Search

The emergence of artificial intelligence (AI) and its progressively wider impact on many sectors requires an assessment of its effect on the achievement of the Sustainable Development Goals. Using a consensus-based expert elicitation process, we find that AI can enable the accomplishment of 134 targets across all the goals, but it may also inhibit 59 targets. However, current research foci overlook important aspects. The fast development of AI needs to be supported by the necessary regulatory insight and oversight for AI-based technologies to enable sustainable development. Failure to do so could result in gaps in transparency, safety, and ethical standards. © 2020, The Author(s). R.V. acknowledges funding provided by KTH Sustainability Office. I.L. acknowledges the Swedish Research Council (registration number 2017-05189) and funding through an Early Career Research Fellowship granted by the Jacobs Foundation. M.B. acknowledges Implicit SSF: Swedish Foundation for Strategic Research project RIT15-0046. V.D. acknowledges the support of the Wallenberg AI, Autonomous Systems, and Software Program (WASP) program funded by the Knut and Alice Wallenberg Foundation. S.D. acknowledges funding from the Leibniz Competition (J45/2018). S.L. acknowledges funding from the European Union’s Horizon 2020 Research and Innovation Programme under the Marie Skłodowska–Curie grant agreement number 748625. M.T. was supported by the Ethics and Governance of AI Fund. F.F.N. acknowledges funding from the Formas grant number 2018-01253.

In this paper, we introduce a new formulation for the value function of a zero-sum Partially Observable Stochastic Game (zs-POSG) in terms of a `plan-time sufficient statistic', a distribution over joint sets of information. We prove that this value function exhibits concavity and convexity with respect to appropriately chosen subspaces of the statistic space. We anticipate that this result is a key pre-cursor for developing solution methods that exploit such structure. Finally, we show that the formulation allow us to reduce a finite zs-POSG to a `centralized' model with shared observations, thereby transferring results for the latter (narrower) class of games to games with individual observations

Combinatorial auctions are mechanisms for allocating bundles of goods to agents who each have preferences over these goods. Finding an economically efficient allocation, the so-called winner determination problem, is computationally intractable in the general case, which is why it is important to identify special cases that are tractable but also sufficiently expressive for applications. We introduce a family of auction problems in which the goods on auction can be rearranged into a sequence, and each bid submitted concerns a bundle of goods corresponding to an interval on this sequence, possibly with multiple gaps of bounded length. We investigate the computational complexity of the winner determination problem for such auctions and explore the frontier between tractability and intractability in detail, identifying tractable, intractable, and fixed-parameter tractable cases.

This paper focuses on computational mechanisms that control the behavior of autonomous systems at runtime without necessarily restricting their autonomy. We build on existing approaches from runtime verification, control automata, and norm-based systems, and define norm-based controllers that enforce norms by modifying system behavior at runtime to make it norm compliant. For many applications, an autonomous system should comply with a set of norms. We extend our approach to a distributed setting, where a set of norm-based controllers jointly modify the runtime behavior of an autonomous system. The norms that a set of norm-based controllers jointly enforce are investigated and characterized in terms of the norms that are enforced by individual norm-based controllers. We show that a set of norm-based controllers is able to modify the runtime behavior of an autonomous system to make it compliant with all norms that the individual norm-based controllers aim at enforcing.

Existing approaches for the verification of normative systems consider limited representations of norms, often neglecting collective imperatives, deadlines and contrary-to-duty obligations. In order to capture the requirements of real-world scenarios, these structures are important. In this paper we propose methods for the specification and formal verification of complex normative systems that include contrary-to-duty, collective and event-driven imperatives with deadlines. We propose an operational syntax and semantics for the specification of such systems. Using Maude and its linear temporal logic model checker, we show how important properties can be verified for such systems, and provide some experimental results for both bounded and unbounded verification.

In the literature, different topic models have been introduced that target the task of viewpoint extraction. Because, generally, these studies do not present thorough validations of the models they introduce, it is not clear in advance which topic modeling technique will work best for our use case of extracting viewpoints of political parties from parliamentary proceedings. We argue that the usefulness of methods like topic modeling depend on whether they yield valid and reliable results on real world data. This means that there is a need for validation studies. In this paper, we present such a study for an existing topic model for viewpoint extraction called cross-perspective topic modeling [11]. The model is applied to Dutch parliamentary proceedings, and the resulting topics and opinions are validated using external data. The results of our validation show that the model yields valid topics (content and criterion validity), and opinions with content validity. We conclude that cross-perspective topic modeling is a promising technique for extracting political parties' positions from parliamentary proceedings. Second, by exploring a number of validation methods, we demonstrate that validating topic models is feasible, even without extensive domain knowledge.

The pervasiveness of open systems raises a range of challenges and opportunities for research and technological development in the area of autonomous agents and multi-agent systems. Open systems comprise loosely coupled entities interacting within a social space. These entities join the social space in order to achieve some goals that are unattainable by agents in isolation. However, when those entities are autonomous, they might misbehave and, furthermore, in open systems one may not know what entities will be active beforehand, when they may become active, or when they may leave the system.

