Project acronym 3D_Tryps
Project The role of three-dimensional genome architecture in antigenic variation
Researcher (PI) Tim Nicolai SIEGEL
Host Institution (HI) LUDWIG-MAXIMILIANS-UNIVERSITAET MUENCHEN
Country Germany
Call Details Starting Grant (StG), LS6, ERC-2016-STG
Summary Antigenic variation is a widely employed strategy to evade the host immune response. It has similar functional requirements even in evolutionarily divergent pathogens. These include the mutually exclusive expression of antigens and the periodic, nonrandom switching in the expression of different antigens during the course of an infection. Despite decades of research the mechanisms of antigenic variation are not fully understood in any organism.
The recent development of high-throughput sequencing-based assays to probe the 3D genome architecture (Hi-C) has revealed the importance of the spatial organization of DNA inside the nucleus. 3D genome architecture plays a critical role in the regulation of mutually exclusive gene expression and the frequency of translocation between different genomic loci in many eukaryotes. Thus, genome architecture may also be a key regulator of antigenic variation, yet the causal links between genome architecture and the expression of antigens have not been studied systematically. In addition, the development of CRISPR-Cas9-based approaches to perform nucleotide-specific genome editing has opened unprecedented opportunities to study the influence of DNA sequence elements on the spatial organization of DNA and how this impacts antigen expression.
I have adapted both Hi-C and CRISPR-Cas9 technology to the protozoan parasite Trypanosoma brucei, one of the most important model organisms to study antigenic variation. These techniques will enable me to bridge the field of antigenic variation research with that of genome architecture. I will perform the first systematic analysis of the role of genome architecture in the mutually exclusive and hierarchical expression of antigens in any pathogen.
The experiments outlined in this proposal will provide new insight, facilitating a new view of antigenic variation and may eventually help medical intervention in T. brucei and in other pathogens relying on antigenic variation for their survival.
Summary
Antigenic variation is a widely employed strategy to evade the host immune response. It has similar functional requirements even in evolutionarily divergent pathogens. These include the mutually exclusive expression of antigens and the periodic, nonrandom switching in the expression of different antigens during the course of an infection. Despite decades of research the mechanisms of antigenic variation are not fully understood in any organism.
The recent development of high-throughput sequencing-based assays to probe the 3D genome architecture (Hi-C) has revealed the importance of the spatial organization of DNA inside the nucleus. 3D genome architecture plays a critical role in the regulation of mutually exclusive gene expression and the frequency of translocation between different genomic loci in many eukaryotes. Thus, genome architecture may also be a key regulator of antigenic variation, yet the causal links between genome architecture and the expression of antigens have not been studied systematically. In addition, the development of CRISPR-Cas9-based approaches to perform nucleotide-specific genome editing has opened unprecedented opportunities to study the influence of DNA sequence elements on the spatial organization of DNA and how this impacts antigen expression.
I have adapted both Hi-C and CRISPR-Cas9 technology to the protozoan parasite Trypanosoma brucei, one of the most important model organisms to study antigenic variation. These techniques will enable me to bridge the field of antigenic variation research with that of genome architecture. I will perform the first systematic analysis of the role of genome architecture in the mutually exclusive and hierarchical expression of antigens in any pathogen.
The experiments outlined in this proposal will provide new insight, facilitating a new view of antigenic variation and may eventually help medical intervention in T. brucei and in other pathogens relying on antigenic variation for their survival.
