Distributed Causal Memory: Modular Specification and Verification in Higher-Order Distributed Separation Logic

Speaker: Léon Gondelman, Radbout University, NL

Tuesday 1.6.2021, 11:00, online

Abstract: In this presentation we are going to talk about modular specification and verification of causally-consistent distributed database, a data structure that guarantees causal consistency among replicas of the database.

With causal consistency, different replicas can observe different data on the same key, yet it is guaranteed that all data are observed in a causally related order: if a node N observes an update X originating at node M, then node N must have also observed the effects of any other update Y that took place on node M before X. Causal consistency can, for instance, be used to ensure in a distributed messaging application that a reply to a message is never seen before the message itself.Read more...

Controlling a random population

Speaker: Pierre Ohlmann, IRIF

Tuesday 6.4.2021, 11:00, online

Abstract: Bertrand et al. introduced a model of parameterised systems, where each agent is represented by a finite state system, and studied the following control problem: for any number of agents, does there exist a controller able to bring all agents to a target state? They showed that the problem is decidable and EXPTIME-complete in the adversarial setting, and posed as an open problem the stochastic setting, where the agent is represented by a Markov decision process. In this paper, we show that the stochastic control problem is decidable. Our solution makes significant uses of well quasi orders, of the max-flow min-cut theorem, and of the theory of regular cost functions. We introduce an intermediate problem of independent interest called the sequential flow problem, and study the complexity of solving it.

Towards synthetic psychology: from phenomenology to cybernetics

Speaker: David Rudrauf

Tuesday 9.3.2021, 11:00, online

Abstract: The role of consciousness in biological cybernetics remains an essential yet open question for science. We introduce the Projective Consciousness Model (PCM) and show how its principles yield a unified model of appraisal and social-affective perspective taking and a method for active inference. We show how the PCM can account for known relationships between appraisal and distance as an inverse distance law, and how it can be generalised to implement Theory of Mind for strategic action planning. We use simulations of artificial agents applied to toy robots to demonstrate how different model parameters can generate a variety of emergent adaptive and maladaptive behaviours: from the ability to be resilient in the face of obstacles through imaginary projections, to the emergence of social approach and joint attention behaviours, and the ability to take advantage of false beliefs attributed to others. The approach opens new paths towards a science of consciousness, and applications, from clinical assessment to the design of artificial (virtual and robotic) agents. We discuss the interest and variety of computational challenges entailed by the approach. Read more...