO KORKEAKOULUSAATIO SR
Country Finland
Call Details Consolidator Grant (CoG), PE3, ERC-2013-CoG
Summary "Large bodies usually follow the classical equations of motion. Deviations from this can be called
macroscopic quantum behavior. These phenomena have been experimentally verified with cavity Quantum
Electro Dynamics (QED), trapped ions, and superconducting Josephson junction systems. Recently, evidence
was obtained that also moving objects can display such behavior. These objects are micromechanical
resonators (MR), which can measure tens of microns in size and are hence quite macroscopic. The degree of
freedom is their vibrations: phonons.
I propose experimental research in order to push quantum mechanics closer to the classical world than ever
before. I will try find quantum behavior in the most classical objects, that is, slowly moving bodies. I will use
MR's, accessed via electrical resonators. Part of it will be in analogy to the previously studied macroscopic
systems, but with photons replaced by phonons. The experiments are done in a cryogenic temperature mostly
in dilution refrigerator. The work will open up new perspectives on how nature works, and can have
technological implications.
The first basic setup is the coupling of MR to microwave cavity resonators. This is a direct analogy to
optomechanics, and can be called circuit optomechanics. The goals will be phonon state transfer via a cavity
bus, construction of squeezed states and of phonon-cavity entanglement. The second setup is to boost the
optomechanical coupling with a Josephson junction system, and reach the single-phonon strong-coupling for
the first time. The third setup is the coupling of MR to a Josephson junction artificial atom. Here we will
access the MR same way as the motion of a trapped ions is coupled to their internal transitions. In this setup,
I am proposing to construct exotic quantum states of motion, and finally entangle and transfer phonons over
mm-distance via cavity-coupled qubits. I believe within the project it is possible to perform rudimentary Bell
measurement with phonons."
Summary
"Large bodies usually follow the classical equations of motion. Deviations from this can be called
macroscopic quantum behavior. These phenomena have been experimentally verified with cavity Quantum
Electro Dynamics (QED), trapped ions, and superconducting Josephson junction systems. Recently, evidence
was obtained that also moving objects can display such behavior. These objects are micromechanical
resonators (MR), which can measure tens of microns in size and are hence quite macroscopic. The degree of
freedom is their vibrations: phonons.
I propose experimental research in order to push quantum mechanics closer to the classical world than ever
before. I will try find quantum behavior in the most classical objects, that is, slowly moving bodies. I will use
MR's, accessed via electrical resonators. Part of it will be in analogy to the previously studied macroscopic
systems, but with photons replaced by phonons. The experiments are done in a cryogenic temperature mostly
in dilution refrigerator. The work will open up new perspectives on how nature works, and can have
technological implications.
The first basic setup is the coupling of MR to microwave cavity resonators. This is a direct analogy to
optomechanics, and can be called circuit optomechanics. The goals will be phonon state transfer via a cavity
bus, construction of squeezed states and of phonon-cavity entanglement. The second setup is to boost the
optomechanical coupling with a Josephson junction system, and reach the single-phonon strong-coupling for
the first time. The third setup is the coupling of MR to a Josephson junction artificial atom. Here we will
access the MR same way as the motion of a trapped ions is coupled to their internal transitions. In this setup,
I am proposing to construct exotic quantum states of motion, and finally entangle and transfer phonons over
mm-distance via cavity-coupled qubits. I believe within the project it is possible to perform rudimentary Bell
measurement with phonons."
Max ERC Funding
2 004 283 €
Duration
Start date: 2015-01-01, End date: 2019-12-31
Project acronym ECLAIR
Project Emulation of subgrid-scale aerosol-cloud interactions in climate models: towards a realistic representation of aerosol indirect effect
Researcher (PI) Sari Hannele Korhonen
Host Institution (HI) ILMATIETEEN LAITOS
Country Finland
Call Details Consolidator Grant (CoG), PE10, ERC-2014-CoG
Summary I propose to develop an innovative interdisciplinary model framework to refine the estimate of aerosol indirect effect (i.e. influence of atmospheric aerosol particles on cloud properties), which remains the single largest uncertainty in the current drivers of climate change.