Max ERC Funding
1 498 175 €
Duration
Start date: 2017-04-01, End date: 2022-03-31
Project acronym 5D Heart Patch
Project A Functional, Mature In vivo Human Ventricular Muscle Patch for Cardiomyopathy
Researcher (PI) Kenneth Randall Chien
Host Institution (HI) KAROLINSKA INSTITUTET
Country Sweden
Call Details Advanced Grant (AdG), LS7, ERC-2016-ADG
Summary Developing new therapeutic strategies for heart regeneration is a major goal for cardiac biology and medicine. While cardiomyocytes can be generated from human pluripotent stem (hPSC) cells in vitro, it has proven difficult to use these cells to generate a large scale, mature human heart ventricular muscle graft on the injured heart in vivo. The central objective of this proposal is to optimize the generation of a large-scale pure, fully functional human ventricular muscle patch in vivo through the self-assembly of purified human ventricular progenitors and the localized expression of defined paracrine factors that drive their expansion, differentiation, vascularization, matrix formation, and maturation. Recently, we have found that purified hPSC-derived ventricular progenitors (HVPs) can self-assemble in vivo on the epicardial surface into a 3D vascularized, and functional ventricular patch with its own extracellular matrix via a cell autonomous pathway. A two-step protocol and FACS purification of HVP receptors can generate billions of pure HVPs- The current proposal will lead to the identification of defined paracrine pathways to enhance the survival, grafting/implantation, expansion, differentiation, matrix formation, vascularization and maturation of the graft in vivo. We will captalize on our unique HVP system and our novel modRNA technology to deliver therapeutic strategies by using the in vivo human ventricular muscle to model in vivo arrhythmogenic cardiomyopathy, and optimize the ability of the graft to compensate for the massive loss of functional muscle during ischemic cardiomyopathy and post-myocardial infarction. The studies will lead to new in vivo chimeric models of human cardiac disease and an experimental paradigm to optimize organ-on-organ cardiac tissue engineers of an in vivo, functional mature ventricular patch for cardiomyopathy
Summary
Developing new therapeutic strategies for heart regeneration is a major goal for cardiac biology and medicine. While cardiomyocytes can be generated from human pluripotent stem (hPSC) cells in vitro, it has proven difficult to use these cells to generate a large scale, mature human heart ventricular muscle graft on the injured heart in vivo. The central objective of this proposal is to optimize the generation of a large-scale pure, fully functional human ventricular muscle patch in vivo through the self-assembly of purified human ventricular progenitors and the localized expression of defined paracrine factors that drive their expansion, differentiation, vascularization, matrix formation, and maturation. Recently, we have found that purified hPSC-derived ventricular progenitors (HVPs) can self-assemble in vivo on the epicardial surface into a 3D vascularized, and functional ventricular patch with its own extracellular matrix via a cell autonomous pathway. A two-step protocol and FACS purification of HVP receptors can generate billions of pure HVPs- The current proposal will lead to the identification of defined paracrine pathways to enhance the survival, grafting/implantation, expansion, differentiation, matrix formation, vascularization and maturation of the graft in vivo. We will captalize on our unique HVP system and our novel modRNA technology to deliver therapeutic strategies by using the in vivo human ventricular muscle to model in vivo arrhythmogenic cardiomyopathy, and optimize the ability of the graft to compensate for the massive loss of functional muscle during ischemic cardiomyopathy and post-myocardial infarction. The studies will lead to new in vivo chimeric models of human cardiac disease and an experimental paradigm to optimize organ-on-organ cardiac tissue engineers of an in vivo, functional mature ventricular patch for cardiomyopathy
Max ERC Funding
2 149 228 €
Duration
Start date: 2017-12-01, End date: 2022-11-30
Project acronym AAA
Project Adaptive Actin Architectures
Researcher (PI) Laurent Blanchoin
Host Institution (HI) CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
Country France
Call Details Advanced Grant (AdG), LS3, ERC-2016-ADG
Summary Although we have extensive knowledge of many important processes in cell biology, including information on many of the molecules involved and the physical interactions among them, we still do not understand most of the dynamical features that are the essence of living systems. This is particularly true for the actin cytoskeleton, a major component of the internal architecture of eukaryotic cells. In living cells, actin networks constantly assemble and disassemble filaments while maintaining an apparent stable structure, suggesting a perfect balance between the two processes. Such behaviors are called “dynamic steady states”. They confer upon actin networks a high degree of plasticity allowing them to adapt in response to external changes and enable cells to adjust to their environments. Despite their fundamental importance in the regulation of cell physiology, the basic mechanisms that control the coordinated dynamics of co-existing actin networks are poorly understood. In the AAA project, first, we will characterize the parameters that allow the coupling among co-existing actin networks at steady state. In vitro reconstituted systems will be used to control the actin nucleation patterns, the closed volume of the reaction chamber and the physical interaction of the networks. We hope to unravel the mechanism allowing the global coherence of a dynamic actin cytoskeleton. Second, we will use our unique capacity to perform dynamic micropatterning, to add or remove actin nucleation sites in real time, in order to investigate the ability of dynamic networks to adapt to changes and the role of coupled network dynamics in this emergent property. In this part, in vitro experiments will be complemented by the analysis of actin network remodeling in living cells. In the end, our project will provide a comprehensive understanding of how the adaptive response of the cytoskeleton derives from the complex interplay between its biochemical, structural and mechanical properties.
