DOCSLIB.ORG

Artificial Life Applied to Adaptive Information Agents

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Artificial Life Applied to Adaptive Information Agents From: AAAI Technical Report SS-95-08. Compilation copyright © 1995, AAAI (www.aaai.org). All rights reserved. Artificial Life Applied to Adaptive Information Agents Filippo Menczer Richard K. Belew Computer Science and Engineering Department, 0114 University of California, San Diego La Jolla, California 92093, USA {fil, rik} @cs.ucsd.edu Wolfram Willuhn CommunicationTechnology Lab, Image Science Group ETH-Zentrum, ETZ F86 8092 Zurich, Switzerland wolfram@ vision.ee.ethz.ch Abstract the complexity of such an environment. Traditional genetic Wepropose a model,inspired by recent artificial life theory, algorithms [Mitchell & Forrest 1994] are characterized by applied to the problemof retrieving information from a exploitation of information, but the distributed information large, distributed collection of documentssuch as the World gathering problem requires adaptation rather than WideWeb. A populationof agents is evolvedunder density optimization. Therefore we propose a model inspired by dependentselection for the task of locating informationfor the endogenous fitness metaphor to search for relevant the user. Theenergy necessary for survivalis obtainedfrom information through a population of intelligent agents both environmentand user in exchangefor appropriate evolving in the WWWenvironment. This paper presents information. By competingfor relevant documents,the various extensions to the work recently reported in agents robustly adapt to their informationenvironment and [Menczer, Willuhn & Belew 1994]. are allocated to efficiently exploit shared resources. We illustrate the roles playedby documentlocality, adaptive searchstrategies, andrelevance feedback, in the information Background gatheringprocess. The traditional solution to the problem of information retrieval (IR) on the WWWis to build a database index Introduction* all the documents, on which to use traditional search and retrieval techniques. Such virtual libraries are built and The World Wide Web (WWW)is an information updated off-line, either in a user-driven fashion or by environment madeof a very large distributed database of automated exhaustive programs, called spiders or robots. A heterogeneous documents, using a wide-area network and a well-known robot is the WWWWorm [McBryan 1994]. client-server protocol. The structure of this environmentis The off-line approachhas several drawbacks:first, the size that of a graph, where the nodes (documents)are connected of the WWWmakes it increasingly difficult to update by hyperlinks. The typical strategy for accessing virtual libraries without inefficient use of network information on the WWWis to navigate across documents resources. Moreover, any database has to abstract away through hyperlinks, retrieving the information of interest important information, concerning both the content of along the way. The dynamic and distributed nature of the documentsand their structure [Belew 1985]. Finally, the environment, however, makes retrieving specific information gathering and retrieval processes are information a hard task [De Bra & Post 1994, McBryan independent and therefore feedback from the latter cannot 1994]. be used to adaptively improveefficiency nor quality of the On the other hand, the WWWrepresents an ideal former. environmentin which to apply techniques recently matured To remedy someof these problems, the client-based Fish in the field of artificial life (ALife). In particular, adaptive Search algorithm has been proposed [De Bra & Post 1994]. and distributed algorithms seem to appropriately capture This approach uses the metaphorof a school of fish, where agents in a population survive, reproduce, and die based on the energy gained from their performance in the retrieval Researchpartly supportedby a fellowshipfrom the Apple task. This approach, however, stops short of solving the AdvancedTechnology Group remaining problems: no mutation occurs at reproduction, queries, just as different behaviors may be optimal for and each agent in the population follows a non-adaptive, different environments. Only the end results of a search exhaustive depth-first search algorithm. While some (the retrieved documents)can be evaluated, and the agents heuristics are used to order the graph traversal, the lack of identify the relevant information from its correlation with intelligent cutting of search branches results in slow speed energy. Therefore populations will evolve strategies and high network consumption (caching being proposed as effective in the different environments specified by both a palliative). information space and queries. Adaptation means for The idea of adaptive IR is not new [Belew 1989]. More agents to concentrate in high energy areas of the Web, recently, learning agents and traditional genetic algorithms where manydocuments are relevant. Each agent’s survival have been successfully applied to information retrieval will be ensured by exchanging an adequate flow of [Yang & Korfhage 1993] and information filtering [Maes information for energy. & Kozierok 1993, Sheth & Maes 1993]. Workin progress indicates that learning is sped up by extending such models to collaborative multi-agent systems [Lashkari, Metral, & An Implementation Maes1994]. Using a distributed population of cooperating best-first search agents has been recently proposed in the Wehave implemented the endogenous fitness model in a WebAntsproject [Mauldin & Leavitt 1994] to overcome simple prototype of IR system for the WWW.The the single-server and single-client bottlenecks. algorithm is illustrated in Table I. The user provides a query consisting of a set of keywords. A population of agents is initialized