Introspective Agents Research Articles

Introspective inference counteracts perceptual distortion

Introspective agents can recognize the extent to which their internal perceptual experiences deviate from the actual states of the external world. This ability, also known as insight, is critically required for reality testing and is impaired in psychosis, yet little is known about its cognitive underpinnings. We develop a Bayesian modeling framework and a psychophysics paradigm to quantitatively characterize this type of insight while people experience a motion after-effect illusion. People can incorporate knowledge about the illusion into their decisions when judging the actual direction of a motion stimulus, compensating for the illusion (and often overcompensating). Furthermore, confidence, reaction-time, and pupil-dilation data all show signatures consistent with inferential adjustments in the Bayesian insight model. Our results suggest that people can question the veracity of what they see by making insightful inferences that incorporate introspective knowledge about internal distortions.

Formula omitted] and [formula omitted] almost output and regulated output synchronization of heterogeneous multi-agent systems: A scale-free protocol design

This paper studies scale-free protocol design for H∞ and H2 almost output and regulated output synchronization of heterogeneous multi-agent systems with linear, right-invertible and introspective agents in presence of external disturbances. The proposed collaborative linear dynamic protocol designs are based on localized information exchange over the same communication network which do not require any knowledge of the directed network topology, spectrum of the associated Laplacian matrix, and the number of agents. Moreover, the proposed scale-free protocols achieve H∞ and H2 almost output and regulated output synchronization with a given arbitrary degree of accuracy for any size of the network.

Modeling correlated information change: from conditional beliefs to quantum conditionals

In this paper, we propose a unified logical framework for representing and analyzing various forms of correlated information change. Our main thesis is that “logical dynamics,” in the sense of van Benthem (Exploring logical dynamics. CSLI Publications, Stanford, 1996; Logical dynamics of information and interaction. Cambridge University Press, Cambridge, 2011), and in particular dynamic epistemic notions of conditional, as developed in Baltag and Smets (Electron Notes Theor Comput Sci 165:5–21, 2006a; Stud Log 89:185–209, 2008a; Texts in logic and games. Amsterdam University Press, Amsterdam, pp 9–58, 2008b), play a central role in understanding and modeling a wide range of apparently very different information-gathering phenomena which do have one specific feature in common, namely the very act of learning new information may directly change the reality that is being learned. On the one hand, we focus on the way in which an introspective agent changes her beliefs when learning new higher-order information, i.e., information that may refer to her own beliefs. On the other hand, we analyze situations in which an observer learns about a phenomenon by performing observations that may perturb the very phenomenon under study, as in the case of quantum measurements, or observations in social sciences, psychology and medicine. Our formal techniques are based on ideas from dynamic logic and on the modeling of “dynamic conditionals.” We offer a semantics based on “test frames,” i.e., Kripke frames labeled by propositional formulae which yields a unified setting for the two types of correlated information change under study. We show how this framework can be used to analyze the ontic and epistemic–informational aspects of quantum measurements and to compare them with other types of observation, testing, belief revision, counterfactual conditionals, etc.

Information dynamics and uniform substitution

The picture of information acquisition as the elimination of possibilities has proven fruitful in many domains, serving as a foundation for formal models in philosophy, linguistics, computer science, and economics. While the picture appears simple, its formalization in dynamic epistemic logic reveals subtleties: given a valid principle of information dynamics in the language of dynamic epistemic logic, substituting complex epistemic sentences for its atomic sentences may result in an invalid principle. In this article, we explore such failures of uniform substitution. First, we give epistemic examples inspired by Moore, Fitch, and Williamson. Second, we answer affirmatively a question posed by van Benthem: can we effectively decide when every substitution instance of a given dynamic epistemic principle is valid? In technical terms, we prove the decidability of this schematic validity problem for public announcement logic (PAL and PAL-RC) over models for finitely many fully introspective agents, as well as models for infinitely many arbitrary agents. The proof of this result illuminates the reasons for the failure of uniform substitution.

Dynamic service composition enabled by introspective agent coordination

Service composition has received much interest from many research communities. The major research efforts published to date propose the use of service orchestration to model this problem. However, the designed orchestration approaches are static since they follow a predefined plan specifying the services to be composed and their data flow, and most of them are centralized around a composition engine. Further, task decomposition is made prior to service composition, whereas it should be performed according to the available competencies. In order to overcome these limitations, we propose to model a dynamic approach for service composition. The studied approach relies on the decentralized and autonomous collaboration of a set of services whose aim is to achieve a specific goal. In our work, this goal is to satisfy requirements in software services that are freely expressed by human users (i.e. not predefined through a composition plan). We propose to enable the service collaborations through a multi-agent coordination protocol. In our model, agents offer services and are endowed with introspective capabilities, that is, they can access and reason on their state and actions at runtime. Thereby, the agents are capable of decomposing a monolithic task according to their service skills, and dynamically coordinating with their acquaintances to cover the whole task achievement. This paper presents the adaptive agent-based approach we propose for dynamic service composition, describes its architecture, specifies the underlying coordination protocol, called omposer, verifies the protocol's main properties, and validates it by unfolding an implemented scenario.

Apprentissage par renforcement d'actes de communication dans un système multi-agent

In this paper, we present a reinforcement learning approach for multi-agent communication in order to learn what to communicate, when and to whom. This method is based on introspective agents, that can reason about their own actions and data so as to construct appropriate communicative acts. We propose an extention of classical reinforcement learning algorithms with multi-agent communication. We show how communicative acts and memory can solve non-markovity and asynchronism issues MAS.

