Széchenyi Plusz RRF
About our partner

The Budapest University of Technology and Economics (BME) is one of the leading engineering higher education institutions in the region. It has 23 000 students and 1500 research and teaching stuff and runs fully fledged degree programs in English. The university has excellent research records earning the title of Research University awarded in 2011, furthermore BME has long standing RDI activities in AI.

Research is pursued in AI in three levels:

  • enhancing the performance of AI by developing new mathematical models in the field of approximation theory, machine learning and stochastic processes,
  • integrating AI systems by using the tools of SW technologies, info-communication networks and data security,
  • application development of AI in a wide range of engineering and economical applications.

BME has managed to secure several governmental grants to fund its AI related RDI activities. It has been awarded twice in a row the “Thematic Excellence Program” in the field AI in “Smart Techniques”. Furthermore, its new national Scientific and Technological Park (presently under development) is also centered around intelligent data acquisition and monitoring supported by AI.

In addition, there are several large industrial RDI projects carried out in collaboration with large multinational companies, such as Nokia, Ericsson, Vodafone, Morgan Stanley, MOL, Knorr-Bremse, Bosch, Siemens and Deutsche Telekom on deploying AI in the fields of communication technologies, Industry 4.0, condition based maintenance in oil and gas industry, financial services and information technologies, energetics.

A brief synopsis of research in AI

The present research directions in AI are listed as follows:

  • We research on novel prediction algorithms by using feedforward neural networks which can outperform traditional mean square estimator and applied them into algo-trading and portfolio optimization.
  • Based on neural networks and large deviation theory, novel algorithms have been proposed for routing and statistical resource management, while neural based optimization was used for optimal cloud storage systems.
  • We use AI in industrial failure prediction, such as the prediction of the Remaining Useful Lifetime (RUL) of Wind turbines or the failures of the components of railway safety equipment using deep neural networks and other neuro-fuzzy systems.
  • We also have several AI research projects in the field of computer vision, such as CrowdMapping and the creation of a robust, model-predictive object detection model via self-supervised and reinforcement learning (RL).
  • We also apply RL to traffic optimization for junction and lane scheduling. Reinforcement methods are applied to find a (sub)optimal dynamic lane architecture in front of junctions in multi-agent environments to increase efficiency.
  • The SmartLabs (Speech Communications and Smart Interactions Laboratories), SmartCom Lab (Smart Communications) and DCLab (Data Science and Content Technologies Laboratory) are responsible for research pursued into cognitive information.
  • We deal with sequential data modeling, self-supervised learning and reinforcement learning as fundamental research. Speech and language technologies, sensor analytics, interaction research, self-driving technology, telecommunications data analysis and media content modeling as applied research.
  • We have been developing a new privacy-preserving biometric template protection scheme for facial recognition, that can be efficiently run at the edge even on relatively moderate hardware
  • We have been analyzing the de-anonymization risks of processing face embedding vectors (unstructured, kilobyte sized identifiers resulting from facial recognition) in CCTV systems. We have shown facial embeddings contain enough information in order to pinpoint the original data subject on a social network platform. Currently, we are looking into ways for mitigating such risks, while also maintaining useful aspects of facial recognition.