with some energy, some random Information Search by Endogenous Fitness strategy, and somedistribution in the Web.The ideal, zero- knowledge assumption is to start with a population at Endogenousfitness models are becoming an increasingly minimaldistance from all nodes. Typical heuristics suggest appreciated and well-understood paradigm in the ALife to initialize the population with a uniformdistribution in a community[Mitchell & Forrest 1994]. While a thorough default set of knownstarting points, or better yet in the discussion of the subject is out of the scope of this paper documentsreturned by a preliminary call to a traditional [see, e.g., Menczer & Belew 1995], we outline in this search engine. section the main aspects that makethis paradigmuseful for the problem of IR in the WWW. INPUT:n>O, e>O, t>O ; queryword(s) A population of agents becomesevolutionary adapted in a dynamicenvironment by a steady-state genetic algorithm. 1 initialize populationof n agentswith random,6, 7, E Energy is the single currency by which agents survive, reproduce, and die, and it must be positively correlated 2 while numberof agents> 0 do for all agents: with someperformance measure for the task defined by the 2.1 computefor eachlink i in current documentthe environment. Agents asynchronously go through a simple estimateL i = sumof occurrencesof eachquery cycle in which they receive input from the environmentas wordin document,weighted inversely to the well as internal state, perform some computation, and numberof headersbetween the positions of i and execute actions. Actions have an energy cost but mayresult the wordin the document in energy intake. Energy is used up and accumulated pick link i to follow accordingto the probability internally throughout an agent’s life; its internal level automatically determines reproduction and death, events in distribution Pi=exp(flLi)//~i exp(~Lj) whichenergy is conserved. Agents that perform the task better than average reproduce more and colonize the population. Indirect 2.2 E ~- E- N + eR whereN = server accesscost measureand current document’srelevance R = interaction among agents occurs without the need of numberof occurrencesof querywords in expensive communication,via competition for the shared, document finite environmental resources. Mutations afford the changes necessary for the evolution of dynamicallyadapted 1 agents. This paradigm enforces density-dependent if ( R + I ) 7 > -ff =:~E ~-- E + where F = user selection: the expected population size is determinedby the feedbackenergy (optional) carrying capacity of the environment. Associating high 2.3 if E>t cloneagent, split parent’senergy with energy costs with expensiveactions intrinsically enforces a offspring, mutateoffspring’s/~, % balanced network load by limiting inefficient uses of bandwidth. else if E < 0 destroyagent In the heterogeneous environment of the WWW,it is Table1: Algorithmfor adaptiveinformation agents. hard to associate a fitness measure with a strategy in general, but it is easy to estimate the results of a strategy In order to keep search strategies simple while allowing applied to a particular query. Different information search adaptivity, stochastic selection is used to navigate across and retrieval strategies may be optimal for different hyperlinks. For each cycle, each agent estimates the 129 hyperlinks from the current node to decide which node to Figure1. Thefact that this collection is closed to the rest of visit next. The estimates, based on a fixed matching the WWWis only one of its limitations. function of the current documentand the user keywords, are scaled by a non-decreasing function to obtain a probability distribution that is in turn used by a stochastic selector. The slope of the non-linear scaling function is determined by the agent’s genetic parameter fl This trait represent the adaptive part of the agent’s search
Load more

Recommended publications
  • Close Engagements with Artificial Companions: Key Social, Psychological, Ethical and Design Issues
    OII / e-Horizons Forum Discussion Paper No. 14, January 2008 Close Engagements with Artificial Companions: Key Social, Psychological, Ethical and Design Issues by Malcolm Peltu Oxford Internet Institute Yorick Wilks Oxford Internet Institute OII / e-Horizons Forum Discussion Paper No. 14 Oxford Internet Institute University of Oxford 1 St Giles, Oxford OX1 3JS United Kingdom Forum Discussion Paper January 2008 © University of Oxford for the Oxford Internet Institute 2008 Close Engagements with Artificial Companions Foreword This paper summarizes discussions at the multidisciplinary forum1 held at the University of Oxford on 26 October 2007 entitled Artificial Companions in Society: Perspectives on the Present and Future, as well as an open meeting the previous day addressed by Sherry Turkle2. The event was organized by Yorick Wilks, Senior Research Fellow for the Oxford Internet Institute (OII)3, on behalf of the e-Horizons Institute4 and in association with the EU Integrated Project COMPANIONS. COMPANIONS is studying conversational software-based artificial agents that will get to know their owners over a substantial period. These could be developed to advise, comfort and carry out a wide range of functions to support diverse personal and social needs, such as to be ‘artificial companions’ for the elderly, helping their owners to learn, or assisting to sustain their owners’ fitness and health. The invited forum participants, including computer and social scientists, also discussed a range of related developments that use advanced artificial intelligence and human– computer interaction approaches. They examined key issues in building artificial companions, emphasizing their social, personal, emotional and ethical implications. This paper summarizes the main issues raised.
