GTI-IA. Grupo de Tecnología Informática

Actuales líneas de investigación del GTI-IA

Activity recognition

Development of tools for automatically recognize the activities performed by an agent (software or human) with the goal of assisting in the performance of future actions.

Many current challenges of society (both at the general and individual levels) require to develop new cyber-physical systems that exhibit a level of autonomy higher than the one of current systems. Examples can be found in many applications, such as: cities management, personal assistants, robotics (home care, household, office, or depots), logistics planning, game development, or security management. These systems must include functionalities such as perception, representation, reasoning, acting and learning, which allow them to act autonomously.

We focus on (1) intelligent digital assistants which can provide better-valued goals or recommendation of alternative higher-utility plans in tourism; (2) robotic-based assistance at home, for helping elders or generic tasks; and (3) traffic management to help smart cities decision making.

Agentes asesores en privacidad

Se pretende diseñar un modelo de agente de privacidad que integre mecanismos para adquirir información sobre el estado emocional del usuario, las normas sociales, y las preferencias y valores del usuario para poder establecer un diálogo argumentado sobre qué acciones son recomendables para disminuir los riesgos asociados al uso de las redes sociales y a la revelación de la información privada. Dicho modelo se integrará en un demostrador de red social en un entorno educativo en la infancia y la adolescencia, que involucre a usuarios reales y sus agentes asesores asociados, así como otros agentes que representen comportamientos predefinidos en redes sociales, con el objetivo de validar el efecto de la integración de dichos agentes asesores sobre la privacidad de los usuarios en determinados escenarios.

Argumentation-based Persuasion

Persuasion technologies are an interdisciplinary field of research that focuses on the design and development of interactive technologies that can create, maintain or change the thinking and human behavior using persuasion techniques. Although the theory of argumentation considers persuasion as one of the most studied types of argumentation dialogues, the application of techniques of argumentation to increase the persuasive potential of intelligent systems is an area of research with numerous challenges still to be addressed. In this context, we are working on the development of argumentation-based persuasion technologies from different perspectives:

• Their application to recommendation and decision-support systems.
• The implications of the emotional state of humans in the persuasive power of arguments.
• Their application in human and virtual agents societies, where agents act as virtual assistants for humans, monitoring the user and providing proactive, personalized and justified support for decision making.
• Their use in social networks as tools to raise awareness on privacy aspects.
• Their ability to persuade users, causing changes in their thinking and behavior through argumentation-based persuasion techniques.

Artificial Intelligent Techniques for Sustainable Manufacturing Systems

Sustainable manufacturing is defined as: “the creation of manufactured products that use processes that minimize negative environmental impacts, conserve energy and natural resources, are safe for employees, communities, and consumers and are economically sound”. Notions as lean, clean, green manufacturing (to name a few) are often used to describe strategies or philosophies that are more or less similar in the context of sustainability in manufacturing. In this research line we focus on the application of artificial intelligent techniques to achieve optimized solutions for manufacturing systems in which sustainable goals are combined with economic, time related and other objectives.

Automated Planning

Development of tools for automated planning and decision support systems for several tasks in a company.

Automated planning is an area of Artificial Intelligence which consists in the task of finding a procedural course of action for a declaratively described system to reach its goals while optimizing overall performance measures. Automated planners find the transformations to apply in each given state out of the possible transformations for that state.

We can find many applications of planning in industry: robots and autonomous systems, service composition, intelligent allocation of resources, optimization of business processes or workflows…

Data mining

Development of tools for analyzing, processing and extracting knowledge from data from different sources such as open data, social networks, etc.

Data mining is the process of sorting through large datasets to identify patterns and establish relationships to solve problems through data analysis. Data mining tools allow enterprises to predict future trends.

We focus on data related to tourism activities, such as information about attractions, visitors in a city, opinions of these visitors, etc.

Detección de emociones y comportamientos empáticos en sistemas inteligentes

Los humanos no siempre nos comportamos de manera racional. Habitualmente nuestro comportamiento está influenciado por las emociones que sentimos. Los sistemas inteligentes que reaccionan de una manera perfectamente racional sin considerar ningún tipo de emoción presentan un comportamiento extremadamente artificial que frecuentemente genera rechazo en sus relaciones con seres humanos. Nuestra línea de trabajo permite construir sistemas inteligentes que adaptan su comportamiento racional en función de las emociones que "sienten". De igual forma, al detectar emociones en las personas, son capaces de mostrar un comportamiento empático que facilita la aproximación y aceptación en sus relaciones con los humanos a medio y largo plazo.

Impacto de las emociones en la toma de decisiones

Se pretende desarrollar un modelo del estado emocional basado en el conocimiento obtenido mediante técnicas de análisis sentimental, análisis de estrés y análisis de información procedente de bioseñales, que permita determinar la situación emocional actual y futura del usuario.

Intelligent Manufacturing Systems

A intelligent manufacturing system, one of the research lines in Factories of the Future and Industry 4.0, is conceived as a distributed intelligent system in which every manufacturing component, element, and/or resource is modeled and controlled by means of software “agents” that can cooperate in order to solve complex problems. Intelligent manufacturing control systems are known to comply with the need for reactivity and adaptability of future sustainable manufacturing systems. The cooperation among agents allows an equitable resource allocation (like energy allocation) that can lead to a “social welfare”. Moreover, Cyber Pyshical Systems, Systems of Systems, Industrial IoT and Digital Twins are related sub-lines in which the group is working.

Recommender systems

Development of tools for recommending different types of items in different contexts.

A recommender system is a personalization tool that attempts to provide people with a list of information items that best fit their individual tastes. It infers the user’s preferences by analyzing the available user data, information about other users and information about the environment and selects the items that better match these preferences. In summary, a recommender system offers the possibility of personalizing the information filtering so that only information tailored to the user’s needs and preferences is shown.

We have a long experience on developing tourism recommender systems, for individuals and groups and based on different input information, such as explicit/implicit preferences, user psychology, etc.

Sistemas inteligentes con comportamientos guiados por emociones

Sistemas normativos

Se pretende desarrollar un modelo normativo donde sea posible inferir normas sociales para facilitar al usuario la adquisición de conocimiento sobre lo socialmente permitido en el contexto de la privacidad.