Project acronym COSYM
Project Computational Symmetry for Geometric Data Analysis and Design
Researcher (PI) Mark Pauly
Host Institution (HI) ECOLE POLYTECHNIQUE FEDERALE DE LAUSANNE
Call Details Starting Grant (StG), PE6, ERC-2010-StG_20091028
Summary The analysis and synthesis of complex 3D geometric data sets is of crucial importance in many scientific disciplines (e.g. bio-medicine, material science, mechanical engineering, physics) and industrial applications (e.g. drug design, entertainment, architecture). We are currently witnessing a tremendous increase in the size and complexity of geometric data, largely fueled by significant advances in 3D acquisition and digital production technology. However, existing computational tools are often not suited to handle this complexity.
The goal of this project is to explore a fundamentally different way of processing 3D geometry. We will investigate a new generalized model of geometric symmetry as a unifying concept for studying spatial organization in geometric data. This model allows exposing the inherent redundancies in digital 3D data and will enable truly scalable algorithms for analysis, processing, and design of large-scale geometric data sets. The proposed research will address a number of fundamental questions: What is the information content of 3D geometric models? How can we represent, store, and transmit geometric data most efficiently? Can we we use symmetry to repair deficiencies and reduce noise in acquired data? What is the role of symmetry in the design process and how can it be used to reduce complexity?
I will investigate these questions with an integrated approach that combines thorough theoretical studies with practical solutions for real-world applications.
The proposed research has a strong interdisciplinary component and will consider the same fundamental questions from different perspectives, closely interacting with scientists of various disciplines, as well artists, architects, and designers.
Summary
The analysis and synthesis of complex 3D geometric data sets is of crucial importance in many scientific disciplines (e.g. bio-medicine, material science, mechanical engineering, physics) and industrial applications (e.g. drug design, entertainment, architecture). We are currently witnessing a tremendous increase in the size and complexity of geometric data, largely fueled by significant advances in 3D acquisition and digital production technology. However, existing computational tools are often not suited to handle this complexity.
The goal of this project is to explore a fundamentally different way of processing 3D geometry. We will investigate a new generalized model of geometric symmetry as a unifying concept for studying spatial organization in geometric data. This model allows exposing the inherent redundancies in digital 3D data and will enable truly scalable algorithms for analysis, processing, and design of large-scale geometric data sets. The proposed research will address a number of fundamental questions: What is the information content of 3D geometric models? How can we represent, store, and transmit geometric data most efficiently? Can we we use symmetry to repair deficiencies and reduce noise in acquired data? What is the role of symmetry in the design process and how can it be used to reduce complexity?
I will investigate these questions with an integrated approach that combines thorough theoretical studies with practical solutions for real-world applications.
The proposed research has a strong interdisciplinary component and will consider the same fundamental questions from different perspectives, closely interacting with scientists of various disciplines, as well artists, architects, and designers.
Max ERC Funding
1 160 302 €
Duration
Start date: 2011-02-01, End date: 2016-01-31
Project acronym GEQIT
Project Generalized (quantum) information theory
Researcher (PI) Renato Renner
Host Institution (HI) EIDGENOESSISCHE TECHNISCHE HOCHSCHULE ZUERICH
Call Details Starting Grant (StG), PE6, ERC-2010-StG_20091028
Summary Information theory is a branch of science that studies, from a mathematical perspective, the processing, transmission, and storage of information. The classical theory has been established in 1948 by Claude Shannon and has later been extended to incorporate processes where information is represented by the state of quantum systems.
A major limitation of the present theory of information is that various of its concepts and methods require, as an assumption, that the processes to be studied are iterated many times. For example, Shannon's well-known result that the Shannon entropy equals the data compression rate assumes a source that repeatedly emits data according to the same given distribution. In addition, such results are often only valid asymptotically as the number of iterations tends to infinity.
While this limitation is normally acceptable when studying classical information-processing tasks such as channel coding (since communication channels are typically used repeatedly), it turns out to be a severe obstacle when analyzing new types of applications such as quantum cryptography. For instance, there is generally no sensible way to describe the attack strategy of an adversary against a quantum key distribution scheme as a recurrent process.
The goal of this project is to overcome this limitation and develop a theory of (classical and quantum) information which is completely general. Among the potential applications are new types of cryptographic schemes providing device-independent security. That is, their security guarantees hold independently of the details (and imperfections) of the actual implementations.
Summary
Information theory is a branch of science that studies, from a mathematical perspective, the processing, transmission, and storage of information. The classical theory has been established in 1948 by Claude Shannon and has later been extended to incorporate processes where information is represented by the state of quantum systems.
A major limitation of the present theory of information is that various of its concepts and methods require, as an assumption, that the processes to be studied are iterated many times. For example, Shannon's well-known result that the Shannon entropy equals the data compression rate assumes a source that repeatedly emits data according to the same given distribution. In addition, such results are often only valid asymptotically as the number of iterations tends to infinity.