    [Show full text]
  • Open Problems in Artificial Life
    Open Problems in Artificial Life Mark A. Bedau,† John S. McCaskill‡ Norman H. Packard§ Steen Rasmussen Chris Adami†† David G. Green‡‡ Takashi Ikegami§§ Abstract This article lists fourteen open problems in Kunihiko Kaneko artificial life, each of which is a grand challenge requiring a Thomas S. Ray††† major advance on a fundamental issue for its solution. Each problem is briefly explained, and, where deemed helpful, some promising paths to its solution are indicated. Keywords artificial life, evolution, self-organi- zation, emergence, self-replication, autopoeisis, digital life, artificial chemistry, selection, evolutionary learning, ecosystems, biosystems, astrobiology, evolvable hardware, dynamical hierarchies, origin of life, simulation, cultural evolution, 1 Introduction information theory At the dawn of the last century, Hilbert proposed a set of open mathematical problems. They proved to be an extraordinarily effective guideline for mathematical research in the following century. Based on a substantial body of existing mathematical theory, the challenges were both precisely formulated and positioned so that a significant body of missing theory needed to be developed to achieve their solution, thereby enriching mathematics as a whole. In contrast with mathematics, artificial life is quite young and essentially interdis- ciplinary. The phrase “artificial life” was coined by C. Langton [11], who envisaged an investigation of life as it is in the context of life as it could be. Although artificial life is fundamentally directed towards both the origins of biology and its future, the scope and complexity of its subject require interdisciplinary cooperation and collabo- ration. This broadly based area of study embraces the possibility of discovering lifelike behavior in unfamiliar settings and creating new and unfamiliar forms of life, and its major aim is to develop a coherent theory of life in all its manifestations, rather than an historically contingent documentation bifurcated by discipline.
    [Show full text]
  • Open-Endedness for the Sake of Open-Endedness
    Open-Endedness for the Sake Arend Hintze Michigan State University of Open-Endedness Department of Integrative Biology Department of Computer Science and Engineering BEACON Center for the Study of Evolution in Action [email protected] Abstract Natural evolution keeps inventing new complex and intricate forms and behaviors. Digital evolution and genetic algorithms fail to create the same kind of complexity, not just because we still lack the computational resources to rival nature, but Keywords because (it has been argued) we have not understood in principle how to create open-ended evolving systems. Much effort has been Open-ended evolution, computational model, complexity, diversity made to define such open-endedness so as to create forms of increasing complexity indefinitely. Here, however, a simple evolving computational system that satisfies all such requirements is presented. Doing so reveals a shortcoming in the definitions for open-ended evolution. The goal to create models that rival biological complexity remains. This work suggests that our current definitions allow for even simple models to pass as open-ended, and that our definitions of complexity and diversity are more important for the quest of open-ended evolution than the fact that something runs indefinitely. 1 Introduction Open-ended evolution has been identified as a key challenge in artificial life research [4]. Specifically, it has been acknowledged that there is a difference between an open-ended system and a system that just has a long run time. Large or slow systems will eventually converge on a solution, but open- ended systems will not and instead keep evolving. Interestingly, a system that oscillates or that is in a dynamic equilibrium [21] would continuously evolve, but does not count as an open-ended evolving system.
    [Show full text]
  • Artificial Intelligence: How Does It Work, Why Does It Matter, and What Can We Do About It?