Summary
Although we have extensive knowledge of many important processes in cell biology, including information on many of the molecules involved and the physical interactions among them, we still do not understand most of the dynamical features that are the essence of living systems. This is particularly true for the actin cytoskeleton, a major component of the internal architecture of eukaryotic cells. In living cells, actin networks constantly assemble and disassemble filaments while maintaining an apparent stable structure, suggesting a perfect balance between the two processes. Such behaviors are called “dynamic steady states”. They confer upon actin networks a high degree of plasticity allowing them to adapt in response to external changes and enable cells to adjust to their environments. Despite their fundamental importance in the regulation of cell physiology, the basic mechanisms that control the coordinated dynamics of co-existing actin networks are poorly understood. In the AAA project, first, we will characterize the parameters that allow the coupling among co-existing actin networks at steady state. In vitro reconstituted systems will be used to control the actin nucleation patterns, the closed volume of the reaction chamber and the physical interaction of the networks. We hope to unravel the mechanism allowing the global coherence of a dynamic actin cytoskeleton. Second, we will use our unique capacity to perform dynamic micropatterning, to add or remove actin nucleation sites in real time, in order to investigate the ability of dynamic networks to adapt to changes and the role of coupled network dynamics in this emergent property. In this part, in vitro experiments will be complemented by the analysis of actin network remodeling in living cells. In the end, our project will provide a comprehensive understanding of how the adaptive response of the cytoskeleton derives from the complex interplay between its biochemical, structural and mechanical properties.
Max ERC Funding
2 349 898 €
Duration
Start date: 2017-09-01, End date: 2022-08-31
Project acronym ACCENT
Project Unravelling the architecture and the cartography of the human centriole
Researcher (PI) Paul, Philippe, Desire GUICHARD
Host Institution (HI) UNIVERSITE DE GENEVE
Country Switzerland
Call Details Starting Grant (StG), LS1, ERC-2016-STG
Summary The centriole is the largest evolutionary conserved macromolecular structure responsible for building centrosomes and cilia or flagella in many eukaryotes. Centrioles are critical for the proper execution of important biological processes ranging from cell division to cell signaling. Moreover, centriolar defects have been associated to several human pathologies including ciliopathies and cancer. This state of facts emphasizes the importance of understanding centriole biogenesis. The study of centriole formation is a deep-rooted question, however our current knowledge on its molecular organization at high resolution remains fragmented and limited. In particular, exquisite details of the overall molecular architecture of the human centriole and in particular of its central core region are lacking to understand the basis of centriole organization and function. Resolving this important question represents a challenge that needs to be undertaken and will undoubtedly lead to groundbreaking advances. Another important question to tackle next is to develop innovative methods to enable the nanometric molecular mapping of centriolar proteins within distinct architectural elements of the centriole. This missing information will be key to unravel the molecular mechanisms behind centriolar organization.
This research proposal aims at building a cartography of the human centriole by elucidating its molecular composition and architecture. To this end, we will combine the use of innovative and multidisciplinary techniques encompassing spatial proteomics, cryo-electron tomography, state-of-the-art microscopy and in vitro assays and to achieve a comprehensive molecular and structural view of the human centriole. All together, we expect that these advances will help understand basic principles underlying centriole and cilia formation as well as might have further relevance for human health.
Summary
The centriole is the largest evolutionary conserved macromolecular structure responsible for building centrosomes and cilia or flagella in many eukaryotes. Centrioles are critical for the proper execution of important biological processes ranging from cell division to cell signaling. Moreover, centriolar defects have been associated to several human pathologies including ciliopathies and cancer. This state of facts emphasizes the importance of understanding centriole biogenesis. The study of centriole formation is a deep-rooted question, however our current knowledge on its molecular organization at high resolution remains fragmented and limited. In particular, exquisite details of the overall molecular architecture of the human centriole and in particular of its central core region are lacking to understand the basis of centriole organization and function. Resolving this important question represents a challenge that needs to be undertaken and will undoubtedly lead to groundbreaking advances. Another important question to tackle next is to develop innovative methods to enable the nanometric molecular mapping of centriolar proteins within distinct architectural elements of the centriole. This missing information will be key to unravel the molecular mechanisms behind centriolar organization.
