Artificial Intelligence Methodology

This book constitutes refereed proceedings of the COST 2102 International Training School on Cognitive Behavioural Systems held in Dresden, Germany, in February 2011. The 39 revised full papers presented were carefully reviewed and selected from various submissions. The volume presents new and original research results in the field of human-machine interaction inspired by cognitive behavioural human-human interaction features. The themes covered are on cognitive and computational social information processing, emotional and social believable Human-Computer Interaction (HCI) systems, behavioural and contextual analysis (...) of interaction, embodiment, perception, linguistics, semantics and sentiment analysis in dialogues and interactions, algorithmic and computational issues for the automatic recognition and synthesis of emotional states. (shrink)

In this chapter, I outline dynamic models of motivation and emotion. These turn out not to be autonomous subsystems, but, instead, are deeply integrated in the basic interactive dynamic character of living systems. Motivation is a crucial aspect of particular kinds of interactive systems -- systems for which representation is a sister aspect. Emotion is a special kind of partially reflective interaction process, and yields its own emergent motivational aspects. In addition, the overall model accounts for some of the crucial (...) properties of consciousness. (shrink)

This paper discusses the relation between intelligence and motivation in artificial agents, developing and briefly arguing for two theses. The first, the orthogonality thesis, holds (with some caveats) that intelligence and final goals (purposes) are orthogonal axes along which possible artificial intellects can freely vary—more or less any level of intelligence could be combined with more or less any final goal. The second, the instrumental convergence thesis, holds that as long as they possess a sufficient level of intelligence, agents having (...) any of a wide range of final goals will pursue similar intermediary goals because they have instrumental reasons to do so. In combination, the two theses help us understand the possible range of behavior of superintelligent agents, and they point to some potential dangers in building such an agent. (shrink)

High-level perception--”the process of making sense of complex data at an abstract, conceptual level--”is fundamental to human cognition. Through high-level perception, chaotic environmen- tal stimuli are organized into the mental representations that are used throughout cognitive pro- cessing. Much work in traditional artificial intelligence has ignored the process of high-level perception, by starting with hand-coded representations. In this paper, we argue that this dis- missal of perceptual processes leads to distorted models of human cognition. We examine some existing artificial-intelligence models--”notably (...) BACON, a model of scientific discovery, and the Structure-Mapping Engine, a model of analogical thought--”and argue that these are flawed pre- cisely because they downplay the role of high-level perception. Further, we argue that perceptu- al processes cannot be separated from other cognitive processes even in principle, and therefore that traditional artificial-intelligence models cannot be defended by supposing the existence of a --œrepresentation module--� that supplies representations ready-made. Finally, we describe a model of high-level perception and analogical thought in which perceptual processing is integrated with analogical mapping, leading to the flexible build-up of representations appropriate to a given context. (shrink)

Emotion theorists tend to separate “arousal” and other bodily events such as “actions” from the evaluative component of emotion known as “appraisal.” This separation, I argue, implies phenomenologically implausible accounts of emotion elicitation and personhood. As an alternative, I attempt a reconceptualization of the notion of appraisal within the so-called “enactive approach.” I argue that appraisal is constituted by arousal and action, and I show how this view relates to an embodied and affective notion of personhood.

For many years emotion theory has been characterized by a dichotomy between the head and the body. In the golden years of cognitivism, during the nineteen-sixties and seventies, emotion theory focused on the cognitive antecedents of emotion, the so-called “appraisal processes.” Bodily events were seen largely as byproducts of cognition, and as too unspecific to contribute to the variety of emotion experience. Cognition was conceptualized as an abstract, intellectual, “heady” process separate from bodily events. Although current emotion theory has moved (...) beyond this disembodied stance by conceiving of emotions as involving both cognitive processes (perception, attention, and evaluation) and bodily events (arousal, behavior, and facial expressions), the legacy of cognitivism persists in the tendency to treat cognitive and bodily events as separate constituents of emotion. Thus the cognitive aspects of emotion are supposedly distinct and separate from the bodily ones. This separation indicates that cognitivism’s disembodied conception of cognition continues to shape the way emotion theorists conceptualize emotion. (shrink)

