Presentation of the High Impact Project “Predicting the density of incidents in the Emergency Medical System (SEM)”

Programme

9.30 am Registration.

Introduction to the event: Marco Orellana, Manager of CIDAI.

10 am Welcome.

• Mr Joan Mas i Albaigès, Director of CIDAI and Digital Scientific Director at Eurecat.
• Ms Anna Fontquerni, Director of the Emergency Medical System (SEM)
• Ms Maria Galindo, Secretary of Digital Policies in the Government of Catalonia

10.15 am High Impact Project – Predicting the density of incidents in the Emergency Medical System (SEM)          

• 10.15 am | Raimon Dalmau i Parés, Head of the Information Systems, ICT and Data Area (CIO) at SEM.
  The variability in SEM’s operations means that the resources available, both in the Coordinating Centre (sizing) and in the territory (sizing and geopositioning), have to be constantly adapted.
  We now have large historical datasets from the Coordinating Centre’s systems and on-board systems. We can harness these data to fully and accurately describe the situation of past but not future activity.
  Initiatives like this one enable us to make accurate predictions based on identifying patterns in historical data to systematise the calculation of SEM resource sizing, thus ensuring the effectiveness of the service for the public without using unnecessary resources.
  The initiative we are presenting today is one of SEM’s first steps in embracing machine learning and AI techniques for service optimisation.
  The project’s outcomes have been extremely encouraging and confirm the administration’s commitment to leveraging these innovative techniques with the medium-term goal of achieving service improvements and resource savings at the same time.         

• 10.30 am Arnau Berenguer Jiménez, Eurecat machine learning researcher Arnau will explain how the proposed challenge was conceptualised and the difficulties encountered. He will talk about the data processing that was performed, the machine learning models and techniques used and their results and conclude by comparing them. He will also show how the components were integrated into an external display platform.

• 11 am | Karla Trejo, Head of Innovation at the i2CAT Foundation Karla Trejo will present the definition of use cases and functional requirements of the SEM incident density prediction model. She will discuss the design and deployment of the cloud infrastructure designed to ensure a robust, scalable and efficient solution to support the SEM’s operational response.

• 11.15 am Roi Rodríguez, Director of Artificial Intelligence Ecosystem Development and Data Spaces at Huawei Huawei has furnished the project with a robust, scalable and efficient infrastructure. Huawei Cloud has been supplied as the main platform, and in particular the Elastic Cloud Server (ECS) service with an NVIDIA A30 GPU which delivers the computation required for incident prediction model inference and the Cloud Search Service (CSS) for data processing and interactive display of results using Kibana. Apart from these two main services, an Elastic IP (EIP) was provided to enable real-time access to Kibana from any device and a backup service generating backups of the environment. Huawei’s role in the project was to deliver technical support to define the infrastructure required for the computing demands of the project, ensuring proper use and supporting i2Cat and Eurecat in infrastructure rollout. Huawei has been part of Working Group 2 Definition of infrastructure architecture and deployment.

11.25 pm Closure

11.30 am Coffee & Networking

12 pm End of the event