This research proposal aims at building a cartography of the human centriole by elucidating its molecular composition and architecture. To this end, we will combine the use of innovative and multidisciplinary techniques encompassing spatial proteomics, cryo-electron tomography, state-of-the-art microscopy and in vitro assays and to achieve a comprehensive molecular and structural view of the human centriole. All together, we expect that these advances will help understand basic principles underlying centriole and cilia formation as well as might have further relevance for human health.
Max ERC Funding
1 498 965 €
Duration
Start date: 2017-01-01, End date: 2021-12-31
Project acronym AgeConsolidate
Project The Missing Link of Episodic Memory Decline in Aging: The Role of Inefficient Systems Consolidation
Researcher (PI) Anders Martin FJELL
Host Institution (HI) UNIVERSITETET I OSLO
Country Norway
Call Details Consolidator Grant (CoG), SH4, ERC-2016-COG
Summary Which brain mechanisms are responsible for the faith of the memories we make with age, whether they wither or stay, and in what form? Episodic memory function does decline with age. While this decline can have multiple causes, research has focused almost entirely on encoding and retrieval processes, largely ignoring a third critical process– consolidation. The objective of AgeConsolidate is to provide this missing link, by combining novel experimental cognitive paradigms with neuroimaging in a longitudinal large-scale attempt to directly test how age-related changes in consolidation processes in the brain impact episodic memory decline. The ambitious aims of the present proposal are two-fold:
(1) Use recent advances in memory consolidation theory to achieve an elaborate model of episodic memory deficits in aging
(2) Use aging as a model to uncover how structural and functional brain changes affect episodic memory consolidation in general
The novelty of the project lies in the synthesis of recent methodological advances and theoretical models for episodic memory consolidation to explain age-related decline, by employing a unique combination of a range of different techniques and approaches. This is ground-breaking, in that it aims at taking our understanding of the brain processes underlying episodic memory decline in aging to a new level, while at the same time advancing our theoretical understanding of how episodic memories are consolidated in the human brain. To obtain this outcome, I will test the main hypothesis of the project: Brain processes of episodic memory consolidation are less effective in older adults, and this can account for a significant portion of the episodic memory decline in aging. This will be answered by six secondary hypotheses, with 1-3 experiments or tasks designated to address each hypothesis, focusing on functional and structural MRI, positron emission tomography data and sleep experiments to target consolidation from different angles.
Summary
Which brain mechanisms are responsible for the faith of the memories we make with age, whether they wither or stay, and in what form? Episodic memory function does decline with age. While this decline can have multiple causes, research has focused almost entirely on encoding and retrieval processes, largely ignoring a third critical process– consolidation. The objective of AgeConsolidate is to provide this missing link, by combining novel experimental cognitive paradigms with neuroimaging in a longitudinal large-scale attempt to directly test how age-related changes in consolidation processes in the brain impact episodic memory decline. The ambitious aims of the present proposal are two-fold:
(1) Use recent advances in memory consolidation theory to achieve an elaborate model of episodic memory deficits in aging
(2) Use aging as a model to uncover how structural and functional brain changes affect episodic memory consolidation in general
The novelty of the project lies in the synthesis of recent methodological advances and theoretical models for episodic memory consolidation to explain age-related decline, by employing a unique combination of a range of different techniques and approaches. This is ground-breaking, in that it aims at taking our understanding of the brain processes underlying episodic memory decline in aging to a new level, while at the same time advancing our theoretical understanding of how episodic memories are consolidated in the human brain. To obtain this outcome, I will test the main hypothesis of the project: Brain processes of episodic memory consolidation are less effective in older adults, and this can account for a significant portion of the episodic memory decline in aging. This will be answered by six secondary hypotheses, with 1-3 experiments or tasks designated to address each hypothesis, focusing on functional and structural MRI, positron emission tomography data and sleep experiments to target consolidation from different angles.