A major reason for this uncertainty is that current climate models are unable to resolve the spatial scales for aerosol-cloud interactions. We will resolve this scale problem by using statistical emulation to build computationally fast surrogate models (i.e. emulators) that can reproduce the effective output of a detailed high-resolution cloud-resolving model. By incorporating these emulators into a state-of-the-science climate model, we will for the first time achieve the accuracy of a limited-area high-resolution model on a global scale with negligible computational cost.
The main scientific outcome of the project will be a highly refined and physically sound estimate of the aerosol indirect effect that enables more accurate projections of future climate change, and thus has high societal relevance. In addition, the developed emulators will help to quantify how the remaining uncertainties in aerosol properties propagate to predictions of aerosol indirect effect. This information will be used, together with an extensive set of remote sensing, in-situ and laboratory data from our collaborators, to improve the process-level understanding of aerosol-cloud interactions.
The comprehensive uncertainty analyses performed during this project will be highly valuable for future research efforts as they point to processes and interactions that most urgently need to be experimentally constrained. Furthermore, our pioneering model framework that incorporates emulators to represent subgrid- scale processes will open up completely new research opportunities also in other fields that deal with heterogeneous spatial scales.
Summary
I propose to develop an innovative interdisciplinary model framework to refine the estimate of aerosol indirect effect (i.e. influence of atmospheric aerosol particles on cloud properties), which remains the single largest uncertainty in the current drivers of climate change.
A major reason for this uncertainty is that current climate models are unable to resolve the spatial scales for aerosol-cloud interactions. We will resolve this scale problem by using statistical emulation to build computationally fast surrogate models (i.e. emulators) that can reproduce the effective output of a detailed high-resolution cloud-resolving model. By incorporating these emulators into a state-of-the-science climate model, we will for the first time achieve the accuracy of a limited-area high-resolution model on a global scale with negligible computational cost.
The main scientific outcome of the project will be a highly refined and physically sound estimate of the aerosol indirect effect that enables more accurate projections of future climate change, and thus has high societal relevance. In addition, the developed emulators will help to quantify how the remaining uncertainties in aerosol properties propagate to predictions of aerosol indirect effect. This information will be used, together with an extensive set of remote sensing, in-situ and laboratory data from our collaborators, to improve the process-level understanding of aerosol-cloud interactions.
The comprehensive uncertainty analyses performed during this project will be highly valuable for future research efforts as they point to processes and interactions that most urgently need to be experimentally constrained. Furthermore, our pioneering model framework that incorporates emulators to represent subgrid- scale processes will open up completely new research opportunities also in other fields that deal with heterogeneous spatial scales.
Max ERC Funding
1 999 511 €
Duration
Start date: 2015-09-01, End date: 2020-08-31
Project acronym ETI
Project Epistemic Transitions in Islamic Philosophy, Theology and Science: From the 12th to the 19th Century
Researcher (PI) Jari Pekka Kaukua
Host Institution (HI) JYVASKYLAN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), SH5, ERC-2015-CoG
Summary Not very long ago, it was still common to hold that little of interest took place in Islamic philosophy, theology and science after the death of the Peripatetic commentator Averroes in 1198. Recent research has produced increasing evidence against this view, and experts now commonly agree that texts from the so-called post-classical period merit serious analysis. That evidence, however, is still fragmentary, and we lack a clear understanding of the large scale and long run development in the various fields of Islamic intellectual culture after the twelfth century.
This project will investigate debates concerning the nature and methods of knowledge in four of the most ambitious strands of Islamic theoretical thought, that is, philosophy, theology, natural science, and philosophically inclined Sufism. Its temporal scope extends from the end of the twelfth century to the beginning of the colonial era, and it focuses on foundational epistemological questions (how knowledge is defined, what criteria are used to distinguish it from less secure epistemic attitudes, what methods are identified as valid in the acquisition of knowledge) as well as questions concerning knowledge as the goal of our existence (in particular, whether perceptual experience is inherently valuable).
