GATE integrates digital twins, superior machine studying, and trusted data-sharing infrastructures to allow sustainable city, well being, and industrial transformations.
GATE Institute is the primary devoted Large Knowledge and synthetic intelligence (AI) Centre of Excellence in Bulgaria and Jap Europe, co-funded by the European Fee (EC) and the Bulgarian Authorities, and established as a joint initiative of Sofia College St. Kliment Ohridski and Chalmers College of Know-how and Chalmers Industrial Applied sciences, Sweden. Having the imaginative and prescient to allow a data-driven good society, GATE is facilitating digital transformation by leveraging the alternatives that Large Knowledge and AI applied sciences create and by accelerating their impression throughout society. GATE performs analysis and creates improvements in utility domains resembling Future Cities and Digital Well being.
Because the Bulgarian hub of the Worldwide Knowledge Areas Affiliation (IDSA), GATE performs a pivotal function in constructing safe, sovereign, and trusted environments for knowledge sharing. Offering superior infrastructure – platform, knowledge, providers and testing and experimentation amenities, GATE goals to be the recognisable nationwide Knowledge Area with credible potential. By implementing federated architectures and IDS-compliant elements, GATE allows organisations to take part in collaborative innovation whereas preserving knowledge sovereignty and privateness. GATE additionally goals at extra democratic and widespread AI, constructing on novel decentralised machine studying (ML) algorithms that may address distributed collectives of native fashions by means of federated studying and data distillation, device-centric AI and ML crowd coaching.
By combining ontological engineering fashions and semantic applied sciences with the specification of logical insurance policies and constraints, GATE develops a brand new stage of intelligence able to coping with each static and dynamic knowledge in quite a lot of enterprise domains.
Aligned with Europe’s regulatory and moral management in AI, GATE contributes to the implementation of the EU AI Act, supporting companies, governments, and academia in understanding and making use of compliance-ready AI. Its experience in accountable and human-centric AI ensures that innovation goes hand in hand with authorized, moral, and societal safeguards.
The Future Cities utility area
The Future Cities utility area explores the potential of digital twin applied sciences to deal with real-world city challenges in planning, design, and repair supply. The pilot undertaking develops a scalable metropolis digital twin as a technical answer for the design, exploration, and experimentation of city environments and processes. The aims embody the manufacturing of a high-quality 3D metropolis mannequin of Sofia, semantic enrichment of the mannequin, and its utility for simulation, evaluation, and visualisation. The digital twin method – ‘design, take a look at, and construct first digitally’ – demonstrates the added worth of data-driven strategies for decision-making and promotes the mixing of digital twins into metropolis processes.
The 3D mannequin of Sofia has been created utilizing cadastral knowledge describing buildings, inexperienced areas, reduction, and the highway community. The mannequin is constantly enhanced with distant sensing knowledge resembling orthophotos and satellite tv for pc imagery to attain increased ranges of element. Constructed on broadly adopted requirements resembling CityGML and prolonged with FIWARE Good Knowledge Fashions or custom-designed fashions for particular use instances, the 3D mannequin is visualised utilizing a Caesium digital globe for interactive exploration.
As a part of ongoing innovation, a conceptual Software Area Extension (ADE) of the CityGML 3.0 Vegetation module has been proposed. This extension enhances the modelling of timber and inexperienced areas by means of the brand new ‘hook’ mechanism in CityGML 3.0 and incorporates dynamic features of vegetation progress and administration utilizing the Dynamizer module. New knowledge varieties, code lists, and enumerations present a unified and complete description of vegetation traits, supporting superior analysis in areas resembling city warmth island results, vegetation dynamics, and environmental impression evaluation.
A number of use instances illustrate the potential of town’s digital twin. In city planning and environmental evaluation, photo voltaic radiation research are performed on the metropolis scale to tell power harvesting with PV panels, passive constructing heating, cooling load administration, and microclimate analysis. These research combine 3D constructing knowledge with distant sensing for an correct illustration of shading in Sofia’s mountainous context. In transportation planning, a novel digital-twin-based micro-traffic simulation method is developed to enhance city mobility and security. Leveraging knowledge from the GATE Metropolis Residing Lab, together with LiDAR monitoring of automobiles and pedestrians, visitors simulations are carried out utilizing SUMO (Simulation of City MObility). Machine studying strategies, resembling Random Forest fashions, are utilized for the reclassification of unrecognised objects, enabling extra correct trajectory and object evaluation. This workflow supplies a scalable answer for visitors administration and concrete planning that may be tailored to completely different city settings.
The Future Cities area demonstrates how digital twins and superior modelling can create actionable insights for sustainable and data-driven city improvement, supporting extra environment friendly planning, safer transportation, and improved high quality of life.
The Digital Well being utility area
Within the Digital Health (DH) utility area, GATE advances analysis and innovation alongside a number of interconnected streams. One necessary line of labor focuses on biomarkers for the early prognosis, prognosis, and remedy of neurological ailments. The purpose is to determine related biomarkers and develop strong strategies for affected person classification and cognitive trajectory estimation. To this finish, GATE applies a spread of AI/ML and segmentation strategies, together with convolutional neural networks (CNNs) and autoencoders, mixed with native knowledge–pushed insights. Present analysis has already demonstrated important progress within the early prognosis of Alzheimer’s illness, and the methodologies at the moment are being prolonged to different neurodegenerative ‘hypersynchrony’ ailments. By incorporating mind connectivity as a basis for mechanistic understanding, AI/ML approaches are additional tailored to detect and refine disease-specific parameters throughout a number of situations.