The incorporation of Moorean type information by introspective agents

The main task is to discuss the issue in belief dynamics in which philosophical beliefs and rational introspective agents incorporate Moorean type new information. First, a brief survey is conducted on Moore’s Paradox, and one of its solutions is introduced with the help of Update Semantics. Then, we present a Dynamic Doxastic Logic (DDL) which revises the belief of introspective agents put forward by Lindstrom & Rabinowicz. Next, we attempt to incorporate Moorean type new information within the DEL (DDL) framework, as advised by van Benthem, Segerberg et al. Though we maintain the principle of “the primacy of new information” from the literature on traditional belief revision theory, several unsuccessful ways are also presented. We then conclude that some special kind of success (weak success) can still be found in those revision processes although absolute success does not hold. At last, the relevant problem of “learnability” is re-considered through weak success.

Cognition and Strategy: A Deliberation Experiment

A theory of deliberation must provide a plausible account both of individuals’ choices to speak or to listen and of how they reinterpret their own views in the aftermath of deliberation. We describe a game-theoretic laboratory experiment in which subjects with diverse interests speak or listen before voting over a common outcome. An important feature of our strategic setting is that introspective agents may, upon hearing an unpersuasive argument, update away from the speaker's preferred position. While subjects are responsive to strategic incentives, they also deviate from Bayesian predictions by “overspeaking” when speech is likelier to alienate than persuade. Subjects thus come closer to the deliberative democratic ideal of a free exchange of arguments than equilibrium predictions suggest. We interpret evidence from subjects’ deliberative choices and policy votes in terms of a cognitive hierarchy among subjects, defined by differing abilities to grasp the strategic implications of different kinds of inf...

My beliefs about your beliefs: a case study in theory of mind and epistemic logic

We model three examples of beliefs that agents may have about other agents’ beliefs, and provide motivation for this conceptualization from the theory of mind literature. We assume a modal logical framework for modelling degrees of belief by partially ordered preference relations. In this setting, we describe that agents believe that other agents do not distinguish among their beliefs (‘no preferences’), that agents believe that the beliefs of other agents are in part as their own (‘my preferences’), and the special case that agents believe that the beliefs of other agents are exactly as their own (‘preference refinement’). This multi-agent belief interaction is frame characterizable. We provide examples for introspective agents. We investigate which of these forms of belief interaction are preserved under three common forms of belief revision.

Maximal Introspection of Agents

This paper concerns the representation of introspective belief and knowledge in multi-agent systems. An introspective agent is an agent that has the ability to refer to itself and reason about its own beliefs. It is well-known that representing introspective beliefs is theoretically very problematic. An agent which is given strong introspective abilities is most likely to have inconsistent beliefs, since it can use introspection to express self-referential beliefs that are paradoxical in the same way as the classical paradoxes of self-reference. In multi-agent systems these paradoxical beliefs can even be expressed as beliefs about the correctness and completeness of other agents' beliefs, i.e., even without the presence of explicit introspection. In this paper we explore the maximal sets of introspective beliefs that an agent can consistently obtain and retain when situated in a dynamic environment, and when treating beliefs “syntactically” (that is, formalizing beliefs as axioms of first-order predicate logic rather than using modal formalisms). We generalize some previous results by Perlis [1985] and des Rivières & Levesque [1988].

Preface

Preface

Ddl Unlimited: Dynamic Doxastic Logic for Introspective Agents

The theories of belief change developed within the AGM-tradition are not logics in the proper sense, but rather informal axiomatic theories of belief change. Instead of characterizing the models of belief and belief change in a formalized object language, the AGM-approach uses a natural language — ordinary mathematical English — to characterize the mathematical structures that are under study. Recently, however, various authors such as Johan van Benthem and Maarten de Rijke have suggested representing doxastic change within a formal logical language: a dynamic modal logic.1 Inspired by these suggestions Krister Segerberg has developed a very general logical framework for reasoning about doxastic change: dynamic doxastic logic (DDL).2 This framework may be seen as an extension of standard Hintikka-style doxastic logic (Hintikka 1962) with dynamic operators representing various kinds of transformations of the agent's doxastic state. Basic DDL describes an agent that has opinions about the external world and an ability to change these opinions in the light of new information. Such an agent is non-introspective in the sense that he lacks opinions about his own belief states. Here we are going to discuss various possibilities for developing a dynamic doxastic logic for introspective agents: full DDL or DDL unlimited. The project of constructing such a logic is faced with difficulties due to the fact that the agent’s own doxastic state now becomes a part of the reality that he is trying to explore: when an introspective agent learns more about the world, then the reality he holds beliefs about

On the Logic of Acceptance and Rejection

The logic of acceptance and rejection (AEL2) is a nonmonotonic formalism to represent states of knowledge of an introspective agent making decisions about available information. Though having much in common, AEL2 differs from Moore's autoepistemic logic (AEL) by the fact that the agent not only can accept or reject a given fact, but he/she also has the possibility not to make any decision in case he/she does not have enough knowledge.

On Perfect Introspection with Quantifying-In1

Agents with perfect introspection may have incomplete beliefs about the world, but they possess complete knowledge about their own beliefs. This fact suggests that the beliefs of introspective agents should be completely determined by their objective beliefs, that is, those beliefs that are only about the domain in question and not about other beliefs. Introspection and logical reasoning alone should suffice to reconstruct all other beliefs from the objective ones. While this property has been shown to hold for propositional belief logics, there have so far only been negative results in the case of first-order belief logics with quantifying-in. In this paper we present a logic of belier with quantifying-in, where the beliefs of a perfectly introspective agent are indeed uniquely determined by the objective beliefs. The result is obtained by weakening the notion of belief of an existing logic that does not have this property.