Content uploaded by Daniel Williams
Author content
All content in this area was uploaded by Daniel Williams on Nov 27, 2018
Content may be subject to copyright.
A preview of the PDF is not available
The Mind as a Predictive Modelling Engine: Generative Models, Structural Similarity, and Mental Representation
Abstract and Figures
I outline and defend a theory of mental representation based on three ideas that I extract from the work of the mid-twentieth century philosopher, psychologist, and cybernetician Kenneth Craik: first, an account of mental representation in terms of idealised models that capitalize on structural similarity to their targets; second, an appreciation of prediction as the core function of such models; and third, a regulatory understanding of brain function. I clarify and elaborate on each of these ideas, relate them to contemporary advances in neuroscience and machine learning, and favourably contrast a generative model-based theory of mental representation with other prominent accounts of the nature, importance, and functions of mental representations in cognitive science and philosophy.
Figures - uploaded by Daniel Williams
Author content
All figure content in this area was uploaded by Daniel Williams
Content may be subject to copyright.
... That evolution reflects a process of shifting responsibility from individuals to the whole organization. A mental model means the intelligence that imitates relation structures of external processes [10][11][12][13][14][15][16]. A shared mental model of a team means a large or complete overlap of the team members' individual mental models [17][18][19][20][21][22][23]. ...
... In organisms, these representations and predictions are based on neural structures. These neural structures only imitate the (causal) relation structure of external process, instead of material changes of external processes [10][11][12][13][14][15][16]. ...
For a video presentation, see https://www.youtube.com/watch?v=SR98-3DWRsM. In this paper, it is shown how second-order adaptive agent-based network models can be used to model a medical team supported by a virtual AI Coach. It is illustrated for the case of a newborn baby in danger. The design of these computational agent models is based on an adaptive self-modeling network modeling approach. It also addresses how the AI Coach can play a central role in organizational learning. The agent models enable representations and processing of all actors' internal mental models and internal simulation of these mental models, adaptive changes of these mental models (learning and forgetting), and the interaction between the actors and the world.
... This contextualization of underdetermined tasks is one of the most fundamental and challenging cognitive tasks that the human brain is capable of. The human brain harnesses powerful prospection capabilities (Williams, 2018;Szpunar et al., 2014;Jeannerod, 2001) to ensure that this contextualization typically succeeds on the first attempt, even for novel objects, tools, and context conditions and complex tasks. A number of researchers have stressed the essential role of prospection for effective agency. ...
Provides a comparative empirical analysis of human-robot interaction in everyday life
Evaluates the social and ethical issues related to robots in human contexts. Brings together an interdisciplinary group of scholars on the key issue of how robots will shape future life
This book is open access via https://link.springer.com/book/10.1007/978-3-031-11447-2#book-header
... This contextualization of underdetermined tasks is one of the most fundamental and challenging cognitive tasks that the human brain is capable of. The human brain harnesses powerful prospection capabilities (Williams, 2018;Szpunar et al., 2014;Jeannerod, 2001) to ensure that this contextualization typically succeeds on the first attempt, even for novel objects, tools, and context conditions and complex tasks. A number of researchers have stressed the essential role of prospection for effective agency. ...
This chapter addresses the legal implications of robotic assistance. Artificial intelligence, which shall make decisions autonomously or act autonomously in interaction with humans, is associated with a substantial potential for conflict that will also and especially become evident from a legal point of view. The more AI diffuses into people’s spheres of life, the more conflicts which are associated with it will become a major theme for both the legislator and the judiciary. Questions which they have to answer include who is liable in the case of an accident and how personal data recorded via robots might be used against the owner and for third parties, including government agencies. If robots carry out actions seemingly based on their own decisions the question arises whether robots are legal persons and acquire “personality” rights as a result. Building on the results from the Delphi expertise on the social conflict scenario, the chapter examines from a legal perspective the challenges that the diffusion of AI and robots brings with it in people’s spheres of life.
... This contextualization of underdetermined tasks is one of the most fundamental and challenging cognitive tasks that the human brain is capable of. The human brain harnesses powerful prospection capabilities (Williams, 2018;Szpunar et al., 2014;Jeannerod, 2001) to ensure that this contextualization typically succeeds on the first attempt, even for novel objects, tools, and context conditions and complex tasks. A number of researchers have stressed the essential role of prospection for effective agency. ...
In addition to areas of application in people’s everyday lives and the area of education and services, robots are primarily envisioned in non-immediate living environments by the society—i.e., in inaccessible or even hostile environments to humans. The results of this population survey clearly demonstrate that such application options come across with a high level of acceptance and application potential among the population. Nevertheless, it is expected that the underlying AI in such systems works reliably and that safety for humans is guaranteed.