Despite the impressive amount of financial resources recently invested in carrying out large-scale brain simulations, it is controversial what the pay-offs are of pursuing this project. One idea is that from designing, building, and running a large-scale neural simulation, scientists acquire knowledge about the computational performance of the simulating system, rather than about the neurobiological system represented in the simulation. It has been claimed that this knowledge may usher in a new era of neuromorphic, cognitive computing systems. This study elucidates (...) this claim and argues that the main challenge this era is facing is not the lack of biological realism. The challenge lies in identifying general neurocomputational principles for the design of artificial systems, which could display the robust flexibility characteristic of biological intelligence. (shrink)

Despite the impressive amount of financial resources invested in carrying out large-scale brain simulations, it is controversial what the payoffs are of pursuing this project. The present paper argues that in some cases, from designing, building, and running a large-scale neural simulation, scientists acquire useful knowledge about the computational performance of the simulating system, rather than about the neurobiological system represented in the simulation. What this means, why it is not a trivial lesson, and how it advances the literature on (...) the epistemology of computer simulation are the three preoccupations addressed by the paper. (shrink)

A number of contributions are been given in recent years to illustrate Herbert Simon’s multidisciplinary approach to the study of behaviour. In this chapter, I give a brief picture of the origins of Simon’s bounded rationality in the framework of rising AI. I show then how seminal it was Simon’s insight on the unifying role of bounded rationality in different fields, from evolutionary theory to domains traditionally difficult for AI decision-making, such as those of real-life and real-world problems.

The year 1943 is customarily considered as the birth of cybernetics. Artificial Intelligence (AI) was officially born thirteen years later, in 1956. This chapter is about two theories on human cognitive processes developed in the context of cybernetics and early AI. The first theory is that of the cyberneticist Donald MacKay, in the framework of an original version of self-organizing systems; the second is that of Allen Newell and Herbert Simon (initially with the decisive support of Clifford Shaw) and is (...) known as information-processing psychology (IPP). The latter represents the human-oriented tendency of early AI, in which the three authors were pioneers. I shall not discuss this topic with the intention, common in this type of reconstruction, of seeking contrasts between opposing paradigms (classical AI vs. cybernetics, symbolic vs. subsymbolic, symbolic vs. situated, and so forth). Randall Beer, referring to one of these battles, the ‘‘battle between computational and dynamical ideologies,’’ decried the fact that the subjects usually examined are not ‘‘experimentally testable predictions, but rather competing intuitions about the sort of theoretical framework that will ultimately be successful in explaining cognition.’’ He concluded, ‘‘the careful study of concrete examples is more likely to clarify the key issues than abstract debate over formal definitions’’. I believe this is a conclusion that should be endorsed. (shrink)

The expression ‘‘artificial intelligence’’ (AI) was introduced by John McCarthy, and the official birth of AI is unanimously considered to be the 1956 Dartmouth Conference. Thus, AI turned fifty in 2006. How did AI begin? Several differently motivated analyses have been proposed as to its origins. In this paper a brief look at those that might be considered steps towards Dartmouth is attempted, with the aim of showing how a number of research topics and controversies that marked the short history (...) of AI were touched on, or fairly well stated, during the year immediately preceding Dartmouth. The framework within which those steps were taken was the development of digital computers. Earlier computer applications in areas such as complex decision making and management, at that time dealt with by operations research techniques, were important in this story. The time was ripe for AI’s intriguingly tumultuous development, marked as it has been by hopes and defeats, successes and difficulties. (shrink)

Heuristic programming was the first area in which AI methods were tested. The favourite case-studies were fairly simple toy- problems, such as cryptarithmetic, games, such as checker or chess, and formal problems, such as logic or geometry theorem-proving. These problems are well-defined, roughly speaking, at least in comparison to real-life problems, and as such have played the role of Drosophila in early AI. In this chapter I will investigate the origins of heuristic programming and the shift to more knowledge-based and (...) real-life problem solving. (shrink)