    Artificial intelligence: How does it work, why does it matter, and what can we do about it? STUDY Panel for the Future of Science and Technology EPRS | European Parliamentary Research Service Author: Philip Boucher Scientific Foresight Unit (STOA) PE 641.547 – June 2020 EN Artificial intelligence: How does it work, why does it matter, and what can we do about it? Artificial intelligence (AI) is probably the defining technology of the last decade, and perhaps also the next. The aim of this study is to support meaningful reflection and productive debate about AI by providing accessible information about the full range of current and speculative techniques and their associated impacts, and setting out a wide range of regulatory, technological and societal measures that could be mobilised in response. AUTHOR Philip Boucher, Scientific Foresight Unit (STOA), This study has been drawn up by the Scientific Foresight Unit (STOA), within the Directorate-General for Parliamentary Research Services (EPRS) of the Secretariat of the European Parliament. To contact the publisher, please e-mail [email protected] LINGUISTIC VERSION Original: EN Manuscript completed in June 2020. DISCLAIMER AND COPYRIGHT This document is prepared for, and addressed to, the Members and staff of the European Parliament as background material to assist them in their parliamentary work. The content of the document is the sole responsibility of its author(s) and any opinions expressed herein should not be taken to represent an official position of the Parliament. Reproduction and translation for non-commercial purposes are authorised, provided the source is acknowledged and the European Parliament is given prior notice and sent a copy.
    [Show full text]
  • How Artificial Intelligence Works
    BRIEFING How artificial intelligence works From the earliest days of artificial intelligence (AI), its definition focused on how its results appear to show intelligence, rather than the methods that are used to achieve it. Since then, AI has become an umbrella term which can refer to a wide range of methods, both current and speculative. It applies equally well to tools that help doctors to identify cancer as it does to self-replicating robots that could enslave humanity centuries from now. Since AI is simultaneously represented as high-risk, low-risk and everything in between, it is unsurprising that it attracts controversy. This briefing provides accessible introductions to some of the key techniques that come under the AI banner, grouped into three sections to give a sense the chronology of its development. The first describes early techniques, described as 'symbolic AI' while the second focuses on the 'data driven' approaches that currently dominate, and the third looks towards possible future developments. By explaining what is 'deep' about deep learning and showing that AI is more maths than magic, the briefing aims to equip the reader with the understanding they need to engage in clear-headed reflection about AI's opportunities and challenges, a topic that is discussed in the companion briefing, Why artificial intelligence matters. First wave: Symbolic artificial intelligence Expert systems In these systems, a human expert creates precise rules that a computer can follow, step by step, to decide how to respond to a given situation. The rules are often expressed in an 'if-then-else' format. Symbolic AI can be said to 'keep the human in the loop' because the decision-making process is closely aligned to how human experts make decisions.
    [Show full text]
  • Biological Computation: a Road
    918 Biological Computation: A Road to Complex Engineered Systems Nelson Alfonso Gómez-Cruz Modeling and Simulation Laboratory Universidad del Rosario Bogotá, Colombia [email protected] Carlos Eduardo Maldonado Research Professor Universidad del Rosario Bogotá, Colombia [email protected] Provided that there is no theoretical frame for complex engineered systems (CES) as yet, this paper claims that bio-inspired engineering can help provide such a frame. Within CES bio- inspired systems play a key role. The disclosure from bio-inspired systems and biological computation has not been sufficiently worked out, however. Biological computation is to be taken as the processing of information by living systems that is carried out in polynomial time, i.e., efficiently; such processing however is grasped by current science and research as an intractable problem (for instance, the protein folding problem). A remark is needed here: P versus NP problems should be well defined and delimited but biological computation problems are not. The shift from conventional engineering to bio-inspired engineering needs bring the subject (or problem) of computability to a new level. Within the frame of computation, so far, the prevailing paradigm is still the Turing-Church thesis. In other words, conventional 919 engineering is still ruled by the Church-Turing thesis (CTt). However, CES is ruled by CTt, too. Contrarily to the above, we shall argue here that biological computation demands a more careful thinking that leads us towards hypercomputation. Bio-inspired engineering and CES thereafter, must turn its regard toward biological computation. Thus, biological computation can and should be taken as the ground for engineering complex non-linear systems.