In this chapter, the results of the study are compared with state-of-the-art systems from classical application environments for robots, like the deep-sea and space. Here, systems have to interact with their environment to a large extent on their own over longer periods of time. Although typically the designs are such that humans are able to intervene in specific situations and so external decisions are possible, the requirements for autonomy are also extremely high. From this perspective one can easily derive what kind of requirements are also necessary, and what challenges are still in front of us, when robots should be acting largely autonomous in our everyday life.
... This contextualization of underdetermined tasks is one of the most fundamental and challenging cognitive tasks that the human brain is capable of. The human brain harnesses powerful prospection capabilities (Williams, 2018;Szpunar et al., 2014;Jeannerod, 2001) to ensure that this contextualization typically succeeds on the first attempt, even for novel objects, tools, and context conditions and complex tasks. A number of researchers have stressed the essential role of prospection for effective agency. ...
The chapter examines the question of whether people must fear for their jobs due to artificial intelligence (AI). A “competitive scenario” with a design to test for this appeared in the Delphi survey. The chapter shows how realistic this scenario is and its sociological implications, with a basis in expert opinions. In addition, the chapter sheds light on how much people see AI affecting themselves in their jobs, their future standard of living, and quality of life. The results in these respects paint a much more positive picture than the public discussion of AI leads us to expect. The chapter deals with ethical concerns that AI could lead to discrimination in the labor market and the perceived need for public policy interventions to ensure that AI develops ethically. An aggregate data analysis reveals substantial variations across EU countries and significant correlations with a country’s prosperity, risk of poverty, multi-ethnicity, and inherent trust in institutions and fellow men. We examine the odds of such concerns in Germany, as a function of socio-structural variables.
In this chapter, it is shown how second-order adaptive agent-based network models can be used to model a medical team supported by a virtual AI Coach. It is illustrated for the case of a newborn baby in danger. The design of these computational agent models is based on an adaptive self-modeling network modeling approach. It also addresses how the AI Coach can play a central role in organizational learning. The agent models enable representations and processing of all actors’ internal mental models and internal simulation of these mental models, adaptive changes of these mental models (learning and forgetting), and the interaction between the actors and the world.
This chapter presents an introduction to this book on a computational analysis approach for safety and security through cyberspace. It briefly introduces main concepts such as mental models, shared mental models, organizational learning, and how dynamics and adaptivity can be modeled by adaptive networks. Finally, it will provide an overview of the parts and chapters of the book.
Structural representations are likely the most talked about representational posits in the contemporary debate over cognitive representations. Indeed, the debate surrounding them is so vast virtually every claim about them has been made. Some, for instance, claimed structural representations are different from indicators. Others argued they are the same. Some claimed structural representations mesh perfectly with mechanistic explanations, others argued they can’t in principle mash. Some claimed structural representations are central to predictive processing accounts of cognition, others rebuked predictive processing networks are blissfully structural representation free. And so forth. Here, I suggest this confusing state of affairs is due to the fact that the term “structural representations” is applied to a number of distinct conceptions of representations. In this paper, I distinguish four such conceptions, argue that these four conceptions are actually distinct, and then show that such a fourfold distinction can be used to clarify some of the most pressing questions concerning structural representations and their role in cognitive theorizing, making these questions more easily answerable.
A video presentation of this paper at BICA'21 can be found at the YouTube Self-Modeling Networks channel at https://www.youtube.com/watch?v=StRGmqY0QD0. Within organisational learning literature, mental models are considered a vehicle for both individual learning and organisational learning. By learning individual mental models (and making them explicit), a basis for formation of shared mental models for the level of the organisation is created, which after its formation can then be adopted by individuals. This provides mechanisms for organisational learning. These mechanisms have been used as a basis for an adaptive computational network model. The model is illustrated by a not too complex but realistic case study.
The chapter focuses on research on robotic assistants and the involved challenge of their manipulating the physical world. It describes the state of the art in this regard and outlines directions for future research. Furthermore, it reports how the Delphi respondents assess various facets of human–robot communication and how specifically the group of scientists from engineering and natural sciences assesses the further technical development of 13 robotic skills. For this aspect, we asked for the experts’ assessment of the points in time when robots will presumably be capable of demonstrating such skills. The list of examples includes cognitive and communicative skills and skills that relate to motion, autonomous navigation, and the performance of everyday activities at home/in elderly care. In addition, the chapter reports on findings from the population survey. It particularly reveals the relative importance that people allocate to the skills of care robots. It underlines the importance of considering the impact of the physical design of a robot on its social perception and acceptance.