This paper deals with the problem of automating the contribution of resources owned by people to do work for others, whilst providing a means for owners of resources to maintain autonomy over how, when and to whom their resources are used with the specification of resource use policies. We give representations of requests for resource usage as a set of conditional norms, and a use policy as specifying what norms should and should not be imposed on a resource, i.e. a set of metanorms. Our main contribution is a reasoner built on the Event Calculus, that takes a request and a use policy and determines if the request is accepted, whether it would would con ict with the use policy.

Socially intelligent systems exhibit, understand, and reason about social behavior, in order to support people in their daily lives. We claim that a fundamental new approach based on social concepts is needed to build these socially intelligent systems. In this paper, we explore how the concepts of social practices and social identities can be used to structure deliberations about actions. We then show the consequences for the architecture and reasoning capabilities of these systems.

Open systems are characterized by a diversity of heterogeneous and autonomous agents that act according to private goals, and with a behavior that is hard to predict. They can be regulated through organizations similar to human organizations, which regulate the agents’ behavior space and describe the expected behavior of the agents. Agents need to be able to reason about the regulations, so that they can act within the expected boundaries and work towards the objectives of the organization. This extended abstract introduces AORTA, a component that can be integrated into agents’ reasoning mechanism, allowing them to reason about (and act upon) regulations specified by an organizational model using simple reasoning rules. The added value is that the organizational model is independent of that of the agents, and that the approach is not tied to a specific organizational model.

Experimental and field research has demonstrated a pervasive tendency toward pairwise conformity among individuals connected by positive social ties, and work using formal models has shown that opinions on connected inuence networks should thus converge toward uniformity. Observing that diversity persists even in small scale groups and organizations, we investigate two empirically grounded mechanisms of social differentiation that may account for this persistence: First, actors may dislike or disrespect peers who diverge too much from their own views, and may change their opinions or behaviors to distance themselves further from those negative referents. Second, when surrounded by similar others, actors may try to maintain a sufficient sense of uniqueness by exploring new opinions or behaviors. Using computational experiments, we demonstrate that these two mechanisms lead to different patterns of polarization, radicalization, and factionalism and also investigate theconditions under which integration occurs. In closing, we discuss the implications for cultural dynamics in organizations.

In this paper, we contend that interpersonal circumplex theories and politeness strategies may be combined to inform the generation of social behaviours for virtual agents. We show how stances from the interpersonal circumplex correspond to certain politeness strategies and present the results of a small pilot study that partially supports our approach. Our goal is to implement this model in a serious game for police training.

Public spaces changed in the last couple of years: abundant use of smart phones and other digitally connecting devices draw peoples' attention away from physical neighbors to virtual peer groups. So, will masses of isolated people be the desired future? We think that current technology involving new display technology and, therein, new interaction possibilities hint at a different future vision. Social connectedness and bonding are important aspects of public interaction that are often overlooked, but can be initiated and supported by technology. This paper reports on research investigating how a publicly displayed application can improve social connectedness by acting in a socially accepted way. Blobulous is a novel interactive installation that interacts with participants through projected avatars, which react to the participants' movement and body signals. A functional prototype was implemented and evaluated.

People are able to comfort others by talking about their problems. In our research, we are exploring whether computers can provide social support in a similar manner. Recently, we proposed a design for an empathic virtual buddy that supports victims of cyberbullying. To validate our approach in providing social support and to gather feedback from potential users, we performed an experiment (N = 30) to compare interaction with the buddy to reading a text. Both the buddy and the text received high scores; scores for the buddy were consistently higher. The difference was significant for the extent to which feelings were taken into account. These results indicate that participants liked to interact with the buddy and that they recognized the emotional cues emitted by the buddy, thus validating our approach in comforting users.

An important task for empathic agents is to provide social support, that is, to help alleviate emotional distress. In this paper we specify verbal support types for a dialogue agent that provides social support, and propose an evaluation method for such an agent in a sensitive domain (cyberbullying) with a vulnerable target audience (children).

Teamwork between humans and computer agents has become increasingly prevalent. This paper presents a behavioral study of fairness and trust in a heterogeneous setting comprising both computer agents and human participants. It investigates people's choice of teammates and their commitment to their teams in a dynamic environment in which actions occur at a fast pace and decisions are made within tightly constrained time frames, under conditions of uncertainty and partial information. In this setting, participants could form teams by negotiating over the division of a reward for the successful completion of a group task. Participants could also choose to defect from their existing teams in order to join or create other teams. Results show that when people form teams, they offer significantly less reward to agents than they offer to people. The most significant factor affecting people's decisions whether to defect from their existing teams is the extent to which they had successful previous interactions with other team members. Also, there is no significant difference in people's rate of defection from agent-led teams as compared to their defection from human-led teams. These results are significant for agent designers and behavioral researchers who study human-agent interactions. © 2011 Elsevier Ltd. All rights reserved.

In dit rapport wordt ingegaan op de organisatorische aspecten van open data. De keus voor een bepaalde organisatorische invulling is uiteraard sterk afhankelijk van de interne besluitvorming, mogelijkheden en ambities van Alliander zelf. Dit rapport moet dan ook gezien worden als een schets van aspecten die een rol spelen bij de organisatorische invulling. De daadwerkelijk keus dient door Alliander zelf te worden gemaakt.