While this limitation is normally acceptable when studying classical information-processing tasks such as channel coding (since communication channels are typically used repeatedly), it turns out to be a severe obstacle when analyzing new types of applications such as quantum cryptography. For instance, there is generally no sensible way to describe the attack strategy of an adversary against a quantum key distribution scheme as a recurrent process.
The goal of this project is to overcome this limitation and develop a theory of (classical and quantum) information which is completely general. Among the potential applications are new types of cryptographic schemes providing device-independent security. That is, their security guarantees hold independently of the details (and imperfections) of the actual implementations.
Max ERC Funding
1 288 792 €
Duration
Start date: 2010-12-01, End date: 2015-11-30
Project acronym KIDNEY CANCER
Project Molecular mechanisms underlying control of renal epithelial proliferative homeostasis
Researcher (PI) Ian James Frew
Host Institution (HI) UNIVERSITAT ZURICH
Call Details Starting Grant (StG), LS4, ERC-2010-StG_20091118
Summary This research grant has two major aspects. The first seeks to understand the molecular and cellular basis of the evolution of clear cell renal cell carcinoma(ccRCC), the most frequent form of kidney cancer. We will utilise an integrated approach based on mouse genetics, the use of primary kidney epithelial cell culture systems, genetic screening approaches using RNA interference libraries and analysis of the genetic and molecular changes that arise in human kidney tumours. The rationale behind these studies is that by better understanding the molecular causes of ccRCC it will be possible to identify new molecules or signaling pathways that could serve as appropriate therapeutic targets. The second aspect of this grant relates to the development of a flexible experimental platform that will allow the rapid and simultaneous up- and down-regulation of gene expression in the mouse kidney in a manner in which the affected cells are marked by a luminescent marker. This system will be based on the injection of modified lentiviral gene overexpression and gene knockdown vectors, allowing us to exploit recently-developed genome-wide cDNA libraries and RNA interference libraries. This experimental system should be equally applicable to other organ systems and will allow for the first time a systematic approach to the manipulation of gene expression in living mice, additionally bypassing the time limitations associated with conventional mouse genetic approaches. We aim to develop this system within the biological context of this grant and will combine it with live-animal imaging approaches to generate a series of mouse models of ccRCC. These will ultimately serve as invaluable tools for testing novel therapeutic approaches against this currently untreatable disease.
Summary
This research grant has two major aspects. The first seeks to understand the molecular and cellular basis of the evolution of clear cell renal cell carcinoma(ccRCC), the most frequent form of kidney cancer. We will utilise an integrated approach based on mouse genetics, the use of primary kidney epithelial cell culture systems, genetic screening approaches using RNA interference libraries and analysis of the genetic and molecular changes that arise in human kidney tumours. The rationale behind these studies is that by better understanding the molecular causes of ccRCC it will be possible to identify new molecules or signaling pathways that could serve as appropriate therapeutic targets. The second aspect of this grant relates to the development of a flexible experimental platform that will allow the rapid and simultaneous up- and down-regulation of gene expression in the mouse kidney in a manner in which the affected cells are marked by a luminescent marker. This system will be based on the injection of modified lentiviral gene overexpression and gene knockdown vectors, allowing us to exploit recently-developed genome-wide cDNA libraries and RNA interference libraries. This experimental system should be equally applicable to other organ systems and will allow for the first time a systematic approach to the manipulation of gene expression in living mice, additionally bypassing the time limitations associated with conventional mouse genetic approaches. We aim to develop this system within the biological context of this grant and will combine it with live-animal imaging approaches to generate a series of mouse models of ccRCC. These will ultimately serve as invaluable tools for testing novel therapeutic approaches against this currently untreatable disease.
Max ERC Funding
1 500 000 €
Duration
Start date: 2010-12-01, End date: 2015-11-30
Project acronym MININEXACT
Project Exact Mining from In-Exact Data
Researcher (PI) Michail Vlachos
Host Institution (HI) IBM RESEARCH GMBH
Call Details Starting Grant (StG), PE6, ERC-2010-StG_20091028
Summary Data exchange and data publishing is an inherent component of our interconnected world. Industrial companies outsource datasets to marketing and mining firms in order to support business intelligence; medical institutions exchange collected clinical experiments; academic institutions create repositories and share datasets for promoting research collaboration. A common denominator in any data exchange is the 'transformation' of the original data, which usually results in 'distortion' of data. While accurate and useful information can be potentially distilled from the original data, operations such as anonymization, rights protection and compression result in modified datasets that very seldom retain the mining capacity of its original source. This proposal seeks to address questions such as the following:
- How can we lossy compress datasets and still guarantee that mining operations are not distorted?
- Is it possible to right protect datasets and provide assurances that this task shall not impair our ability to distill useful knowledge?
- To what extent can we resolve data anonymization issues and yet retain the mining capacity of the original dataset?