    [Show full text]
  • What Was Life? Answers from Three Limit Biologies Author(S): Stefan Helmreich Source: Critical Inquiry, Vol
    What Was Life? Answers from Three Limit Biologies Author(s): Stefan Helmreich Source: Critical Inquiry, Vol. 37, No. 4 (Summer 2011), pp. 671-696 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/660987 . Accessed: 25/08/2011 13:15 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Critical Inquiry. http://www.jstor.org What Was Life? Answers from Three Limit Biologies Stefan Helmreich “What was life? No one knew.” —THOMAS MANN, The Magic Mountain What is life? A gathering consensus in anthropology, science studies, and philosophy of biology suggests that the theoretical object of biology, “life,” is today in transformation, if not dissolution. Proliferating repro- ductive technologies, along with genomic reshufflings of biomatter in such practices as cloning, have unwound the facts of life.1 Biotechnology, bio- This paper grew from a presentation at “Vitalism Revisited: History, Philosophy, Biology,” at the Center for Interdisciplinary Studies in Science and Cultural Theory, Duke University, 22 Mar. 2008. I thank Barbara Herrnstein Smith for inviting me. The paper went through revision for “Extreme: Histories and Economies of Humanness Inside Outerspaces,” at the American Anthropological Association meeting, 2–6 Dec.
    [Show full text]
  • Artificial Intelligence in Computer Graphics: a Constructionist Approach
    Artificial Intelligence in Computer Graphics: A Constructionist Approach Kristinn R.Thórisson,Christopher ology aims to help coordinate the effort. approach to building AI systems. The Pennock,Thor List,John DiPirro There is also a lack of incremental accumula- Constructionist AI Methodology (CAIM) is an Communicative Machines tion of knowledge in AI and related computer emerging set of principles for designing and graphics. By supporting reuse of prior work implementing interactive intelligences, Introduction we enable the building of increasingly speeding up implementation of relatively The ultimate embodiment of artificial intelli- powerful systems, as core system elements complex, multi-functional systems with full, gence (AI) has traditionally been seen as do not need to be built from scratch. real-time perception-action loop capabilities. physical robots. However, hardware is expen- Background and Motivation It helps novices as well as experts with the sive to construct and difficult to modify. The The creation of believable, interactive charac- creation of embodied, virtual agents that can connection between graphics and AI is ters is a challenge familiar to many game perceive and act in real time and interact with becoming increasingly strong, and, on the one developers. Systems that have to respond to virtual as well as real environments. The hand, it is now clear that AI research can asynchronous inputs in real time, where the embodiment of such humanoids can come in benefit tremendously from embodiment in inputs’ timing is largely unpredictable, can various levels of completeness, from just a virtual worlds [1, 4, 5]. On the other hand, pose a serious challenge to the system design.
    [Show full text]
  • Art and Artificial Life – a Primer
    Art and Artificial Life – a Primer Simon Penny University of California, Irvine [email protected] ABSTRACT December 2009) which is a testament to the enduring and It was not until the late 1980s that the term ‘Artificial Life’ arose inspirational intellectual significance of ideas associated with as a descriptor of a range of (mostly) computer based research Artificial Life. practices which sought alternatives to conventional Artificial Intelligence methods as a source of (quasi-) intelligent behavior in technological systems and artifacts. These practices included Artificial Life could not have emerged as a persuasive paradigm reactive and bottom-up robotics, computational systems which without the easy availability of computation. This is not simply to simulated evolutionary and genetic processes, and are range of proclaim, as did Christopher Langton, that Artificial Life was an other activities informed by biology and complexity theory. A exploration of life on a non-carbon substrate, but that Artificial general desire was to capture, harness or simulate the generative Life is ‘native’ to computing in the sense that large scale iterative and ‘emergent’ qualities of ‘nature’ - of evolution, co-evolution process is crucial to the procedures which generate (most) and adaptation. ‘Emergence’ was a keyword in the discourse. Two artificial life phenomena. The notion that Artificial Life is life decades later, the discourses of Artificial Life continues to have created an ethico-philosophical firestorm concerning intelligence, intellectual force,
    [Show full text]
  • Artificial Life
    ARTIFICIAL LIFE Mark A. Bedau Contemporary artificial life (also known as “ALife”) is an interdisciplinary study of life and life-like processes. Its two most important qualities are that it focuses on the essential rather than the contingent features of living systems and that it attempts to understand living systems by artificially synthesizing extremely simple forms of them. These two qualities are connected. By synthesizing simple systems that are very life-like and yet very unfamiliar, artificial life constructively explores the boundaries of what is possible for life. At the moment, artificial life uses three different kinds of synthetic methods. “Soft” artificial life creates computer simula- tions or other purely digital constructions that exhibit life-like behavior. “Hard” artificial life produces hardware implementations of life-like systems. “Wet” artifi- cial life involves the creation of life-like systems in a laboratory using biochemical materials. Contemporary artificial life is vigorous and diverse. So this chapter’s first goal is to convey what artificial life is like. It first briefly reviews the history of artificial life and illustrates the current research thrusts in contemporary “soft”, “hard”, and “wet” artificial life with respect to individual cells, whole organisms, and evolving populations. Artificial life also raises and informs a number of philosophical is- sues concerning such things as emergence, evolution, life, mind, and the ethics of creating new forms of life from scratch. This chapter’s second goal is to illustrate these philosophical issues, discuss some of their complexities, and suggest the most promising avenues for making further progress. 1 HISTORY AND METHODOLOGY Contemporary artificial life became known as such when Christopher Langton coined the phrase “artificial life” in the 1980s.
