Development of a universal European AI platform has started. Slovak company Sygic is one of the industrial partners

15. 12. 2020

Europe has a consortium for the development of a universal platform for the processing of Artificial Intelligence. The international project called EVEREST gets funding from the Horizon 2020 Programme for research & innovation action. Coordinated by IBM Research GmbH and the university Politecnico Di Milano, Sygic is one of the project partners to prove the platform's application in the industry – smart mobility in cities.

The European data market is showing a significant boost in recent years. In the last five years, a compound annual growth rate is 15.7%, while the volume of data has a growth of 236% per year. Data processing has a profound impact on the industry, society, and the environment. The new European platform has to be comparable to those developed by giants like Google (Google Cloud), Microsoft (Microsoft Azure), or Amazon (AWS). In Horizon 2020, the platform should help bring solutions for sustainable energy, low-carbon economy, or smart city transport.

Science and industry joined forces

The three-year EVEREST project put together scientists and IT experts. Their common goal is to develop a holistic approach for co-designing computation and communication in a heterogeneous, distributed, scalable, and secure system for High-Performance Big Data Analytics (HPDA) applications. They can achieve it by simplifying architectures’ programmability through a “data-driven” design approach, the use of hardware-accelerated AI, and efficient monitoring of the execution.

“Artificial Intelligence is promoting a resurgence of research in computer architecture, especially with the combination of general-purpose processors and specialized hardware. The programmability of such heterogeneous architectures is an unsolved problem. With this framework, application developers and architecture designers can sit at the same table and efficiently co-optimize systems and applications,” said Christian Pilato from Politecnico Di Milano.

Prove by energetics, environmental and transportation cases

A project team will validate the approach with industrial partners, who prepare three scenarios for business applications. The market for trading with weather-dependent renewable energies could use a weather analysis-based prediction model. Therefore, the production, utilization, and commercialization of energy become more foreseeable. Air-quality monitoring of industrial sites helps forecast the environmental impacts of chemical pollutants and enables sites to delay their production or activate emission reduction treatments. Real-time traffic modeling framework for intelligent transportation in smart cities has to reduce congestions in the traffic infrastructure thanks to traffic simulator, traffic prediction models, and smart routing methods.

An essential step to smart and safe mobility

Smart mobility is one of the use cases that Sygic could bring to make city transport more effective in the future. Therefore, the Sygic project team covers the last-mentioned application for real-time traffic modeling framework. On this task, they cooperate with IT4Innovations, the National Supercomputing Centre at The Technical University of Ostrava.

"We have long struggled with a cost-efficient processing power for highly demanding algorithms as those of AI. This project enables us to predict traffic trends thanks to terabytes of our historic floating car data. When we put it all together, we can come up with relevant solutions for cities," says Radim Cmar, Solution Architect at Sygic.

In other words, Sygic needs to store, transfer, and process historical location-based data collected from navigation app directly by moving vehicles and create new traffic views. Combining it with collected data on acceleration, braking, and cornering, Sygic aims to learn accident risk profiles. These can be a basis for calculating the safest routes, interesting for the insurance segment.

The EVEREST project has recently launched its webpage, where you can follow their news.