[agents] PhD position on Logics for Ethical Reasoning in Social Robots - IRIT, Toulouse University, France

[lorini]

PhD position in Logics for Ethical Reasoning in Social Robots

Institut de Recherche en Informatique de Toulouse (IRIT), Toulouse University, France


The International Center for Mathematics and Computer Science in Toulouse (https://cimi.univ-toulouse.fr/en <https://cimi.univ-toulouse.fr/en>), named CIMI, offers a 3-year support grant for students starting a PhD in October 2021. 
Recruited doctoral students will be paid € 1900 gross per month. They will have the opportunity to sign a teaching endorsement for the duration of their doctoral studies. 
Umberto Grandi (https://www.irit.fr/~Umberto.Grandi/ <https://www.irit.fr/~Umberto.Grandi/>) and Emiliano Lorini (https://www.irit.fr/~Emiliano.Lorini/ <https://www.irit.fr/~Emiliano.Lorini/>) are seeking a candidate for a PhD position at CIMI to work 
on the research project “Logics for Ethical Reasoning in Social Robots” in close cooperation with Rachid Alami (https://homepages.laas.fr/rachid/ <https://homepages.laas.fr/rachid/>) and Aurélie Clodic (https://homepages.laas.fr/aclodic/ <https://homepages.laas.fr/aclodic/>).

Description of the research project

An autonomous agent is, by definition, endowed with endogenous motivations, commonly called goals, which determine her preferences, thereby indirectly influencing her decision-making process. 
The connection between an agent’s goals and preferences is highly relevant for machine ethics, one of the central areas of AI nowadays (Allen et al. 2000; Etzioni & Etzioni 2017; Wallach & Allen 2008). 
Indeed, for an autonomous agent to be ethical and to behave responsibly, some of her goals must reflect values and norms with which she is expected to comply and which take other agents and their welfare into consideration. 
This includes both abstract values such as justice, fairness, reciprocity, equity and honesty and more concrete ones such as “greenhouse gas emissions are reduced”. A typical example of ethical autonomous agent 
is a robot whose set of values includes the respect for human integrity (Winfield et al. 2014). In order to supply her expected functionality, an ethical agent should be capable of computing her preference ordering over 
the alternatives directly from her values and then use it, together with her knowledge and belief, as input of her decision-making process.

There have been some attempts to formalize ethical reasoning with the aid of logical tools. There are approaches based on preference logic (Hansson 2001), event calculus (ASP) (Berreby et al. 2017), 
temporal-epistemic logic (Lorini 2015), BDI (belief, desire, intention) agent language (Dennis et al. 2016) and classical higher-order logic (HOL) (Benzmüller et al. 2002). 
The focus of this PhD thesis is the formalization of the relationship between ethical values and preferences as well as the influence of ethical values on decision-making. 
The methodology used in the project is a combination of epistemic logic (Fagin et al. 1995), dynamic epistemic logic (van Ditmarsch et al. 2007) and preference logic (van Benthem & Liu 2007) 
interpreted on a variety of formal semantics including relational semantics (Blackburn et al. 2001), neighbourhood semantics (Chellas 1980) and belief base semantics (Lorini 2020). 
The output of the PhD thesis will be a family of logics for ethical reasoning aimed at modelling interactive situations in which: (i) an agent’s value may concern other agents’ well-being, 
safety and integrity, and (ii) agents’ decisions are interdependent so that the possibility for an agent to achieve her values may depend on what other agents decide to do. 
The latter are the typical situations studied in game-theory. The logics developed in the context of the PhD thesis will allow us to express solution concepts from game theory and 
to elucidate the strategic aspects of ethical reasoning. Their semantics will borrow from well-studied concepts in social choice (most notably fairness criteria) and compact languages for preference, 
goals, and values representation (Loreggia et al., 2018) and their aggregation (Novaro et al., 2019, Haret et al. 2018).  Decision procedures for their satisfiability checking and model checking problems will be devised.

We will focus on social robotics as a pertinent context to investigate a potential algorithmic implementation of the framework. Indeed, human-robot joint action opens very challenging decisional problems 
for the robot to elaborate strategies which are not only pertinent, but also acceptable and legible by its human partner. Architectures, models and algorithms (Clodic 2017, Lemaignan 2017, Kruse 2013) 
have been proposed to reason about human mental state, to generate human-aware plans which allow to conduct collaborative human-robot task achievement. One objective would be to combine and 
enrich such systems with ethical reasoning.

References

C. Allen, G. Varner, and J. Zinser (2000). Prolegomena to any future artificial moral agent. Journal of Experimental and Theoretical Artificial Intelligence, 12, 3, 251-261.

J. van Benthem and F. Liu (2007). Dynamic logic of preference upgrade. Journal of Applied Non- Classical Logics, 17, 2, 157–182.

C. Benzmüller, X. Parent, and L. W. N. van der Torre (2020). Designing normative theories for ethical and legal reasoning: LogiKEy framework, methodology, and tool support. Artificial Intelligence, 287.