Max ERC Funding
1 999 482 €
Duration
Start date: 2017-05-01, End date: 2022-04-30
Project acronym ALEXANDRIA
Project Large-Scale Formal Proof for the Working Mathematician
Researcher (PI) Lawrence PAULSON
Host Institution (HI) THE CHANCELLOR MASTERS AND SCHOLARS OF THE UNIVERSITY OF CAMBRIDGE
Country United Kingdom
Call Details Advanced Grant (AdG), PE6, ERC-2016-ADG
Summary Mathematical proofs have always been prone to error. Today, proofs can be hundreds of pages long and combine results from many specialisms, making them almost impossible to check. One solution is to deploy modern verification technology. Interactive theorem provers have demonstrated their potential as vehicles for formalising mathematics through achievements such as the verification of the Kepler Conjecture. Proofs done using such tools reach a high standard of correctness.
However, existing theorem provers are unsuitable for mathematics. Their formal proofs are unreadable. They struggle to do simple tasks, such as evaluating limits. They lack much basic mathematics, and the material they do have is difficult to locate and apply.
ALEXANDRIA will create a proof development environment attractive to working mathematicians, utilising the best technology available across computer science. Its focus will be the management and use of large-scale mathematical knowledge, both theorems and algorithms. The project will employ mathematicians to investigate the formalisation of mathematics in practice. Our already substantial formalised libraries will serve as the starting point. They will be extended and annotated to support sophisticated searches. Techniques will be borrowed from machine learning, information retrieval and natural language processing. Algorithms will be treated similarly: ALEXANDRIA will help users find and invoke the proof methods and algorithms appropriate for the task.
ALEXANDRIA will provide (1) comprehensive formal mathematical libraries; (2) search within libraries, and the mining of libraries for proof patterns; (3) automated support for the construction of large formal proofs; (4) sound and practical computer algebra tools.
ALEXANDRIA will be based on legible structured proofs. Formal proofs should be not mere code, but a machine-checkable form of communication between mathematicians.
Summary
Mathematical proofs have always been prone to error. Today, proofs can be hundreds of pages long and combine results from many specialisms, making them almost impossible to check. One solution is to deploy modern verification technology. Interactive theorem provers have demonstrated their potential as vehicles for formalising mathematics through achievements such as the verification of the Kepler Conjecture. Proofs done using such tools reach a high standard of correctness.
However, existing theorem provers are unsuitable for mathematics. Their formal proofs are unreadable. They struggle to do simple tasks, such as evaluating limits. They lack much basic mathematics, and the material they do have is difficult to locate and apply.
ALEXANDRIA will create a proof development environment attractive to working mathematicians, utilising the best technology available across computer science. Its focus will be the management and use of large-scale mathematical knowledge, both theorems and algorithms. The project will employ mathematicians to investigate the formalisation of mathematics in practice. Our already substantial formalised libraries will serve as the starting point. They will be extended and annotated to support sophisticated searches. Techniques will be borrowed from machine learning, information retrieval and natural language processing. Algorithms will be treated similarly: ALEXANDRIA will help users find and invoke the proof methods and algorithms appropriate for the task.
ALEXANDRIA will provide (1) comprehensive formal mathematical libraries; (2) search within libraries, and the mining of libraries for proof patterns; (3) automated support for the construction of large formal proofs; (4) sound and practical computer algebra tools.
ALEXANDRIA will be based on legible structured proofs. Formal proofs should be not mere code, but a machine-checkable form of communication between mathematicians.
Max ERC Funding
2 430 140 €
Duration
Start date: 2017-09-01, End date: 2022-08-31
Project acronym ALFA
Project Shaping a European Scientific Scene : Alfonsine Astronomy
Researcher (PI) Matthieu Husson
Host Institution (HI) CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRS
Country France
Call Details Consolidator Grant (CoG), SH6, ERC-2016-COG
Summary Alfonsine astronomy is arguably among the first European scientific achievements. It shaped a scene for actors like Regiomontanus or Copernicus. There is however little detailed historical analysis encompassing its development in its full breadth. ALFA addresses this issue by studying tables, instruments, mathematical and theoretical texts in a methodologically innovative way relying on approaches from the history of manuscript cultures, history of mathematics, and history of astronomy.
ALFA integrates these approaches not only to benefit from different perspectives but also to build new questions from their interactions. For instance the analysis of mathematical practices in astral sciences manuscripts induces new ways to analyse the documents and to think about astronomical questions.