Since the second half of the XXth century, researchers in cybernetics and AI, neural nets and connectionism, Artificial Life and new robotics have endeavoured to build different machines that could simulate functions of living organisms, such as adaptation and development, problem solving and learning. In this book these research programs are discussed, particularly as regard the epistemological issues of the behaviour modelling. One of the main novelty of this book consists of the fact that certain projects involving the building of (...) simulative machine models before the advent of cybernetics are been investigated for the first time, on the basis of little known, and sometimes completely forgotten or unpublished, texts and figures. These pre-cybernetics projects can be considered as steps toward the “discovery” of a modelling methodology that has been fully developed by those more recent research programs, and that shares some of their central goals and key methodological proposals. -/- More info in Springer link: http://www.springer.com/new+%26+forthcoming+titles+%28default%29/book/978-1-4020-0606-7 -/- This book is the English translation of La scoperta dell'artificiale, Dunod/Masson, Milan, 1998. (shrink)

My aim here is to raise a few questions concerning the problem of representation in scientific discovery computer programs. Representation, as Simon says in his paper, "imposes constraints upon the phenomena that allow the mechanisms to be inferred from the data". The issue is obviously barely outlined by Simon in his paper, while it is addressed in detail in the book by Langley, Simon, Bradshaw and Zytkow (1987), to which I shall refer in this note. Nevertheless, their analysis would appear (...) to leave open certain issues related to the nature of the "constraints" imposed by representation on problem solving strategies. (shrink)

In this chapter the early history of Computer Science, Cybernetics and Artificial Intelligence is sketched. More recent developments of AI and the philosophy of Cognitive Science are also discussed.

Since the early eighties, computationalism in the study of the mind has been “under attack” by several critics of the so-called “classic” or “symbolic” approaches in AI and cognitive science. Computationalism was generically identified with such approaches. For example, it was identified with both Allen Newell and Herbert Simon’s Physical Symbol System Hypothesis and Jerry Fodor’s theory of Language of Thought, usually without taking into account the fact ,that such approaches are very different as to their methods and aims. Zenon (...) Pylyshyn, in his influential book Computation and Cognition, claimed that both Newell and Fodor deeply influenced his ideas on cognition as computation. This probably added to the confusion, as many people still consider Pylyshyn’s book as paradigmatic of the computational approach in the study of the mind. Since then, cognitive scientists, AI researchers and also philosophers of the mind have been asked to take sides on different “paradigms” that have from time to time been proposed as opponents of (classic or symbolic) computationalism. Examples of such oppositions are: -/- computationalism vs. connectionism, computationalism vs. dynamical systems, computationalism vs. situated and embodied cognition, computationalism vs. behavioural and evolutionary robotics. -/- Our preliminary claim in section 1 is that computationalism should not be identified with what we would call the “paradigm (based on the metaphor) of the computer” (in the following, PoC). PoC is the (rather vague) statement that the mind functions “as a digital computer”. Actually, PoC is a restrictive version of computationalism, and nobody ever seriously upheld it, except in some rough versions of the computational approach and in some popular discussions about it. Usually, PoC is used as a straw man in many arguments against computationalism. In section 1 we look in some detail at PoC’s claims and argue that computationalism cannot be identified with PoC. In section 2 we point out that certain anticomputationalist arguments are based on this misleading identification. In section 3 we suggest that the view of the levels of explanation proposed by David Marr could clarify certain points of the debate on computationalism. In section 4 we touch on a controversial issue, namely the possibility of developing a notion of analog computation, similar to the notion of digital computation. A short conclusion follows in section 5. (shrink)

This paper examines the hypothesis that analogies may play a role in the generation of new ideas that are built into new explanatory theories. Methods of theory construction by analogy, by failed analogy, and by modular components from several analogies are discussed. Two different analyses of analogy are contrasted: direct mapping (Mary Hesse) and shared abstraction (Michael Genesereth). The structure of Charles Darwin's theory of natural selection shows various analogical relations. Finally, an "abstraction for selection theories" is shown to be (...) the structure of a number of theories. (shrink)

Artificial intelligence, conceived either as an attempt to provide models of human cognition or as the development of programs able to perform intelligent tasks, is primarily interested in theuses of language. It should be concerned, therefore, withpragmatics. But its concern with pragmatics should not be restricted to the narrow, traditional conception of pragmatics as the theory of communication (or of the social uses of language). In addition to that, AI should take into account also the mental uses of language (in (...) reasoning, for example) and the existential dimensions of language as a determiner of the world we (and our computers) live in. In this paper, the relevance of these three branches of pragmatics-sociopragmatics, psychopragmatics, and ontopragmatics-for AI are explored. (shrink)