Projects

  • European AI On-Demand Platform (2018-2021 – 20M EUR est. project cost): AI4EU is the European Union’s landmark Artificial Intelligence project, which seeks to develop a European AI ecosystem, bringing together the knowledge, algorithms, tools and resources available and making it a compelling solution for users. Involving more than 70 partners, covering 21 countries, the €20m project kicked off in January 2019 and runs for three years. AI4EU will unify Europe’s Artificial Intelligence community. It will facilitate collective work in AI research, innovation and business in Europe. By sharing AI expertise, knowledge and tools with the Platform, AI4EU will make AI available to all.
  • Comprehensive safety solution for people with Aphasia (APH-ALARM – aal-2019-6-131-CP): The aim of the project is to create an alarm system for older people (55+) after stroke after aphasia in order to 1) regain and keep their independence, abilities and dignity while 2) feeling safe and supported. This would help them to 3) live an active and assisted life even after stroke. It is a comprehensive safety solution. The manual, gesture and the automatic alert triggering allows users to connect with safety contacts
  • in almost every medical, or safety related emergency. The gesture control works while the smartphone is in the pocket or bag and without wearable devices and manual intervention. Partners: Academic and SME partners from Hungary, Austria and Portugal.
  • ECSEL Productive Intelligence: Productive Intelligence (PI) – boosting digital transformation in manufacturing Productive Intelligence (PI) will further boost the digital transformation achieved by its predecessing lighthouse initiative Productive4.0, in 9 versatile industrial domains, by making full use of the potentials Artificial Intelligence (AI) has in store.Partners: Infineon, BMW, Skoda, Philips, Metso, Bosch, Danobat, TTTech, NXP, SKF, ST, AIT, KIT, TUWien, BME, etc. (89 partners)
  • The European Sector Skills Alliences, leader DIGITALEUROPE: The growing demand for skilled employees within the European Software Sector cannot be met by current education and training programmes. Europe needs an innovative new Software Skills Strategy that can fast-track the upskilling and reskilling of Europe’s workforce to address this ever-increasing skills gap. Technology innovation has changed the paradigm of the way software and IT infrastructure are being designed, delivered and managed (vide: automation, instant availability of services, growing software support functions), allowing for shorter and more efficient education cycles. Vocational education is considered as an applicable format, which enables better alignment with industry and employers’ real time needs and more flexible learning pathways.
  • The primary objective of this project is to design and implement a highly innovative (especially AI and deep learning related) effective and sustainable Software Skills Strategy for Europe that will ensure the skills needs of the rapidly expanding and evolving Software Sector can be met in the short, medium and long term.
  • ECSEL MANTIS Cyber Physical System based Proactive Collaborative Maintenance: The proactive service maintenance platform and its associated architecture of MANTIS draw inspiration from the Cyber Physical System approach. Sophisticated distributed sensing and decision making functions are performed at different levels in a collaborative way ranging from local nodes that pre-process raw sensor data and extract relevant information before transmitting it, thereby reducing bandwidth requirements of communication, over intermediate nodes that offer asset-specific analytics to locally optimise performance and maintenance, to cloud-based platforms that integrate information from ERP, CRM and CMMS systems and execute distributed processing and analytics algorithms for global decision making
  • Partners: Mondragon, Ikerlan, Fagor, Tekniker, Acciona, VTT, Wapice, Danfoss, Philips, S&T, Bosch, Liebherr, Still, TU/e, BME, etc. (47 partners)
  • COALA-Phonetics Psychological Status Monitoring by Computerised Analysis of Language phenomena EUROPEAN SPACE AGENCY_No.4000108003/13A/ KLM
  • URBMOBI: Within the framework of two EIT (European Institute of Innovation & Technology) Climate-KIC [EIT Climate-KIC] the URBMOBI (Urban Mobile Instruments for Environmental Monitoring, i.e. a Mobile Measurement Device for Urban Environmental Monitoring) projectintegrates a mobile measurement unit for operation on vehicles in urban areas (i.e. local buses and trams), with data post-processing, inclusion in enhanced environmental models and visualization techniques for climate related services, environmental monitoring, planning and research needs. URBMOBI is a mobile environmental sensor that provides temporally and spatially distributed environmental data,  fulfills the need for monitoring at various places without the costs for a large number of fixed measurement stations, integrates small and precise sensors in a system that can be operated on buses, trams or other vehicles, focusses on urban heat and thermal comfort, and aims at providing climate services and integration with real-time climate models. Consortium: RWTH Aachen University (Germany), Netherlands Organisation for Applied Scientific Research TNO (Netherlands), ARIA Technologies (France), Budapest University of Technology and Economics (Hungary), MEEO S.r.l - Meteorological and Environmental Earth Observation (Italy), and Aacener Straßenbahn und Energieversorgungsbetrieb (Germany).
  • SOLSUN: Within the frame of two EIT (European Institute of Innovation & Technology) Climate-KIC [EIT Climate-KIC] SOLSUN (Sustainable Outdoor Lighting & Sensory Urban Networks) project is about to demonstrate how intelligent city infrastructure can be created in a cost-effective and sustainable way by re-using existing street lighting as the communications backbone. We apply different technologies and methods to reduce energy consumption at the same time as turning streetlights into nodes on a scalable network that is also expandable for other applications. Sensors capture data on air pollution, noise pollution and traffic density; information gathered are used to address traffic congestion, another key contributor of greenhouse gas emissions in cities.
  • HU-MATHS-IN: The Hungarian Service Network for Mathematics in Industry and Innovations. HU-MATHS-IN promotes the participation of its members in projects supported within Horizon 2020. To achieve these goals, the network participates as a member of the EU-MATHS-IN, the European Service Network for Mathematics in Industry and Innovations. EU-MATHS-IN aims to leverage the impact of mathematics on innovations in key technologies by enhanced communication and information exchange between and among the involved stakeholders on a European level. It shall become a dedicated one-stop shop to coordinate and facilitate the required exchanges in the field of application-driven mathematical research and its exploitation for innovations in industry, science, and society.
  • NOKIA-Bell Labs: This cooperation has resulted in a number of joint papers and innovations. Here we mention the topic of our last few projects with NOKIA:  Analyzing telecommunication user data with statistical and predictive analytical tools; Using statistics, stochastic modeling and machine learning techniques to obtain application fingerprints for resource usage of data processing units; Predictive network state trajectory modeling; Anomaly detection for early detection of network anomalies. The success of one our most recent projects is also reported by the international webpage of Nokia Bell Labs.
  • MELLODDY seeks to accelerate drug discovery and increase efficiencies using machine learning and pharma industry data. The project leverages the world’s largest collection of small molecules with known biochemical/cellular activity to enable more accurate and efficient drug discovery. MELLODDY aims to train machine learning models across multi-partner datasets while ensuring privacy preservation of both the data and the models by developing a platform using federated learning. The MELLODDY consortium consists of 17 partners: 10 pharmaceutical companies: Amgen, Astellas, AstraZeneca, Bayer, Boehringer Ingelheim, GSK, Janssen Pharmaceutica NV, Merck KgaA, Novartis, and Institut de Recherches Servie; 2 academic universities: KU Leuven, BME; 4 subject matter experts: Owkin, Substra Foundation, Loodse, Iktos; one AI computing company: NVIDIA. MELLODDY is an IMI funded project. 
  • SECREDAS consortium – 69 partners from 16 European countries – has kicked-off the 50 MEuro ECSEL Joint Undertaking research and innovation project, to build a reference architecture for Secure and Safe Automated systems compliant with the new GDPR Regulation. The focus will be on automotive, rail and personal healthcare, all of which demand high security and safety, covering technologies such as radar, lidar, Vehicle-to-Infrastructure and in-vehicle networks.
  • CLAWAR Project focuses on generic issues of applied mobile robotics. This collective body was able to carry out technical tasks, perform state-of-the-art surveys, and disseminate the group’s work very effectively via the CLAWAR conferences and newsletters. The technical work was carried out largely by sharing experiences and expertise between the partners so that the individual level knowledge could be generalized to make it more widely relevant. In total, 20 technical tasks were carried out in this manner over four years. Our researchers were responsible for sensing and perception (mainly vision). Read more.