Relying on these approaches the main objectives of ALFA are thus to:
- Retrace the development of the corpus of Alfonsine texts from its origin in the second half of the 13th century to the end of the 15th century by following, on the manuscript level, the milieus fostering it;
- Analyse the Alfonsine astronomers’ practices, their relations to mathematics, to the natural world, to proofs and justification, their intellectual context and audiences;
- Build a meaningful narrative showing how astronomers in different milieus with diverse practices shaped, also from Arabic materials, an original scientific scene in Europe.
ALFA will shed new light on the intellectual history of the late medieval period as a whole and produce a better understanding of its relations to related scientific periods in Europe and beyond. It will also produce methodological breakthroughs impacting the ways history of knowledge is practiced outside the field of ancient and medieval sciences. Efforts will be devoted to bring these results not only to the relevant scholarly communities but also to a wider audience as a resource in the public debates around science, knowledge and culture.
Summary
Alfonsine astronomy is arguably among the first European scientific achievements. It shaped a scene for actors like Regiomontanus or Copernicus. There is however little detailed historical analysis encompassing its development in its full breadth. ALFA addresses this issue by studying tables, instruments, mathematical and theoretical texts in a methodologically innovative way relying on approaches from the history of manuscript cultures, history of mathematics, and history of astronomy.
ALFA integrates these approaches not only to benefit from different perspectives but also to build new questions from their interactions. For instance the analysis of mathematical practices in astral sciences manuscripts induces new ways to analyse the documents and to think about astronomical questions.
Relying on these approaches the main objectives of ALFA are thus to:
- Retrace the development of the corpus of Alfonsine texts from its origin in the second half of the 13th century to the end of the 15th century by following, on the manuscript level, the milieus fostering it;
- Analyse the Alfonsine astronomers’ practices, their relations to mathematics, to the natural world, to proofs and justification, their intellectual context and audiences;
- Build a meaningful narrative showing how astronomers in different milieus with diverse practices shaped, also from Arabic materials, an original scientific scene in Europe.
ALFA will shed new light on the intellectual history of the late medieval period as a whole and produce a better understanding of its relations to related scientific periods in Europe and beyond. It will also produce methodological breakthroughs impacting the ways history of knowledge is practiced outside the field of ancient and medieval sciences. Efforts will be devoted to bring these results not only to the relevant scholarly communities but also to a wider audience as a resource in the public debates around science, knowledge and culture.
Max ERC Funding
1 871 250 €
Duration
Start date: 2017-09-01, End date: 2022-08-31
Project acronym AlgoRNN
Project Recurrent Neural Networks and Related Machines That Learn Algorithms
Researcher (PI) Juergen Schmidhuber
Host Institution (HI) UNIVERSITA DELLA SVIZZERA ITALIANA
Country Switzerland
Call Details Advanced Grant (AdG), PE6, ERC-2016-ADG
Summary Recurrent neural networks (RNNs) are general parallel-sequential computers. Some learn their programs or weights. Our supervised Long Short-Term Memory (LSTM) RNNs were the first to win pattern recognition contests, and recently enabled best known results in speech and handwriting recognition, machine translation, etc. They are now available to billions of users through the world's most valuable public companies including Google and Apple. Nevertheless, in lots of real-world tasks RNNs do not yet live up to their full potential. Although universal in theory, in practice they fail to learn important types of algorithms. This ERC project will go far beyond today's best RNNs through novel RNN-like systems that address some of the biggest open RNN problems and hottest RNN research topics: (1) How can RNNs learn to control (through internal spotlights of attention) separate large short-memory structures such as sub-networks with fast weights, to improve performance on many natural short-term memory-intensive tasks which are currently hard to learn by RNNs, such as answering detailed questions on recently observed videos? (2) How can such RNN-like systems metalearn entire learning algorithms that outperform the original learning algorithms? (3) How to achieve efficient transfer learning from one RNN-learned set of problem-solving programs to new RNN programs solving new tasks? In other words, how can one RNN-like system actively learn to exploit algorithmic information contained in the programs running on another? We will test our systems existing benchmarks, and create new, more challenging multi-task benchmarks. This will be supported by a rather cheap, GPU-based mini-brain for implementing large RNNs.