Perception has been for philosophers in the last few decades an area of compelling interest and intense investigation. In large part, the catalyst for this activity has come from contemporary cognitive science and neuroscience, which has been progressing at an accelerating pace, throwing up new information about the brain and new conceptions of how sensory information is processed and used. These new conceptions offer philosophers opportunities for reconceptualizing the senses—what they tell us, how we use them, and the nature of the knowledge they give us. Today, the philosophy of perception resonates with ideas that had not even been articulated in the 1970s and 1980s. This Handbook is a survey by leading philosophical thinkers of contemporary issues and new thinking in philosophy of perception. It includes sections on the history of the subject, introductions to contemporary issues in the epistemology, ontology, and aesthetics of perception, treatments of the individual sense modalities and of the things we perceive by means of them, and a consideration of how perceptual information is integrated and consolidated. New analytic tools and applications to other areas of philosophy are discussed in depth. Each of the forty-five entries is written by a leading expert, some collaborating with younger figures; each seeks to introduce the reader to a broad range of issues. All contain new ideas on the topic covered; together they are a portrait of the vigour and innovative zeal of a young field. The book is accessible to anybody who has an intellectual interest in issues concerning perception.
Recent research in cognitive and computational neuroscience portrays the neocortex as a hierarchically structured prediction machine. Several theorists have drawn on this research to challenge the traditional distinction between perception and cognition – specifically, to challenge the very idea that perception and cognition constitute useful kinds from the perspective of cognitive neuroscience. In place of this traditional taxonomy, such theorists advocate a unified inferential hierarchy subject to substantial bi-directional message passing. I outline the nature of this challenge and then raise two objections to the cognitive architecture it proposes as a replacement: first, standard ways of characterising this inferential hierarchy are in tension with the representational reach of conceptual thought; second, there is compelling evidence that commonsense reasoning is structured around highly domain-specific intuitive theories that are difficult to situate within a single hierarchy.
New essays by leading philosophers and cognitive scientists that present recent findings and theoretical developments in the study of concepts.
The study of concepts has advanced dramatically in recent years, with exciting new findings and theoretical developments. Core concepts have been investigated in greater depth and new lines of inquiry have blossomed, with researchers from an ever broader range of disciplines making important contributions. In this volume, leading philosophers and cognitive scientists offer original essays that present the state-of-the-art in the study of concepts. These essays, all commissioned for this book, do not merely present the usual surveys and overviews; rather, they offer the latest work on concepts by a diverse group of theorists as well as discussions of the ideas that should guide research over the next decade. The book is an essential companion volume to the earlier Concepts: Core Readings, the definitive source for classic texts on the nature of concepts.
The essays cover concepts as they relate to animal cognition, the brain, evolution, perception, and language, concepts across cultures, concept acquisition and conceptual change, concepts and normativity, concepts in context, and conceptual individuation. The contributors include such prominent scholars as Susan Carey, Nicola Clayton, Jerry Fodor, Douglas Medin, Joshua Tenenbaum, and Anna Wierzbicka.
ContributorsAurore Avarguès-Weber, Eef Ameel, Megan Bang, H. Clark Barrett, Pascal Boyer, Elisabeth Camp, Susan Carey, Daniel Casasanto, Nicola S. Clayton, Dorothy L. Cheney, Vyvyan Evans, Jerry A. Fodor, Silvia Gennari, Tobias Gerstenberg, Martin Giurfa, Noah D. Goodman, J. Kiley Hamlin, James A. Hampton, Mutsumi Imai, Charles W. Kalish, Frank Keil, Jonathan Kominsky, Stephen Laurence, Gary Lupyan, Edouard Machery, Bradford Z. Mahon, Asifa Majid, Barbara C. Malt, Eric Margolis, Douglas Medin, Nancy J. Nersessian, bethany ojalehto, Anna Papafragou, Joshua M. Plotnik, Noburo Saji, Robert M. Seyfarth, Joshua B. Tenenbaum, Sandra Waxman, Daniel A. Weiskopf, Anna Wierzbicka
Perception is concerned with how we use the information reaching our senses to guide and control our behaviour and create our particular, subjective experiences of the world. Perception: A Very Short Introduction discusses the philosophical question of what it means to perceive, and describes how we are able to perceive the particular characteristics of objects and scenes such as their lightness, colour, form, depth, and motion. The study of illusions can be useful in telling us something about the nature and limitations of our perceptual processes. This VSI explores perception from an evolutionary perspective, explaining how evolutionary pressures have shaped the perceptual systems of humans and other animals.