F. Berreby, G. Bourgne, and J.-G. Ganascia (2017). A Declarative Modular Framework for Representing and Applying Ethical Principles. In Proceedings of the 16th Conference on 
Autonomous Agents and MultiAgent Systems (AAMAS 2017), ACM, 96-104.

P. Blackburn, M. de Rijke, and Y. Venema (2001) Modal Logic. Cambridge University Press, Cambridge.

G. Buisan, G. Sarthou, R. Alami (2020). Human Aware Task Planning Using Verbal Communication Feasibility and Costs. International Conference on Social Robotics, Golden, United States. pp. 554-565.

B. Chellas (1980). Modal logic: an introduction. Cambridge University Press, Cambridge.

A. Clodic, J. Vázquez-Salceda, F. Dignum, S. Mascarenhas, V. Dignum, et al. (2018). On the Pertinence of Social Practices for Social Robotics. IOS Press. Envisioning Robots in Society – Power, Politics, and Public Space, pp. 36-74.

A. Clodic, E. Pacherie, R. Alami, R. Chatila (2017). Key Elements for Human Robot Joint Action. Sociality and Normativity for Robots Philosophical Inquiries into Human-Robot Interactions, 
Springer, pp.159-177, Studies in the Philosophy of Sociality.

L. A. Dennis, M. Fisher, M. Slavkovik, and M. Webster (2016). Formal verification of ethical choices in autonomous systems. Robotics and Autonomous Systems, 77, 1-14.

H. P. van Ditmarsch, W. van der Hoek, and B. Kooi (2007). Dynamic Epistemic Logic. Kluwer Academic Publishers.

A. Etzioni and O. Etzioni (2017). Incorporating Ethics into Artificial Intelligence. 
The Journal of Ethics, 21, 403-418. 


R. Fagin, J. Halpern, Y. Moses, and M. Vardi (1995). Reasoning about Knowledge. MIT Press, Cambridge.

S. O. Hansson (2001). The Structure of Values and Norms. Cambridge University 
Press.

A. Haret, A. Novaro, U. Grandi. Preference Aggregation with Incomplete CP-nets. Proceedings of the 16th International Conference on Principles of Knowledge Representation and Reasoning (KR), 2018. 


T. Kruse, A. Pandey, R. Alami, A. Kirsch (2013). Human-Aware Robot Navigation: A Survey. Robotics and Autonomous Systems, Elsevier, 61 (12), pp.1726-1743.

S. Lemaignan, M. Warnier, E. A. Sisbot, A. Clodic, R. Alami (2017). Artificial Cognition for Social Human-Robot Interaction: An Implementation. Artificial Intelligence, Elsevier, 247, pp. 45-69.

A. Loreggia, N. Mattei, F. Rossi, K. B. Venable: Preferences and Ethical Principles in Decision Making. AAAI Spring Symposia 2018

E. Lorini (2015). A logic for reasoning about moral agents. Logique & Analyse, 58, 
230, 177-218. 


E. Lorini (2019). Reasoning about cognitive attitudes in a qualitative setting. In Proceedings of the 16th European Conference on Logics in Artificial Intelligence (ECAI 2019), LNCS, vol. 11468, Springer, 726-743.

E. Lorini (2020). Rethinking epistemic logic with belief bases. Artificial Intelligence, 282.

A. Novaro, U. Grandi, D. Longin, and E. Lorini. Goal-Based Collective Decisions: Axiomatics and Computational Complexity. Proceedings of the 27th International Joint Conference on Artificial Intelligence (IJCAI), 2018.

W. Wallach and C. Allen (2008). Moral Machines: Teaching Robots Right from Wrong. Oxford University Press.

F. T. Winfield, C. Blum, and W. Liu (2014). Towards an Ethical Robot: Internal Models, Consequences and Ethical Action Selection. 
In Proceedings of the 15th Annual Conference on Advances in Autonomous Robotics Systems (TAROS 2014), LNCS, Vol. 8717, Springer, 85-96.

Candidate profile

The PhD is at the intersection of logic, knowledge and preference representation, game theory, social choice theory and social robotics. The ideal candidate should have a strong mathematical background and 
a master’s degree in Computer Science, Logic or Mathematics. She/he should also have previous experience in programming. 
Ideally, she/he should be familiar with propositional logic, modal logic as well as with the theory of static and sequential games.

Further information and how to apply

For further information about the application and the CIMI competition please email to Emiliano.Lorini at irit.fr <mailto:Emiliano.Lorini at irit.fr> and Umberto.Grandi at irit.fr <mailto:Umberto.Grandi at irit.fr>
For application, please email your detailed CV, a motivation letter, and transcripts of bachelor's degree and master’s degree to the previous e-mail addresses. 
Samples of published research by the candidate and reference letters will be a plus.

APPLICATION DEADLINE FOR FULL CONSIDERATION: February 21st 2021
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.umbc.edu/pipermail/agents/attachments/20210108/0f789d30/attachment-0001.html>