Keynote Lectures

Socially Intelligent Robotics
Vanessa Evers, University of Twente, Netherlands, Netherlands

Efficient Reasoning with Rules and Ontologies
João Leite, Universidade Nova de Lisboa, Portugal, Portugal

Learning Tasks in Robotics: Problems and Solutions
Nuno Lau, Universidade de Aveiro, Portugal, Portugal

Ethical Embodied Decision Making
Francesca Rossi, IBM, United States, United States

Abstract
The classic image in the psychology of Human-Robot Interaction is that of a person who is focused and eager to learn how to work with or control a robot. The job of the roboticist then is primarily to avoid mistakes in accuracy of detection, manipulation, navigation, decision making, planning and so on to optimize human robot collaboration. 
Right?

In this talk I will argue that social norms embedded in people, robots and the context in which the robots are used make this approach obsolete. Specifically, I will address the following questions:
  • How do people understand robot behaviours?
  • What do we know about people and robots collaborating?
  • Can a robot understand human social behaviours?
  • How does knowledge about human social relationships necessitate a change in our thinking about how humans should be modelled?
  • How can the design of robots and their behaviour improve acceptance of robots in everyday environments such as our homes, airports, museums, schools, roads, and hospitals? 

Through examples of practical deployment of robots, I will explore the fundamentally social relationship people have with autonomous robots and offer essential rules for effective human-robot collaboration.

Abstract
Ontology languages based on Description Logics and reasoning rules based on Logic Programming are both well-known formalisms in knowledge representation and reasoning, each with its own distinct benefits and features, which are quite orthogonal to each other. Both appear in the Semantic Web stack in distinct standards – OWL and RIF – and over the last decade a considerable research effort has been put into trying to provide a framework that integrates the two, which would provide the grounds for the development of modern real-world applications that need to integrate and efficiently reason with hybrid knowledge bases written in these distinct formalisms. This has proved to be quite challenging, not only because of both their semantical mismatch, but also because of complexity issues. Additionally, these new applications require mechanisms for keeping such hybrid knowledge base up to date, by incorporating new and possibly conflicting information, which poses yet another significant challenge, mostly because of the fundamental mismatch between existing belief revision methods employed in Description Logics and Logic Programming which make them seem irreconcilable.
In this talk I will overview recent developments that provide the formal foundation for the efficient integration of rules and ontologies, present NoHR (http://nohr.di.fct.unl.pt/) - a tool that automates query answering over knowledge bases composed of both an Ontology written in OWL 2 EL or QL and a set of Reasoning Rules - and discus some recent promising results towards the development of automated update mechanisms for these hybrid knowledge bases.

Abstract
Machine Learning and Optimization techniques are now quite widely used in many scientific disciplines. Joining Robotics and Artificial Intelligence provides the framework to actually perform challenging tasks in real non-structured environments. Using Machine Learning in Robotics is a very challenging task as robots may be quite expensive and fragile, and the time and effort to collect data is, in general, quite high. Considering these premises, we have developed several techniques that through the use of simulators, and adapted learning/optimization algorithms, that use the data very efficiently, make the use of Learning in Robotics an effective option for hand-coded approaches.
This talk will present some of these techniques, namely those related to Q-Batch update rule, model-based learning, black-box optimization for several contexts and adapted interfaces. Although the focus will be on the application of these techniques to develop skills/interaction for robotic agents, these techniques can also be used in other types of agents.

Abstract
Decision making is an ubiquitous task in our life. However, most of the times humans are not very good at it, because of cognitive biases and data handling difficulties. Intelligent decision support systems are intended to help us in this respect.
To build an effective human-machine symbiotic system with the capability to make optimal decisions, we advocate for an embodied environment where humans are immersed. Moreover, we need humans to trust such systems.
To build the right level of trust, we need to be sure that they act in a morally acceptable way. Therefore, we need to be able to embed ethical principles (as well as social norms, professional codes, etc.) into these systems.
Existing preference modelling and reasoning framework can be a starting point, since they define priorities over actions, just like an ethical theory does. However, much work is still needed to understand how to mix preferences (that are at the core of decision making) and morality, both at the individual level and in a social context.