Cognitive Gadgets is a book about the cultural evolution of distinctively human cognitive mechanisms. Responding to commentators with different and broader interests, I argue that intelligent design has been more important in the formation of grist (technologies, practices and ideas) than of mills (cognitive mechanisms), and that embracing genetic accommodation would leave research on the origins of human cognition empirically unconstrained. I also underline the need to assess empirical methods; query the value of theories that merely accommodate existing data; and ask whether acquiring literacy is more laborious than learning to imitate, to talk and to read minds.
A cognitive science perspective on scientific development, drawing on philosophy, psychology, neuroscience, and computational modeling.
Many disciplines, including philosophy, history, and sociology, have attempted to make sense of how science works. In this book, Paul Thagard examines scientific development from the interdisciplinary perspective of cognitive science. Cognitive science combines insights from researchers in many fields: philosophers analyze historical cases, psychologists carry out behavioral experiments, neuroscientists perform brain scans, and computer modelers write programs that simulate thought processes.
Thagard develops cognitive perspectives on the nature of explanation, mental models, theory choice, and resistance to scientific change, considering disbelief in climate change as a case study. He presents a series of studies that describe the psychological and neural processes that have led to breakthroughs in science, medicine, and technology. He shows how discoveries of new theories and explanations lead to conceptual change, with examples from biology, psychology, and medicine. Finally, he shows how the cognitive science of science can integrate descriptive and normative concerns; and he considers the neural underpinnings of certain scientific concepts.
A novel proposal that the unified nature of our cognition can be partially explained by a cognitive architecture based on graphical models.
Our ordinary, everyday thinking requires an astonishing range of cognitive activities, yet our cognition seems to take place seamlessly. We move between cognitive processes with ease, and different types of cognition seem to share information readily. In this book, David Danks proposes a novel cognitive architecture that can partially explain two aspects of human cognition: its relatively integrated nature and our effortless ability to focus on the relevant factors in any particular situation. Danks argues that both of these features of cognition are naturally explained if many of our cognitive representations are understood to be structured like graphical models.
The computational framework of graphical models is widely used in machine learning, but Danks is the first to offer a book-length account of its use to analyze multiple areas of cognition. Danks demonstrates the usefulness of this approach by reinterpreting a variety of cognitive theories in terms of graphical models. He shows how we can understand much of our cognition—in particular causal learning, cognition involving concepts, and decision making—through the lens of graphical models, thus clarifying a range of data from experiments and introspection. Moreover, Danks demonstrates the important role that cognitive representations play in a unified understanding of cognition, arguing that much of our cognition can be explained in terms of different cognitive processes operating on a shared collection of cognitive representations. Danks's account is mathematically accessible, focusing on the qualitative aspects of graphical models and separating the formal mathematical details in the text.
In Reconstructing the Cognitive World, Michael Wheeler argues that we should turn away from the generically Cartesian philosophical foundations of much contemporary cognitive science research and proposes instead a Heideggerian approach. Wheeler begins with an interpretation of Descartes. He defines Cartesian psychology as a conceptual framework of explanatory principles and shows how each of these principles is part of the deep assumptions of orthodox cognitive science (both classical and connectionist). Wheeler then turns to Heidegger's radically non-Cartesian account of everyday cognition, which, he argues, can be used to articulate the philosophical foundations of a genuinely non-Cartesian cognitive science. Finding that Heidegger's critique of Cartesian thinking falls short, even when supported by Hubert Dreyfus's influential critique of orthodox artificial intelligence, Wheeler suggests a new Heideggerian approach. He points to recent research in "embodied-embedded" cognitive science and proposes a Heideggerian framework to identify, amplify, and clarify the underlying philosophical foundations of this new work. He focuses much of his investigation on recent work in artificial intelligence-oriented robotics, discussing, among other topics, the nature and status of representational explanation, and whether (and to what extent) cognition is computation rather than a noncomputational phenomenon best described in the language of dynamical systems theory.
Wheeler's argument draws on analytic philosophy, continental philosophy, and empirical work to "reconstruct" the philosophical foundations of cognitive science in a time of a fundamental shift away from a generically Cartesian approach. His analysis demonstrates that Heideggerian continental philosophy and naturalistic cognitive science need not be mutually exclusive and shows further that a Heideggerian framework can act as the "conceptual glue" for new work in cognitive science.
Bradford Books imprint