ERC FUNDED PROJECTS

Natural language understanding (NLU) by the machine is of large scientific, economic and social value. Humans perform the NLU task in an efficient way by relying on their capability to imagine or anticipate situations. They engage commonsense and world knowledge that is often acquired through perceptual experiences to make explicit what is left implicit in language. Inspired by these characteristics CALCULUS will design, implement and evaluate innovative paradigms supporting NLU, where it will combine old but powerful ideas for language understanding from the early days of artificial intelligence with new approaches from machine learning. The project focuses on the effective learning of anticipatory, continuous, non-symbolic representations of event frames and narrative structures of events that are trained on language and visual data. The grammatical structure of language is grounded in the geometric structure of visual data while embodying aspects of commonsense and world knowledge. The reusable representations are evaluated in a selection of NLU tasks requiring efficient real-time retrieval of the representations and parsing of the targeted written texts. Finally, we will evaluate the inference potential of the anticipatory representations in situations not seen in the training data and when inferring spatial and temporal information in metric real world spaces that is not mentioned in the processed language. The machine learning methods focus on learning latent variable models relying on Bayesian probabilistic models and neural networks and focus on settings with limited training data that are manually annotated. The best models will be integrated in a demonstrator that translates the language of stories to events happening in a 3-D virtual world. The PI has interdisciplinary expertise in natural language processing, joint processing of language and visual data, information retrieval and machine learning needed for the successful realization of the project.

2 227 500 €

Start date: 2018-09-01, End date: 2023-08-31

With this research programme, inspired by open problems within noncommutative algebraic geometry (NCAG) as well as by actual developments in algebraic topology, it is our aim to lay out new foundations for NCAG. On the one hand, the categorical approach to geometry put forth in NCAG has seen a wide range of applications both in mathematics and in theoretical physics. On the other hand, algebraic topology has received a vast impetus from the development of higher topos theory by Lurie and others. The current project is aimed at cross-fertilisation between the two subjects, in particular through the development of “higher linear topos theory”. We will approach the higher structure on Hochschild type complexes from two angles. Firstly, focusing on intrinsic incarnations of spaces as large categories, we will use the tensor products developed jointly with Ramos González and Shoikhet to obtain a “large version” of the Deligne conjecture. Secondly, focusing on concrete representations, we will develop new operadic techniques in order to endow complexes like the Gerstenhaber-Schack complex for prestacks (due to Dinh Van-Lowen) and the deformation complexes for monoidal categories and pasting diagrams (due to Shrestha and Yetter) with new combinatorial structure. In another direction, we will move from Hochschild cohomology of abelian categories (in the sense of Lowen-Van den Bergh) to Mac Lane cohomology for exact categories (in the sense of Kaledin-Lowen), extending the scope of NCAG to “non-linear deformations”. One of the mysteries in algebraic deformation theory is the curvature problem: in the process of deformation we are brought to the boundaries of NCAG territory through the introduction of a curvature component which disables the standard approaches to cohomology. Eventually, it is our goal to set up a new framework for NCAG which incorporates curved objects, drawing inspiration from the realm of higher categories.

1 171 360 €

Start date: 2019-06-01, End date: 2024-05-31

Immunotherapy holds the potential to dramatically improve the curative prognosis of cancer patients. However, despite significant progress, a huge gap remains to be bridged to gain board success in the clinic. A first limiting factor in cancer immunotherapy is the low response rate in large fraction of the patients and an unmet need exists for more efficient - potentially synergistic - immunotherapies that improve upon or complement existing strategies. The second limiting factor is immune-related toxicity that can cause live-threatening situations as well as seriously impair the quality of life of patients. Therefore, there is an urgent need for safer immunotherapies that allow for a more target-specific engineering of the immune system. Strategies to engineer the immune system via a materials chemistry approach, i.e. immuno-engineering, have gathered major attention over the past decade and could complement or replace biologicals, and holds promise to contribute to resolving the current issues faced by the immunotherapy field. I hypothesize that synthetic biomaterials can play an important role in anti-cancer immunotherapy with regard to synergistic, safe, but potent, instruction of innate and adaptive anti-cancer immunity and to revert the tumor microenvironment from an immune-suppressive into an immune-susceptible state. Hereto, the overall scientific objective of this proposal is to fully embrace the potential of immuno-engineering and develop several highly synergistic biomaterials strategies to engineer the immune system to fight cancer. I will develop a series of biomaterials and address a number of fundamental questions with regard to optimal biomaterial design for immuno-engineering. Based on these findings, I will elucidate those therapeutic strategies that lead to synergistic engineering of innate and adaptive immunity in combination with remodeling the tumor microenvironment from an immune-suppressive into an immune-susceptible state.

2 000 000 €

Start date: 2019-04-01, End date: 2024-03-31

In our current society, therapeutic strategies against cancer suffer from dose-limiting toxicity, lack of specificity and high morbidity. To overcome this, the use of nanomaterials (NMs) is rising, where several NM formulations are undergoing clinical trials or are used in clinics where the NMs are used as drug delivery vehicles or as mediators in physical anticancer methods (e.g. hyperthermia), where to date, the success rate is limited due to low tumor targeting efficacy, lack of specificity and frequent re-use of classical toxicity mechanisms. To overcome these issues, this research program aims to exploit the intrinsic toxicity of certain types of metal-based, degradation-prone NMs (Fe-doped ZnO, Fe-doped CuO and Ag of different sizes and coatings) towards only cancer cells as a novel and generic anti-cancer tool with 1) improved efficacy against difficult to treat cancers such as multidrug-resistant cancer cells, 2) enhanced specificity and selectivity of the treatment by the intrinsic cancer cell-specific toxicity of NMs towards cancer cells. To overcome the issues related to selective delivery of the NMs, tumor-homing cells will be used that have been shown to efficiently home to primary tumors and their metastases. In practice, the NMs used show distinct degradation kinetics that primarily induce cancer-selective toxicity. To obtain efficient tumor targeting, suicide gene-expressing tumor-homing cells will be loaded with the NMs in their cytoplasm, hereby impeding premature NM degradation. The tumor homing efficacy of these cells will be monitored via optical imaging and once at the target site these cells will be chemically destroyed using the suicide gene strategy. This will release the NMs into the tumor site, where they can selectively destroy the cancer cells. This research program will be the first to explore the full potential of cancer-specific toxicity of NMs and the use of cytoplasmic loading of cells as biological carriers for efficient delivery.

1 947 519 €

Start date: 2018-01-01, End date: 2022-12-31