One other stream of analysis addresses real-time affected person monitoring and alerting for hazardous occasions. Right here, GATE develops methods that use distant optical sensors to detect seizures, life-sign disruptions, or deadly syndromes resembling SUDEP, SIDS, and central apnea. In collaboration with the Centre of Experience for Epilepsy and Sleep Issues SEIN (The Netherlands), a real-time motor seizure detection and warning machine has been developed. Whereas the Dutch companion is specializing in a clinical-use model, GATE is making a home-use machine, which integrates AI/ML for personalisation along with a patented GATE subsystem for optical affected person monitoring and PTZ digicam management. Each variations at the moment are within the ultimate phases of testing and validation.
Constructing on this basis, GATE additionally explores the mixing of biomarkers with real-time detection and alerting within the context of telemedicine, the place the mixture of those capabilities contributes to simpler distant healthcare options. On the similar time, strategies initially developed for affected person monitoring are being prolonged to functions past well being. For example, real-time fall detection is being tailored for good metropolis environments, creating synergies between the Digital Well being and Future Cities domains and demonstrating how large knowledge can generate real-time societal impression.
Throughout all these analysis areas, GATE is advancing improvements in machine studying itself. Present work emphasises on-line studying, personalisation, and adaptive methods – approaches that, though nonetheless in early improvement, have the potential to considerably broaden the applicability and impression of the applied sciences. Importantly, all algorithms, subsystems, and modules developed throughout the DH area are designed in a modular manner, permitting them to be reused and built-in into various methods, notably in contexts the place robust cross-domain synergies exist.
GATE Knowledge Area Lab
Because the Bulgarian hub of the Worldwide Knowledge Areas Affiliation (IDSA), GATE drives the event of sovereign and standards-based infrastructures for reliable knowledge sharing. A key initiative is the GATE Knowledge Area Lab, an innovation atmosphere for experimentation, demonstration, and coaching. The lab capabilities as a totally operational testbed, designed in alignment with worldwide requirements and outfitted with IDSA-certified elements, together with a sandbox for implementing and validating data-sharing eventualities. Complemented by providers resembling technical and authorized coaching and enterprise mannequin improvement, the lab accelerates the deployment of real-world knowledge house use instances.
On the core of this infrastructure is the GATE Dataspace Connector, the primary in Jap Europe to attain IDSA Assurance Stage 1 Certification. The connector ensures interoperability, compliance, and safe alternate of information, whereas safeguarding full knowledge sovereignty. It allows organisations to outline and implement utilization insurance policies, monitor provenance securely, and handle entry with precision.
These capabilities underpin the event of Bulgaria’s first City Knowledge Area, supporting advanced city functions in areas resembling air high quality monitoring, sustainable mobility, power effectivity, and data-driven city planning. Leveraging federated software program architectures and federated machine studying, this infrastructure permits insights to be drawn from distributed datasets with out transferring delicate data exterior safe environments. Such an method enhances predictive modelling – optimising visitors flows, lowering air pollution publicity, or managing power demand – whereas strengthening belief amongst stakeholders managing important or private knowledge.
Via its licensed connector, demonstrator lab, and lively contribution to European initiatives, GATE is constructing a resilient, interoperable ecosystem for data-driven innovation in Large Knowledge and synthetic intelligence.
The GATE Knowledge Platform
The GATE Knowledge Platform underpins analysis and innovation by offering each infrastructure for inside initiatives and a internet hosting atmosphere for exterior companions and shoppers. Constructed on the fashionable idea of a non-public cloud, the platform integrates a variety of system sources, together with GPUs, utility servers, and superior software program instruments. These sources are accessible domestically in consumer/server mode in addition to remotely by way of the Web, guaranteeing flexibility and scalability.
The platform is totally virtualised, supporting a number of ranges of operation – digital machines, containers, utility servers, and standalone providers. It incorporates 4 distinct database methods able to storing knowledge in codecs resembling SQL, JSON, RDF, and CityGML. Direct integration with the native distributed file system of a Cloudera Hadoop cluster allows environment friendly processing of Large Knowledge in each offline and real-time modes.
Designed to accommodate each software-centric and data-centric initiatives, the platform aligns with trendy engineering practices, together with DevOps, DataOps, AIOps, and MLOps. It gives a full stack of preinstalled instruments to assist knowledge administration throughout your complete lifecycle – from knowledge acquisition and transport, by means of platform-level processing and repository storage, to superior evaluation and interpretation. Slicing-edge AI strategies resembling semantic integration, knowledge enrichment, deep machine studying, reinforcement studying, and immediate engineering are natively supported.
The platform additionally hosts knowledge providers for main worldwide initiatives. Notably, it helps DiverSea, a European initiative finding out the dynamics of marine biodiversity throughout coastal areas from the Black Sea to the North Sea, demonstrating its capability to allow large-scale, data-intensive analysis with societal and environmental impression.
The Challenge “Large Knowledge for Good Society” (GATE) is funded by Horizon 2020 Teaming programme and Operational Programme “Analysis, Innovation and Digitalisation for Clever Transformation“ 2021-2027 r., co-funded by the European Union.
Please notice, this text may even seem within the twenty fourth version of our quarterly publication.