    [Show full text]
  • Applying Evolutionary Computation to Mitigate Uncertainty in Dynamically-Adaptive, High-Assurance Middleware
    J Internet Serv Appl (2012) 3:51–58 DOI 10.1007/s13174-011-0049-4 SI: FOME - THE FUTURE OF MIDDLEWARE Applying evolutionary computation to mitigate uncertainty in dynamically-adaptive, high-assurance middleware Philip K. McKinley · Betty H.C. Cheng · Andres J. Ramirez · Adam C. Jensen Received: 31 October 2011 / Accepted: 12 November 2011 / Published online: 3 December 2011 © The Brazilian Computer Society 2011 Abstract In this paper, we explore the integration of evolu- systems need to perform complex, distributed tasks despite tionary computation into the development and run-time sup- lossy wireless communication, limited resources, and uncer- port of dynamically-adaptable, high-assurance middleware. tainty in sensing the environment. However, even systems The open-ended nature of the evolutionary process has been that operate entirely within cyberspace need to account for shown to discover novel solutions to complex engineering changing network and load conditions, component failures, problems. In the case of high-assurance adaptive software, and exposure to a wide variety of cyber attacks. however, this search capability must be coupled with rigor- A dynamically adaptive system (DAS) monitors itself and ous development tools and run-time support to ensure that its execution environment at run time. This monitoring en- the resulting systems behave in accordance with require- ables a DAS not only to detect conditions warranting a re- ments. Early investigations are reviewed, and several chal- configuration, but also to determine when and how to safely lenging problems and possible research directions are dis- reconfigure itself in order to deliver acceptable behavior be- cussed. fore, during, and after adaptation [2].
    [Show full text]
  • On the Possibility of Robots Having Emotions
    Georgia State University ScholarWorks @ Georgia State University Philosophy Theses Department of Philosophy Summer 8-12-2014 On the Possibility of Robots Having Emotions Cameron Hamilton Follow this and additional works at: https://scholarworks.gsu.edu/philosophy_theses Recommended Citation Hamilton, Cameron, "On the Possibility of Robots Having Emotions." Thesis, Georgia State University, 2014. https://scholarworks.gsu.edu/philosophy_theses/150 This Thesis is brought to you for free and open access by the Department of Philosophy at ScholarWorks @ Georgia State University. It has been accepted for inclusion in Philosophy Theses by an authorized administrator of ScholarWorks @ Georgia State University. For more information, please contact [email protected]. ON THE POSSIBILITY OF ROBOTS HAVING EMOTIONS by CAMERON REID HAMILTON Under the Direction of Andrea Scarantino ABSTRACT I argue against the commonly held intuition that robots and virtual agents will never have emotions by contending robots can have emotions in a sense that is functionally similar to hu- mans, even if the robots' emotions are not exactly equivalent to those of humans. To establish a foundation for assessing the robots' emotional capacities, I first define what emotions are by characterizing the components of emotion consistent across emotion theories. Second, I dissect the affective-cognitive architecture of MIT's Kismet and Leonardo, two robots explicitly de- signed to express emotions and to interact with humans, in order to explore whether they have emotions.
    [Show full text]