We describe and try to motivate our project to build systems using both a knowledge based and a neural network approach. These two approaches are used at different stages in the solution of a problem, instead of using knowledge bases exclusively on some problems, and neural nets exclusively on others. The knowledge base (KB) is defined first in a declarative, symbolic language that is easy to use. It is then compiled into an efficient neural network (NN) representation, run, and the (...) results from run time and (eventually) from learning are decompiled to a symbolic description of the knowledge contained in the network. After inspecting this recovered knowledge, a designer would be able to modify the KB and go through the whole cycle of compiling, running, and decompiling again. The central question with which this project is concerned is, therefore, How do we go from a KB to an NN, and back again? We are investigating this question by building tools consisting of a repertoire of language/translation/network types, and trying them on problems in a variety of domains. (shrink)

This paper defines the form of prior knowledge that is required for sound inferences by analogy and single-instance generalizations, in both logical and probabilistic reasoning. In the logical case, the first order determination rule defined in Davies (1985) is shown to solve both the justification and non-redundancy problems for analogical inference. The statistical analogue of determination that is put forward is termed 'uniformity'. Based on the semantics of determination and uniformity, a third notion of "relevance" is defined, both logically and (...) probabilistically. The statistical relevance of one function in determining another is put forward as a way of defining the value of information: The statistical relevance of a function F to a function G is the absolute value of the change in one's information about the value of G afforded by specifying the value of F. This theory provides normative justifications for conclusions projected by analogy from one case to another, and for generalization from an instance to a rule. The soundness of such conclusions, in either the logical or the probabilistic case, can be identified with the extent to which the corresponding criteria (determination and uniformity) actually hold for the features being related. (shrink)

We analyze the logical form of the domain knowledge that grounds analogical inferences and generalizations from a single instance. The form of the assumptions which justify analogies is given schematically as the "determination rule", so called because it expresses the relation of one set of variables determining the values of another set. The determination relation is a logical generalization of the different types of dependency relations defined in database theory. Specifically, we define determination as a relation between schemata of first (...) order logic that have two kinds of free variables: (1) object variables and (2) what we call "polar" variables, which hold the place of truth values. Determination rules facilitate sound rule inference and valid conclusions projected by analogy from single instances, without implying what the conclusion should be prior to an inspection of the instance. They also provide a way to specify what information is sufficiently relevant to decide a question, prior to knowledge of the answer to the question. (shrink)

DeLancey shows that our understanding of emotion provides essential insight on key issues in philosophy of mind and artificial intelligence. He offers us a bold new approach to the study of the mind based on the latest scientific research and provides an accessible overview of the science of emotion.

Concern over the nature of AI is, for the tastes many AI scientists, probably overdone. In this they are like all other scientists. Working scientists worry about experiments, data, and theories, not foundational issues such as what their work is really about or whether their discipline is methodologically healthy. However, most scientists aren’t in a field that is approximately fifty years old. Even relatively new fields such as nonlinear dynamics or branches of biochemistry are in fact advances in older established (...) sciences and are therefore much more settled. Of course, by stretching things, AI can be said to have a history reaching back t o Charles Babbage, and possibly back beyond that to Leibnitz. However, all of that is best viewed as prelude. AI’s history is punctuated with the invention of the computer (and, if one wants t o stretch our history back to the 1930s, the development of the notion of computation by Turing, Church, and others). Hence, AI really began (or began in earnest) sometime in the late 1940s or early 1950s (some mark the conference a t Dartmouth in the summer of 1957 as the moment of our birth). And since those years we simply have not had time to settle into a routine science attacking reasonably well understood questions (for example, many of the questions some of us regard as supreme are regarded by others as inconsequential or mere excursions). (shrink)

In an earlier paper (Elster, 1989 a), I discussed the relation between rationality and social norms. Although I did mention the role of the emotions in sustaining social norms, I did not focus explicitly on the relation between rationality and the emotions. That relation is the main topic of the present paper, with social norms in a subsidiary part.

