Invited Speakers

Autonomous Agents and Game Theory

Speaker: Jeff Rosenschein
The School of Engineering and Computer Science
The Hebrew University of Jerusalem, Israel
 
Wednesday, 09:00 – 10:00
 

Abstract

Game theory provides a solid theoretical foundation for the building of autonomous agents, but only when used appropriately. In this talk I will present three pieces of recent work, all of which are “game-theoretic” in flavor, but each of which takes a different approach to the use of that theory:

1. Truth-bias as a technique to improve the match between Nash Equlibria and other solution criteria;
2. Iterative Voting and the use of equilibria as a stability criterion for a dynamic process;
3. Local Dominance, and the recasting of rationality criteria.

Along with descriptions of this research, I will discuss my evolving view of game theory and its relation to the building of autonomous systems.

This is joint work with Omer Lev, Kevin Leyton-Brown, Evangelos Markakis, Reshef Meir, Svetlana Obraztsov, David Parkes, Maria Polukarov, Zinovi Rabinovich, and David R. M. Thompson.

Bio (click to show)

Acting and Thinking Together

Speaker: Michael Bratman
University of Stanford
 
Thursday, 09:00 – 10:15
 

Abstract

We have an important capacity to act together, a capacity for shared intentional activity. Such shared activities are important in our lives, both instrumentally and intrinsically. A theory of human thought and action needs to be, in part, a theory of human thinking and acting together. Shared intentional activity of the sort of interest here is not merely strategic interaction; but it also need not be embedded in a framework of promises or the like.

My initial focus is on small-scale cases of shared intentional activity that need not be embedded in a larger institution and that do not involve asymmetric authority relations. I focus on duets rather than orchestras with conductors. I argue that if we begin with the planning theory of individual
agency I sketched in Intention, Plans, and Practical Reason (1987), we can construct a model of robust forms of shared agency, a model that is broadly continuous – conceptually, metaphysically, and normatively – with resources already present in our theory of individual planning agency. Our capacity for individual planning agency and related forms of individual cross-temporal organization is also at the bottom of our capacity for robust forms of sociality.

In this lecture I develop and defend this approach, drawing on my recent book Shared Agency: A Planning Theory of Acting Together (2014). I discuss how this approach helps us understand important forms of shared deliberation in which we think together. I go on to consider kinds of group intention that may seem to fall outside this framework. And I consider some recent, important challenges to this plan-theoretic approach.

Bio (click to show)

On the current state of vision in robotics

Speaker: Michael Zillich
Technical University of Vienna
 
Thursday, 13:30 – 14:30
 

Abstract

Computer vision has come a long way since its beginnings. In this talk I want to review some of the recent successes, which seem to indicate that many aspects of vision have indeed been solved and that the way should now be paved for robotic systems that can operate freely in the real world. On closer inspection though that is not the case just yet. A set of specialised solutions in different sub areas, however impressive individually, does not constitute something like a unified theory of vision. I will point out some of the problems of current approaches related to robotic systems, like lack of abstraction and dealing with uncertainty. Finally I will suggest where I think research should and should not focus on in order to advance on a broader basis.

Bio (click to show)

Datalog revisited for reasoning in Linked Data

Speaker: Marie-Christine Rousset
Laboratoire d’Informatique de Grenoble
 
Friday, 09:00 – 10:00
 

Abstract

Linked Data provides access to huge, continuously growing amounts of open data and ontologies in RDF format that describe entities, links and properties on those entities. Equipping Linked Data with inference paves the way to make the Semantic Web a reality.
In this presentation, I will describe a unifying framework for RDF ontologies and databases that we call deductive RDF triplestores. It consists in equipping RDF triplestores with Datalog inference rules. This rule language allows to capture in a uniform manner OWL constraints that are useful in practice, such as property transtivity or symmetry, but also domain-specific rules with practical relevance for users in many domains of interest. I will illustrate the expressivity of this framework for modeling Linked Data applications and its genericity for developing inference algorithms. In particular, we will show how it allows to model the problem of data linkage in Linked Data as a reasoning problem on possibly decentralized data. I will also explain how it makes possible to efficiently extract expressive modules from Semantic Web ontologies and databases with formal guarantees, whilst effectively controlling their succinctness. Experiments conducted on real-world datasets have demonstrated the feasibility of this approach and its usefulness in practice for data integration and information extraction.

Bio (click to show)

Social Robotics: New forms of collective intentionality, new types of responsibility?

Speaker: Johanna Seibt
Aarhus University
 
Friday, 16:00 – 17:00
 

Abstract

‘Social robots’ are artificial systems that act in physical space in accordance with social norms and are designed to afford processes of human social cognition. In short, social robots are robots that people are disposed to understand as social agents. According to recent projections, during the next decade ‘social robotics’ will engender a technological revolution of possibly unprecedented disruptive potential, both at the socio-economic and the socio-cultural level—by 2035 a third of the workforce of industrialized countries may be replaced by robots.

In order to address the challenges of the impending “robot revolution” proactively philosophers need to move from “roboethics” (G. Veruggio, 2002) to the new stance of “robophilosophy” (Seibt, J., Nørskov, M., Hakli, R., 2014). Robophilosophy engages both practical and theoretical disciplines in philosophy, but also conceives of itself as a part of Integrative Social Robotics, a wide-scope interdisciplinary approach to developing culturally sustainable human-robot interaction scenarios.

In this lecture I present a contribution to robophilosophy that aims to highlight the potential significance of theoretical work in social ontology for the urgent task of developing legislation and policies for the quickly growing social robotics market. I shall raise the question whether we need new concepts of sociality and responsibility in order to understand (i) which types of plural actions we are engaged in when we interact with ‘social’ robots, and (ii) which obligations are carried by these plural actions. Using a process-based ontological framework for social interactions I will clarify under which conditions it is justified to speak of ‘social’ robots or ‘robot sociality’—even if one does not subscribe to functionalism in the philosophy of mind–and which kinds of asymmetric social interactions generate which kinds of plural responsibilities. The resulting taxonomy has some interesting implications. It can be used to argue that (a) a new variety of “pro-group I-intentions” (R. Hakli, K. Miller, and R. Tuomela 2010) will deserve our particular attention in the future, (b) that the deployment of social robots creates new types of collective remedial and ameliorative responsibilities for those who build, sell, and ‘use’ social robots, and (c) that roboticists face new types of individual ameliorative responsibilities that they can best discharge in collaboration with philosophers.

Bio (click to show)