We will examine a fundamental and hard problem in the area of knowledge discovery, which is the delicate balance between data transformation and data utility under mining operations. The problem lies at the confluence of many areas, such as machine and statistical learning, information theory, data representation and optimization. We will focus on studying data transformation methods (compression, anonymization, right protection) that guarantee the preservation of the salient dataset characteristics, such that data mining operations on original and transformed dataset are retained as well as possible. We will investigate how graph-centric approaches, clustering, classification and visualization algorithms can be ported to work under the proposed mining-preservation paradigm. Additional research challenges i
Summary
Data exchange and data publishing is an inherent component of our interconnected world. Industrial companies outsource datasets to marketing and mining firms in order to support business intelligence; medical institutions exchange collected clinical experiments; academic institutions create repositories and share datasets for promoting research collaboration. A common denominator in any data exchange is the 'transformation' of the original data, which usually results in 'distortion' of data. While accurate and useful information can be potentially distilled from the original data, operations such as anonymization, rights protection and compression result in modified datasets that very seldom retain the mining capacity of its original source. This proposal seeks to address questions such as the following:
- How can we lossy compress datasets and still guarantee that mining operations are not distorted?
- Is it possible to right protect datasets and provide assurances that this task shall not impair our ability to distill useful knowledge?
- To what extent can we resolve data anonymization issues and yet retain the mining capacity of the original dataset?
We will examine a fundamental and hard problem in the area of knowledge discovery, which is the delicate balance between data transformation and data utility under mining operations. The problem lies at the confluence of many areas, such as machine and statistical learning, information theory, data representation and optimization. We will focus on studying data transformation methods (compression, anonymization, right protection) that guarantee the preservation of the salient dataset characteristics, such that data mining operations on original and transformed dataset are retained as well as possible. We will investigate how graph-centric approaches, clustering, classification and visualization algorithms can be ported to work under the proposed mining-preservation paradigm. Additional research challenges i
Max ERC Funding
1 499 999 €
Duration
Start date: 2011-04-01, End date: 2016-03-31
Project acronym MULTIJEDI
Project Multilingual Joint Word Sense Disambiguation
Researcher (PI) Roberto Navigli
Host Institution (HI) UNIVERSITA DEGLI STUDI DI ROMA LA SAPIENZA
Call Details Starting Grant (StG), PE6, ERC-2010-StG_20091028
Summary In the information society the language barrier represents one of the main obstacles to the automatic use, integration and manipulation of knowledge, and this is manifested in the lack of intelligent systems able to perform unified semantic processing of textual resources in a multitude of different languages. To create such systems, a necessary step is to assign the appropriate meanings to the words in documents, a task referred to as Word Sense Disambiguation (WSD). But while WSD is typically performed in a monolingual setting, in order to enable multilingual processing, the semantic connections between word senses (i.e. meanings) in different languages need to be exploited. However, current state-of-the-art systems mainly rely on the existence of bilingual aligned text collections or limited-coverage multilingual resources to perform cross-lingual disambiguation, an unrealistic requirement when working with an arbitrary number of language pairs.
Here we propose a research program that will investigate radically new directions for performing multilingual WSD. The key intuition underlying our proposal is that WSD can be performed globally to exploit at the same time knowledge available in many languages. The first stage will involve the development of a methodology for automatically creating a large-scale, multilingual knowledge base. In a second stage, using this lexical resource, novel graph-based algorithms for jointly performing disambiguation across different languages will be designed and experimented. Crucially, we aim to show that these two tasks are mutually beneficial for going beyond current state-of-the-art WSD systems. The proposed project will have an impact not only on WSD research, but also on related areas such as Information Retrieval and Machine Translation.
Summary
In the information society the language barrier represents one of the main obstacles to the automatic use, integration and manipulation of knowledge, and this is manifested in the lack of intelligent systems able to perform unified semantic processing of textual resources in a multitude of different languages. To create such systems, a necessary step is to assign the appropriate meanings to the words in documents, a task referred to as Word Sense Disambiguation (WSD). But while WSD is typically performed in a monolingual setting, in order to enable multilingual processing, the semantic connections between word senses (i.e. meanings) in different languages need to be exploited. However, current state-of-the-art systems mainly rely on the existence of bilingual aligned text collections or limited-coverage multilingual resources to perform cross-lingual disambiguation, an unrealistic requirement when working with an arbitrary number of language pairs.
Here we propose a research program that will investigate radically new directions for performing multilingual WSD. The key intuition underlying our proposal is that WSD can be performed globally to exploit at the same time knowledge available in many languages. The first stage will involve the development of a methodology for automatically creating a large-scale, multilingual knowledge base. In a second stage, using this lexical resource, novel graph-based algorithms for jointly performing disambiguation across different languages will be designed and experimented. Crucially, we aim to show that these two tasks are mutually beneficial for going beyond current state-of-the-art WSD systems. The proposed project will have an impact not only on WSD research, but also on related areas such as Information Retrieval and Machine Translation.
Max ERC Funding
1 288 400 €
Duration
Start date: 2011-02-01, End date: 2016-01-31