A Bayesian network (BN) is a graphical model of uncertainty that is especially well suited to legal arguments. It enables us to visualize and model dependencies between different hypotheses and pieces of evidence and to calculate the revised probability beliefs about all uncertain factors when any piece of new evidence is presented. Although BNs have been widely discussed and recently used in the context of legal arguments, there is no systematic, repeatable method for modeling legal arguments as BNs. Hence, where (...) BNs have been used in the legal context, they are presented as completed pieces of work, with no insights into the reasoning and working that must have gone into their construction. This means the process of building BNs for legal arguments is ad hoc, with little possibility for learning and process improvement. This article directly addresses this problem by describing a method for building useful legal arguments in a consistent and repeatable way. The method complements and extends recent work by Hepler, Dawid, and Leucari (2007) on object-oriented BNs for complex legal arguments and is based on the recognition that such arguments can be built up from a small number of basic causal structures (referred to as idioms). We present a number of examples that demonstrate the practicality and usefulness of the method. (shrink)

Concept representation is still an open problem in the field of ontology engineering and, more generally, of knowledge representation. In particular, the issue of representing “non classical” concepts, i.e. concepts that cannot be defined in terms of necessary and sufficient conditions, remains unresolved. In this paper we review empirical evidence from cognitive psychology, according to which concept representation is not a unitary phenomenon. On this basis, we sketch some proposals for concept representation, taking into account suggestions from psychological research. In (...) particular, it seems that human beings employ both prototype-based and exemplar-based representations in order to represent non classical concepts. We suggest that a similar, hybrid prototype-exemplar based approach could also prove useful in the field of knowledge representation technology. Finally, we propose conceptual spaces as a suitable framework for developing some aspects of this proposal. (shrink)

The problem of concept representation is relevant for many sub-fields of cognitive research, including psychology and philosophy, as well as artificial intelligence. In particular, in recent years it has received a great deal of attention within the field of knowledge representation, due to its relevance for both knowledge engineering as well as ontology-based technologies. However, the notion of a concept itself turns out to be highly disputed and problematic. In our opinion, one of the causes of this state of affairs (...) is that the notion of a concept is, to some extent, heterogeneous, and encompasses different cognitive phenomena. This results in a strain between conflicting requirements, such as compositionality, on the one hand and the need to represent prototypical information on the other. In some ways artificial intelligence research shows traces of this situation. In this paper, we propose an analysis of this current state of affairs. Since it is our opinion that a mature methodology with which to approach knowledge representation and knowledge engineering should also take advantage of the empirical results of cognitive psychology concerning human abilities, we outline some proposals for concept representation in formal ontologies, which take into account suggestions from psychological research. Our basic assumption is that knowledge representation systems whose design takes into account evidence from experimental psychology may therefore give better results in many applications. (shrink)

We argued [Since this argument appeared in other journals, I am reprising it here, almost verbatim.] (Fulda in J Law Info Sci 2:230–232, 1991/AI & Soc 8(4):357–359, 1994) that the paradox of the preface suggests a reason why machines cannot, will not, and should not be allowed to judge criminal cases. The argument merely shows that they cannot now and will not soon or easily be so allowed. The author, in fact, now believes that when—and only when—they are ready they (...) actually should be so allowed, in the interests of justice. Both the original argument applied and this detailed reconsideration applies exclusively to trial courts, and both specifically exclude(d) sentencing. The argument highlights some key relevant differences between minds and machines and attempts, also, to explain why automation is of far greater import for the first-level justice system (trial courts) than for higher courts. A final section discusses why sentencing was, is, and should be excluded. (shrink)

This has been made available gratis by the publisher. -/- This piece gives the raison d'etre for the development of the converters mentioned in the title. Three reasons are given, one linguistic, one philosophical, and one practical. It is suggested that at least /two/ independent converters are needed. -/- This piece ties together the extended paper "Abstracts from Logical Form I/II," and the short piece providing the comprehensive theory alluded to in the abstract of that extended paper in "Pragmatics, Montague, (...) and 'Abstracts from Logical Form'" by motivating the entire project from beginning to end. (shrink)

Humans are often extraordinary at performing practical reasoning. There are cases where the human computer, slow as it is, is faster than any artificial intelligence system. Are we faster because of the way we perceive knowledge as opposed to the way we represent it? -/- The authors address this question by presenting neural network models that integrate the two most fundamental phenomena of cognition: our ability to learn from experience, and our ability to reason from what has been learned. This (...) book is the first to offer a self-contained presentation of neural network models for a number of computer science logics, including modal, temporal, and epistemic logics. By using a graphical presentation, it explains neural networks through a sound neural-symbolic integration methodology, and it focuses on the benefits of integrating effective robust learning with expressive reasoning capabilities. -/- The book will be invaluable reading for academic researchers, graduate students, and senior undergraduates in computer science, artificial intelligence, machine learning, cognitive science and engineering. It will also be of interest to computational logicians, and professional specialists on applications of cognitive, hybrid and artificial intelligence systems. (shrink)