Our study of the four strands is based on the hypothesis that the post-classical period is witness to a sophisticated discussion of knowledge, in which epistemic realism, intuitionism, phenomenalism, and subjectivism are pitted against each other in a nuanced manner. Hence, the project will result in a well-founded reassessment of the common view according to which post-classical Islamic intellectual culture is authoritarian and stuck to an epistemic paradigm that stifles insight and creativity. Thereby it will provide new ingredients for projects of endogenous reform and reorientation in Islam, and corroborate the view that our future histories of philosophy should incorporate the Islamic tradition.
Summary
Not very long ago, it was still common to hold that little of interest took place in Islamic philosophy, theology and science after the death of the Peripatetic commentator Averroes in 1198. Recent research has produced increasing evidence against this view, and experts now commonly agree that texts from the so-called post-classical period merit serious analysis. That evidence, however, is still fragmentary, and we lack a clear understanding of the large scale and long run development in the various fields of Islamic intellectual culture after the twelfth century.
This project will investigate debates concerning the nature and methods of knowledge in four of the most ambitious strands of Islamic theoretical thought, that is, philosophy, theology, natural science, and philosophically inclined Sufism. Its temporal scope extends from the end of the twelfth century to the beginning of the colonial era, and it focuses on foundational epistemological questions (how knowledge is defined, what criteria are used to distinguish it from less secure epistemic attitudes, what methods are identified as valid in the acquisition of knowledge) as well as questions concerning knowledge as the goal of our existence (in particular, whether perceptual experience is inherently valuable).
Our study of the four strands is based on the hypothesis that the post-classical period is witness to a sophisticated discussion of knowledge, in which epistemic realism, intuitionism, phenomenalism, and subjectivism are pitted against each other in a nuanced manner. Hence, the project will result in a well-founded reassessment of the common view according to which post-classical Islamic intellectual culture is authoritarian and stuck to an epistemic paradigm that stifles insight and creativity. Thereby it will provide new ingredients for projects of endogenous reform and reorientation in Islam, and corroborate the view that our future histories of philosophy should incorporate the Islamic tradition.
Max ERC Funding
1 526 429 €
Duration
Start date: 2016-09-01, End date: 2021-08-31
Project acronym FoTran
Project Found in Translation – Natural Language Understanding with Cross-Lingual Grounding
Researcher (PI) Joerg TIEDEMANN
Host Institution (HI) HELSINGIN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), PE6, ERC-2017-COG
Summary "Natural language understanding is the ""holy grail"" of computational linguistics and a long-term goal in research on artificial intelligence. Understanding human communication is difficult due to the various ambiguities in natural languages and the wide range of contextual dependencies required to resolve them. Discovering the semantics behind language input is necessary for proper interpretation in interactive tools, which requires an abstraction from language-specific forms to language-independent meaning representations. With this project, I propose a line of research that will focus on the development of novel data-driven models that can learn such meaning representations from indirect supervision provided by human translations covering a substantial proportion of the linguistic diversity in the world. A guiding principle is cross-lingual grounding, the effect of resolving ambiguities through translation. The beauty of that idea is the use of naturally occurring data instead of artificially created resources and costly manual annotations. The framework is based on deep learning and neural machine translation and my hypothesis is that training on increasing amounts of linguistically diverse data improves the abstractions found by the model. Eventually, this will lead to universal sentence-level meaning representations and we will test our ideas with multilingual machine translation and tasks that require semantic reasoning and inference."