Summary
Recurrent neural networks (RNNs) are general parallel-sequential computers. Some learn their programs or weights. Our supervised Long Short-Term Memory (LSTM) RNNs were the first to win pattern recognition contests, and recently enabled best known results in speech and handwriting recognition, machine translation, etc. They are now available to billions of users through the world's most valuable public companies including Google and Apple. Nevertheless, in lots of real-world tasks RNNs do not yet live up to their full potential. Although universal in theory, in practice they fail to learn important types of algorithms. This ERC project will go far beyond today's best RNNs through novel RNN-like systems that address some of the biggest open RNN problems and hottest RNN research topics: (1) How can RNNs learn to control (through internal spotlights of attention) separate large short-memory structures such as sub-networks with fast weights, to improve performance on many natural short-term memory-intensive tasks which are currently hard to learn by RNNs, such as answering detailed questions on recently observed videos? (2) How can such RNN-like systems metalearn entire learning algorithms that outperform the original learning algorithms? (3) How to achieve efficient transfer learning from one RNN-learned set of problem-solving programs to new RNN programs solving new tasks? In other words, how can one RNN-like system actively learn to exploit algorithmic information contained in the programs running on another? We will test our systems existing benchmarks, and create new, more challenging multi-task benchmarks. This will be supported by a rather cheap, GPU-based mini-brain for implementing large RNNs.
Max ERC Funding
2 500 000 €
Duration
Start date: 2017-10-01, End date: 2022-09-30
Project acronym ALGSTRONGCRYPTO
Project Algebraic Methods for Stronger Crypto
Researcher (PI) Ronald John Fitzgerald CRAMER
Host Institution (HI) STICHTING NEDERLANDSE WETENSCHAPPELIJK ONDERZOEK INSTITUTEN
Country Netherlands
Call Details Advanced Grant (AdG), PE6, ERC-2016-ADG
Summary Our field is cryptology. Our overarching objective is to advance significantly the frontiers in
design and analysis of high-security cryptography for the future generation.
Particularly, we wish to enhance the efficiency, functionality, and, last-but-not-least, fundamental understanding of cryptographic security against very powerful adversaries.
Our approach here is to develop completely novel methods by
deepening, strengthening and broadening the
algebraic foundations of the field.
Concretely, our lens builds on
the arithmetic codex. This is a general, abstract cryptographic primitive whose basic theory we recently developed and whose asymptotic part, which relies on algebraic geometry, enjoys crucial applications in surprising foundational results on constant communication-rate two-party cryptography. A codex is a linear (error correcting) code that, when endowing its ambient vector space just with coordinate-wise multiplication, can be viewed as simulating, up to some degree, richer arithmetical structures such as finite fields (or products thereof), or generally, finite-dimensional algebras over finite fields. Besides this degree, coordinate-localities for which simulation holds and for which it does not at all are also captured.
Our method is based on novel perspectives on codices which significantly
widen their scope and strengthen their utility. Particularly, we bring
symmetries, computational- and complexity theoretic aspects, and connections with algebraic number theory, -geometry, and -combinatorics into play in novel ways. Our applications range from public-key cryptography to secure multi-party computation.
Our proposal is subdivided into 3 interconnected modules:
(1) Algebraic- and Number Theoretical Cryptanalysis
(2) Construction of Algebraic Crypto Primitives
(3) Advanced Theory of Arithmetic Codices
Summary
Our field is cryptology. Our overarching objective is to advance significantly the frontiers in
design and analysis of high-security cryptography for the future generation.
Particularly, we wish to enhance the efficiency, functionality, and, last-but-not-least, fundamental understanding of cryptographic security against very powerful adversaries.
Our approach here is to develop completely novel methods by
deepening, strengthening and broadening the
algebraic foundations of the field.
Concretely, our lens builds on
the arithmetic codex. This is a general, abstract cryptographic primitive whose basic theory we recently developed and whose asymptotic part, which relies on algebraic geometry, enjoys crucial applications in surprising foundational results on constant communication-rate two-party cryptography. A codex is a linear (error correcting) code that, when endowing its ambient vector space just with coordinate-wise multiplication, can be viewed as simulating, up to some degree, richer arithmetical structures such as finite fields (or products thereof), or generally, finite-dimensional algebras over finite fields. Besides this degree, coordinate-localities for which simulation holds and for which it does not at all are also captured.
Our method is based on novel perspectives on codices which significantly
widen their scope and strengthen their utility. Particularly, we bring
symmetries, computational- and complexity theoretic aspects, and connections with algebraic number theory, -geometry, and -combinatorics into play in novel ways. Our applications range from public-key cryptography to secure multi-party computation.