In this chapter, we explore the development and importance of the connection between argumentation and artificial intelligence. Specifically, we show that the influence of argumentation on AI has occurred within a framework that is consistent with the basic approach of Pragma-Dialectics. While the pragma-dialectical approach is typically conceived of as applying primarily to argumentation occurring between human agents, we show that the basic features of this approach can consistently be applied in a virtual context, whereby the goal-directed activities of, and (...) exchanges of information between, artificial agents are regulated by procedural rules. (shrink)

The declared goal of this paper is to fill this gap: “... cognitive systems research needs questions or challenges that define progress. The challenges are not (yet more) predictions of the future, but a guideline to what are the aims and what would constitute progress.” – the quotation being from the project description of EUCogII, the project for the European Network for Cognitive Systems within which this formulation of the ‘challenges’ was originally developed (http://www.eucognition.org). So, we stick out our neck (...) and formulate the challenges for artificial cognitive systems. These challenges are articulated in terms of a definition of what a cognitive system is: a system that learns from experience and uses its acquired knowledge (both declarative and practical) in a flexible manner to achieve its own goals. (shrink)

The Pleadings Game is a major contribution to artificial intelligence and legal theory. The book draws on jurisprudence and moral philosophy to develop a formal model of argumentation called the pleadings game. From a technical perspective, the work can be viewed as an extension of recent argumentation-based approaches to non-monotonic logic: (1) the game is dialogical rather than mono-logical; (2) the validity and priority of defeasible rules is subject to debate; and (3) resource limitations are acknowledged by rules for fairly (...) dividing the burdens of representation and proof among the players. Gordon's work evaluates important jurisprudential theories of argumentation and reasoning in the context of the U.S. commercial law on secured transactions. Audience: It is not necessary to have a formal background in law to appreciate The Pleadings Game. It will be of equal interest to both the artificial intelligence community and legal theorists. (shrink)

Paul E Griffiths Biohumanities Project University of Queensland St Lucia 4072 Australia paul.griffiths@uq.edu.au.

We present an approach to subjective computing for the design of future robots that exhibit more adaptive and flexible behavior in terms of subjective intelligence. Instead of encapsulating subjectivity into higher order states, we show by means of a relational approach how subjective intelligence can be implemented in terms of the reciprocity of autonomous self-referentiality and direct world-coupling. Subjectivity concerns the relational arrangement of an agent’s cognitive space. This theoretical concept is narrowed down to the problem of coaching a reinforcement (...) learning agent by means of binary feedback. Algorithms are presented that implement subjective computing. The relational characteristic of subjectivity is further confirmed by a questionnaire on human perception of robot’s behavior. The results imply that subjective intelligence cannot be externally observed. In sum, we conclude that subjective intelligence in relational terms is fully tractable and therefore implementable in artificial agents. (shrink)

In this paper, we model a relational notion of subjectivity by means of two experiments in subjective computing. The goal is to determine to what extent a cognitive and social robot can be regarded to act subjectively. The system was implemented as a reinforcement learning agent with a coaching function. To analyze the robotic agent we used the method of levels of abstraction in order to analyze the agent at four levels of abstraction. At one level the agent is described (...) in mentalistic or subjective language respectively. By mapping this mentalistic to an algorithmic, functional, and relational level, we can show to what extent the agent behaves subjectively as we make use of a relational concept of subjectivity that draws upon the relations that hold between the agent and its environment. According to a relational notion of subjectivity, an agent is supposed to be subjective if it exhibits autonomous relations to itself and others, i.e. the agent is not fully determined by a given input but is able to operate on its input and decide what to do with it. This theoretical notion is confirmed by the technical implementation of self-referentiality and social interaction in that the agent shows improved behavior compared to agents without the ability of subjective computing. On the one hand, a relational concept of subjectivity is confirmed, whereas on the other hand, the technical framework of subjective computing is being theoretically founded. (shrink)