Summary
"Natural language understanding is the ""holy grail"" of computational linguistics and a long-term goal in research on artificial intelligence. Understanding human communication is difficult due to the various ambiguities in natural languages and the wide range of contextual dependencies required to resolve them. Discovering the semantics behind language input is necessary for proper interpretation in interactive tools, which requires an abstraction from language-specific forms to language-independent meaning representations. With this project, I propose a line of research that will focus on the development of novel data-driven models that can learn such meaning representations from indirect supervision provided by human translations covering a substantial proportion of the linguistic diversity in the world. A guiding principle is cross-lingual grounding, the effect of resolving ambiguities through translation. The beauty of that idea is the use of naturally occurring data instead of artificially created resources and costly manual annotations. The framework is based on deep learning and neural machine translation and my hypothesis is that training on increasing amounts of linguistically diverse data improves the abstractions found by the model. Eventually, this will lead to universal sentence-level meaning representations and we will test our ideas with multilingual machine translation and tasks that require semantic reasoning and inference."
Max ERC Funding
1 817 622 €
Duration
Start date: 2018-09-01, End date: 2023-08-31
Project acronym IPTheoryUnified
Project Inverse boundary problems: toward a unified theory
Researcher (PI) Mikko SALO
Host Institution (HI) JYVASKYLAN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), PE1, ERC-2017-COG
Summary This proposal is concerned with the mathematical theory of inverse problems. This is a vibrant research field at the intersection of pure and applied mathematics, drawing techniques from PDE, geometry, and harmonic analysis as well as generating new research questions inspired by applications. Prominent questions include the Calderón problem related to electrical imaging, the Gel'fand problem related to seismic imaging, and geometric inverse problems such as inversion of the geodesic X-ray transform.
Recently, exciting new connections between these different topics have begun to emerge in the work of the PI and others, such as:
- The explicit appearance of the geodesic X-ray transform in the Calderón problem.
- An unexpected connection between the Calderón and Gel’fand problems involving control theory.
- Pseudo-linearization as a potential unifying principle for reducing nonlinear problems to linear ones.
- The introduction of microlocal normal forms in inverse problems for PDE.
These examples strongly suggest that there is a larger picture behind various different inverse problems, which remains to be fully revealed.
This project will explore the possibility of a unified theory for several inverse boundary problems. Particular objectives include:
1. The use of normal forms and pseudo-linearization as a unified point of view, including reductions to questions in integral geometry and control theory.
2. The solution of integral geometry problems, including the analysis of convex foliations, invertibility of ray transforms, and a systematic Carleman estimate approach to uniqueness results.
3. A theory of inverse problems for nonlocal models based on control theory arguments.
Such a unified theory could have remarkable consequences even in other fields of mathematics, including controllability methods in transport theory, a solution of the boundary rigidity problem in geometry, or a general pseudo-linearization approach for solving nonlinear operator equations.
Summary
This proposal is concerned with the mathematical theory of inverse problems. This is a vibrant research field at the intersection of pure and applied mathematics, drawing techniques from PDE, geometry, and harmonic analysis as well as generating new research questions inspired by applications. Prominent questions include the Calderón problem related to electrical imaging, the Gel'fand problem related to seismic imaging, and geometric inverse problems such as inversion of the geodesic X-ray transform.
Recently, exciting new connections between these different topics have begun to emerge in the work of the PI and others, such as:
- The explicit appearance of the geodesic X-ray transform in the Calderón problem.
- An unexpected connection between the Calderón and Gel’fand problems involving control theory.
- Pseudo-linearization as a potential unifying principle for reducing nonlinear problems to linear ones.
- The introduction of microlocal normal forms in inverse problems for PDE.
These examples strongly suggest that there is a larger picture behind various different inverse problems, which remains to be fully revealed.
This project will explore the possibility of a unified theory for several inverse boundary problems. Particular objectives include:
1. The use of normal forms and pseudo-linearization as a unified point of view, including reductions to questions in integral geometry and control theory.
2. The solution of integral geometry problems, including the analysis of convex foliations, invertibility of ray transforms, and a systematic Carleman estimate approach to uniqueness results.
3. A theory of inverse problems for nonlocal models based on control theory arguments.
Such a unified theory could have remarkable consequences even in other fields of mathematics, including controllability methods in transport theory, a solution of the boundary rigidity problem in geometry, or a general pseudo-linearization approach for solving nonlinear operator equations.