Our proposal is subdivided into 3 interconnected modules:
(1) Algebraic- and Number Theoretical Cryptanalysis
(2) Construction of Algebraic Crypto Primitives
(3) Advanced Theory of Arithmetic Codices
Max ERC Funding
2 447 439 €
Duration
Start date: 2017-10-01, End date: 2022-09-30
Project acronym ALLERGUT
Project Mucosal Tolerance and Allergic Predisposition: Does it all start in the gut?
Researcher (PI) Caspar OHNMACHT
Host Institution (HI) HELMHOLTZ ZENTRUM MUENCHEN DEUTSCHES FORSCHUNGSZENTRUM FUER GESUNDHEIT UND UMWELT GMBH
Country Germany
Call Details Starting Grant (StG), LS6, ERC-2016-STG
Summary Currently, more than 30% of all Europeans suffer from one or more allergic disorder but treatment is still mostly symptomatic due to a lack of understanding the underlying causality. Allergies are caused by type 2 immune responses triggered by recognition of harmless antigens. Both genetic and environmental factors have been proposed to favour allergic predisposition and both factors have a huge impact on the symbiotic microbiota and the intestinal immune system. Recently we and others showed that the transcription factor ROR(γt) seems to play a key role in mucosal tolerance in the gut and also regulates intestinal type 2 immune responses.
Based on these results I postulate two major events in the gut for the development of an allergy in the lifetime of an individual: First, a failure to establish mucosal tolerance or anergy constitutes a necessity for the outbreak of allergic symptoms and allergic disease. Second, a certain ‘core’ microbiome or pathway of the intestinal microbiota predispose certain individuals for the later development of allergic disorders. Therefore, I will address the following aims:
1) Influence of ROR(γt) on mucosal tolerance induction and allergic disorders
2) Elucidate the T cell receptor repertoire of intestinal Th2 and ROR(γt)+ Tregs and assess the role of alternative NFκB pathway for induction of mucosal tolerance
3) Identification of ‘core’ microbiome signatures or metabolic pathways that favour allergic predisposition
ALLERGUT will provide ground-breaking knowledge on molecular mechanisms of the failure of mucosal tolerance in the gut and will prove if the resident ROR(γt)+ T(reg) cells can function as a mechanistic starting point for molecular intervention strategies on the background of the hygiene hypothesis. The vision of ALLERGUT is to diagnose mucosal disbalance, prevent and treat allergic disorders even before outbreak and thereby promote Public Health initiative for better living.
Summary
Currently, more than 30% of all Europeans suffer from one or more allergic disorder but treatment is still mostly symptomatic due to a lack of understanding the underlying causality. Allergies are caused by type 2 immune responses triggered by recognition of harmless antigens. Both genetic and environmental factors have been proposed to favour allergic predisposition and both factors have a huge impact on the symbiotic microbiota and the intestinal immune system. Recently we and others showed that the transcription factor ROR(γt) seems to play a key role in mucosal tolerance in the gut and also regulates intestinal type 2 immune responses.
Based on these results I postulate two major events in the gut for the development of an allergy in the lifetime of an individual: First, a failure to establish mucosal tolerance or anergy constitutes a necessity for the outbreak of allergic symptoms and allergic disease. Second, a certain ‘core’ microbiome or pathway of the intestinal microbiota predispose certain individuals for the later development of allergic disorders. Therefore, I will address the following aims:
1) Influence of ROR(γt) on mucosal tolerance induction and allergic disorders
2) Elucidate the T cell receptor repertoire of intestinal Th2 and ROR(γt)+ Tregs and assess the role of alternative NFκB pathway for induction of mucosal tolerance
3) Identification of ‘core’ microbiome signatures or metabolic pathways that favour allergic predisposition
ALLERGUT will provide ground-breaking knowledge on molecular mechanisms of the failure of mucosal tolerance in the gut and will prove if the resident ROR(γt)+ T(reg) cells can function as a mechanistic starting point for molecular intervention strategies on the background of the hygiene hypothesis. The vision of ALLERGUT is to diagnose mucosal disbalance, prevent and treat allergic disorders even before outbreak and thereby promote Public Health initiative for better living.
Max ERC Funding
1 498 175 €
Duration
Start date: 2017-07-01, End date: 2022-06-30