A Formal Analysis of the Concept of Behavioral Individuation of Mental States in the Functionalist Framework

The functionalist theory of mind proposes to analyze mental states in terms of internal states of Turing machine, and states of the machine’s tape and head. In the paper, I perform a formal analysis of this approach. I define the concepts of behavioral equivalence of Turing machines, and of behavioral individuation of internal states. I prove a theorem saying that for every Turing machine T there exists a Turing machine T’ which is behaviorally equivalent to T, and all of whose internal states of T’ can be behaviorally individuated. Finally, I discuss some applications of this theorem to computational theories of mind.

How Could We Know When a Robot was a Moral Patient?

There is growing interest in machine ethics in the question of whether and under what circumstances an artificial intelligence would deserve moral consideration. This paper explores a particular type of moral status that the author terms psychological moral patiency, focusing on the epistemological question of what sort of evidence might lead us to reasonably conclude that a given artificial system qualified as having this status. The paper surveys five possible criteria that might be applied: intuitive judgments, assessments of intelligence, the presence of desires and autonomous behavior, evidence of sentience, and behavioral equivalence. The author suggests that despite its limitations, the latter approach offers the best way forward, and defends a variant of that, termed the cognitive equivalence strategy. In short, this holds that an artificial system should be considered a psychological moral patient to the extent that it possesses cognitive mechanisms shared with other beings such as nonhuman animals whom we also consider to be psychological moral patients.

Towards a Generic Framework for Trustworthy Program Refactoring

Refactoring has to preserve the dynamics of the transformed program with respect to a particular definition of semantics and behavioral equivalence. Apparently, it is always challenging to relate executable refactoring implementations with the formal semantics of the transformed language. There are a number of approaches to specifying program transformations on various kinds of program models, but trustworthiness of refactoring is still to be improved by means of formal verification. We propose a specification formalism and a generic framework for its processing, which claims to allow semi-automatic execution and formal verification, as well as to be adaptable to multiple paradigms.

Scalable Methods for Computing State Similarity in Deterministic Markov Decision Processes

We present new algorithms for computing and approximating bisimulation metrics in Markov Decision Processes (MDPs). Bisimulation metrics are an elegant formalism that capture behavioral equivalence between states and provide strong theoretical guarantees on differences in optimal behaviour. Unfortunately, their computation is expensive and requires a tabular representation of the states, which has thus far rendered them impractical for large problems. In this paper we present a new version of the metric that is tied to a behavior policy in an MDP, along with an analysis of its theoretical properties. We then present two new algorithms for approximating bisimulation metrics in large, deterministic MDPs. The first does so via sampling and is guaranteed to converge to the true metric. The second is a differentiable loss which allows us to learn an approximation even for continuous state MDPs, which prior to this work had not been possible.