Max ERC Funding
920 880 €
Duration
Start date: 2018-05-01, End date: 2023-04-30
Project acronym KETJU
Project Post-Newtonian modelling of the dynamics of supermassive black holes in galactic-scale hydrodynamical simulations (KETJU)
Researcher (PI) Peter Hilding JOHANSSON
Host Institution (HI) HELSINGIN YLIOPISTO
Country Finland
Call Details Consolidator Grant (CoG), PE9, ERC-2018-COG
Summary Supermassive black holes (SMBHs) with masses in the range ~10^6-10^10 M⊙ are found at the centres of all massive galaxies in the Local Universe. In the ΛCDM picture of structure formation galaxies grow bottom-up through mergers and gas accretion, leading to multiple SMBHs in the same stellar system. Current simulation codes are unable to resolve in a single simulation the full SMBH merging process, which involves dynamical friction, three-body interactions and finally gravitational wave (GW) emission. KETJU will provide a significant breakthrough in SMBH research by following for the first time accurately global galactic-scale dynamical and gaseous astrophysical processes, while simultaneously solving the dynamics of SMBHs, SMBH binaries and surrounding stellar systems at sub-parsec scales. Our code KETJU (the word for 'chain' in Finnish) is built on the GADGET-3 code and it includes regions around every SMBH in which the dynamics of SMBHs and stellar particles is modelled using a non-softened Post-Newtonian algorithmic chain regularisation technique. The remaining simulation particles far from the SMBHs are evolved using softened GADGET-3. Using KETJU we can study at unprecedented accuracy the dynamics of SMBHs to separations of ~10 Schwarzschild radii, the formation of cores in massive galaxies, the formation of nuclear stellar clusters and finally provide a realistic prediction for the amplitude and frequency distribution of the cosmological gravitational wave background. The UH theoretical extragalactic team is ideally suited for this project, as it has an unusually versatile background in modelling the dynamics, feedback and merging of SMBHs. KETJU is also particularly timely, as the spectacular direct detection of GWs in 2016 is paving the way for a new era in gravitational wave astronomy. Future space-borne GW observatories, such as the European Space Agency's LISA, require accurate global GW predictions in order to fully realise their science goals.
Summary
Supermassive black holes (SMBHs) with masses in the range ~10^6-10^10 M⊙ are found at the centres of all massive galaxies in the Local Universe. In the ΛCDM picture of structure formation galaxies grow bottom-up through mergers and gas accretion, leading to multiple SMBHs in the same stellar system. Current simulation codes are unable to resolve in a single simulation the full SMBH merging process, which involves dynamical friction, three-body interactions and finally gravitational wave (GW) emission. KETJU will provide a significant breakthrough in SMBH research by following for the first time accurately global galactic-scale dynamical and gaseous astrophysical processes, while simultaneously solving the dynamics of SMBHs, SMBH binaries and surrounding stellar systems at sub-parsec scales. Our code KETJU (the word for 'chain' in Finnish) is built on the GADGET-3 code and it includes regions around every SMBH in which the dynamics of SMBHs and stellar particles is modelled using a non-softened Post-Newtonian algorithmic chain regularisation technique. The remaining simulation particles far from the SMBHs are evolved using softened GADGET-3. Using KETJU we can study at unprecedented accuracy the dynamics of SMBHs to separations of ~10 Schwarzschild radii, the formation of cores in massive galaxies, the formation of nuclear stellar clusters and finally provide a realistic prediction for the amplitude and frequency distribution of the cosmological gravitational wave background. The UH theoretical extragalactic team is ideally suited for this project, as it has an unusually versatile background in modelling the dynamics, feedback and merging of SMBHs. KETJU is also particularly timely, as the spectacular direct detection of GWs in 2016 is paving the way for a new era in gravitational wave astronomy. Future space-borne GW observatories, such as the European Space Agency's LISA, require accurate global GW predictions in order to fully realise their science goals.
Max ERC Funding
1 953 569 €
Duration
Start date: 2019-07-01, End date: 2024-06-30