Actionable AI goes beyond the traditional uses of AI focused on providing predictions or recommendations, since it is focused on making decisions, taking action and executing specific tasks autonomously thanks to agent-based or multi-agent systems (MAS).

MAS present environments where several autonomous agents interact and work together to achieve various objectives that may or may not be shared. These systems can be used in areas such as robotics, logistics, games or resource or infrastructure management.

In optimal control scenarios, the aim is to find the best decisions or actions to achieve a specific objective. This involves the use of algorithms and control and planning techniques to minimise or maximise a multi-objective function, which will aggregate various metrics (e.g. process improvement, resource optimisation, cost reduction, climate impact reduction and so on).

In this session we will see how actionable AI and multi-agent systems are key concepts to search for and apply artificial intelligence, especially in contexts where optimal decisions, autonomous systems and collaboration between different agents are required.

Presents:

• Xavier Domingo, Director of the Applied Artificial Intelligence (AAI) unit, Eurecat

Participants:

• Alejandro Espinosa, AI project engineer, Fundació i2CAT

• Bernat Joseph i Duran, Researcher and project manager, CETAQUA

• Lluís Santcliment, Innovation Promoter, Centre for Telecommunications and Information Technologies (CTTI). Government of Catalonia.

• Monika Falk, Artificial intelligence – Research Technician, Mosquito Alert Project, The Blanes Center for Advanced Studies (CEAB) – CSIC

https://i2cat.net | @i2CAT

Alejandro Espinosa is a Computer Scientist currently pursuing a Master’s degree in Data Science. Alejandro is currently leading the technical development of AI solutions through his participation in two Horizon Europe projects that aim at using Decentralised AI techniques to optimize edge-cloud continuum operations. Before taking on this challenge, Alejandro worked as a data analyst and completed his Bachelor’s degree thesis on AI applied to the field of cybersecurity.

This talk will explore the usage of artificial intelligence in the edge-cloud continuum by showcasing the results of a custom Deep Reinforcement Learning solution implemented within the framework of a Horizon Europe project (CODECO) to improve energy consumption in Kubernetes clusters. A Multi-Agent Reinforcement Learning (MARL) approach that extends the original DRL model with the aim of guaranteeing data privacy and providing more robust solutions in multi-cluster scenarios will also be presented.

www.cetaqua.com | Linkedin

Bernat Joseph i Duran has a degree in mathematics (2007) and a master in applied mathematics (2009) from the Technical University of Catalonia (UPC). In 2014 he obtained his PhD from the Institute of Robotics and Industrial Informatics (IRI, UPC-CSIC) where he studied the application of model predictive control to sewer networks. Since 2015 he is a researcher and project manager at Cetaqua where he is involved in innovation and research projects about the urban water cycle with special emphasis on the control and monitoring of water networks.

The operation of transport and distribution water networks is of capital importance for the companies operating the service. The way the elements of these neworks (pumps, valves, tanks) are operated defines the quality of the service to the final user as well as the safety of the infrastructure, the hydraulic efficiency or the operational costs. In this session we will explain how we have addressed, by means of deep reinforcement learning techniques, two of the main problems in the management of drinking water networks: pressure regulation and pump scheduling.

https://ctti.gencat.cat/ca/inici | Linkedin

Degree in Computer Science from the UAB 1987. Master’s Degree in Quality Systems Integration, PRL, Environment and Innovation and Postgraduate Degree in Business Management (UPC). ESADE Executive Course in Business Management.
I have been Director of Information Systems and Innovation at Aigües de Barcelona, Director of Quality and Innovation at Aigües de Mataró, CEO of Sigic Tecnologies, Director of Operations and Processes in Distribution companies. Director and advisor to different companies. As well as a member of management committees and technology and innovation committees.

In the current moment marked by constant and dynamic changes, the Public Administration is considering facing the challenges and opportunities that society requires, through technological innovation as a lever for change in digital transformation.I am pleased to present the Voice Assistant to fill in the vineyard weighing register, a pioneering initiative in the field of voice recognition applied to automatically fill in forms, through voice and with natural and plain language. The objective of the project is to implement a technological solution by the Department of Climate Action, Food and Rural Agenda of the Government of Catalonia, thus extending the proof of concept already carried out on a farm, to 10 farms throughout the Catalan territory.
This project opens the door for citizens to be able to communicate with the administration in the immediate future to obtain services through their voice

www.ceab.csic.es | Linkedin

Monika Falk is an IT professional with a strong background in web development and artificial intelligence. With dual master’s degrees, one in Computer Science focusing on Multi-media Technologies and Artificial Intelligence Methods and another in Environmental Engineering concentrating on Heating, Ventilation, and Atmosphere Protection, Monika boasts a diverse educational background. She transitioned her career focus from web development to AI several years ago. She is currently working on leveraging artificial intelligence techniques to enhance mosquito surveillance and control efforts with Mosquito Alert, driven by her passion for AI4GOOD. Her aim: AI-driven solutions for effective disease vector management.

Dive into the practical application of actionable AI for mosquito surveillance and public health management with Mosquito Alert (MA), a citizen science initiative for monitoring disease vectors. Explore how MA utilises AI for rapid mosquito detection and classification. Discover how the platform provides open access to real-time data on mosquito distribution and breeding sites via dynamic maps, empowering public institutions to reduce breeding sites. Uncover MA’s role as an early warning system, delivering information to national supervisors upon discovering new mosquito species. This session delves into the pragmatic implementation of AI for autonomous mosquito control and public health governance.

www.eurecat.org | X

Xavier Domingo is the director of the Applied Artificial Intelligence (AAI) unit at the Eurecat Technology Centre. He holds a Diploma of Advanced Studies (DIOSA) in artificial intelligence from the University of Lleida (UdL) and a degree in Computer Engineering from the Polytechnic University of Catalonia (UPC). He has extensive experience in the application of artificial intelligence methodologies and techniques supported by advanced data architectures. These include intelligent platforms with Industry 4.0 technologies, solutions for predictive maintenance, resource planning and optimisation, monitoring and control of machinery, or traceability applied to the manufacturing industry, as well as applications in other sectors such as resources (water and energy), aerospace, logistics, environment and climate change, or agri-food. He collaborates as a senior consultant in companies related to workflow management, logistics optimisation, fleet management and personal and transport fleet security systems, with 20 years of professional experience. He is a collaborator of ESADE, and associate professor in the Department of Computer Engineering and Digital Design (DEIDD) of the University of Lleida (UdL).