The first decade of this century has seen the nascency of the first mathematical theory of general artificial intelligence. This theory of Universal Artificial Intelligence (UAI) has made significant contributions to many theoretical, philosophical, and practical AI questions. In a series of papers culminating in book (Hutter, 2005), an exciting sound and complete mathematical model for a super intelligent agent (AIXI) has been developed and rigorously analyzed. While nowadays most AI researchers avoid discussing intelligence, the award-winning PhD thesis (Legg, (...) 2008) provided the philosophical embedding and investigated the UAI-based universal measure of rational intelligence, which is formal, objective and non-anthropocentric. Recently, effective approximations of AIXI have been derived and experimentally investigated in JAIR paper (Veness et al. 2011). This practical breakthrough has resulted in some impressive applications, finally muting earlier critique that UAI is only a theory. For the first time, without providing any domain knowledge, the same agent is able to self-adapt to a diverse range of interactive environments. For instance, AIXI is able to learn from scratch to play TicTacToe, Pacman, Kuhn Poker, and other games by trial and error, without even providing the rules of the games. These achievements give new hope that the grand goal of Artificial General Intelligence is not elusive. This article provides an informal overview of UAI in context. It attempts to gently introduce a very theoretical, formal, and mathematical subject, and discusses philosophical and technical ingredients, traits of intelligence, some social questions, and the past and future of UAI. (shrink)

Many virtue epistemologists conceive of epistemic competence on the model of skill —such as archery, playing baseball, or chess. In this paper, I argue that this is a mistake: epistemic competences and skills are crucially and relevantly different kinds of capacities. This, I suggest, undermines the popular attempt to understand epistemic normativity as a mere special case of the sort of normativity familiar from skilful action. In fact, as I argue further, epistemic competences resemble virtues rather than skills—a claim that (...) is based on an important, but often overlooked, difference between virtue and skill. The upshot is that virtue epistemology should indeed be based on virtue, not on skill. (shrink)

Epistemic constitutivism (EC) holds that the nature of believing is such that it gives rise to a standard of correctness and that other epistemic normative notions (e.g., reasons for belief) can be explained in terms of this standard. If defensible, this view promises an attractive and unifying account of epistemic normativity. However, EC faces a forceful objection: that constitutive standards of correctness are never enough for generating normative reasons. This paper aims to defend EC in the face of this objection. (...) I do so in two steps. First, I dispute a crucial assumption underlying the case against EC: that constitutive standards of correctness in general are ‘reason-giving’ only if and because there is also a prior reason to comply with them. Second, I outline a strategy of how EC can meet the challenge of explaining what’s special about the activity of believing such that, unlike other standard-governed activities, it is capable of generating normative reasons. (shrink)

Many hold that theoretical reasoning aims at truth. In this paper, I ask what it is for reasoning to be thus aim-directed. Standard answers to this question explain reasoning’s aim-directedness in terms of intentions, dispositions, or rule-following. I argue that, while these views contain important insights, they are not satisfactory. As an alternative, I introduce and defend a novel account: reasoning aims at truth in virtue of being the exercise of a distinctive kind of cognitive power, one that, unlike ordinary (...) dispositions, is capable of fully explaining its own exercises. I argue that this account is able to avoid the difficulties plaguing standard accounts of the relevant sort of mental teleology. (shrink)

Can an agent's intelligence level be negative? We extend the Legg-Hutter agent-environment framework to include punishments and argue for an affirmative answer to that question. We show that if the background encodings and Universal Turing Machine (UTM) admit certain Kolmogorov complexity symmetries, then the resulting Legg-Hutter intelligence measure is symmetric about the origin. In particular, this implies reward-ignoring agents have Legg-Hutter intelligence 0 according to such UTMs.

Automated reasoning about uncertain knowledge has many applications. One difficulty when developing such systems is the lack of a completely satisfactory integration of logic and probability. We address this problem directly. Expressive languages like higher-order logic are ideally suited for representing and reasoning about structured knowledge. Uncertain knowledge can be modeled by using graded probabilities rather than binary truth-values. The main technical problem studied in this paper is the following: Given a set of sentences, each having some probability of being (...) true, what probability should be ascribed to other (query) sentences? A natural wish-list, among others, is that the probability distribution (i) is consistent with the knowledge base, (ii) allows for a consistent inference procedure and in particular (iii) reduces to deductive logic in the limit of probabilities being 0 and 1, (iv) allows (Bayesian) inductive reasoning and (v) learning in the limit and in particular (vi) allows confirmation of universally quantified hypotheses/sentences. We translate this wish-list into technical requirements for a prior probability and show that probabilities satisfying all our criteria exist. We also give explicit constructions and several general characterizations of probabilities that satisfy some or all of the criteria and various (counter) examples. We also derive necessary and sufficient conditions for extending beliefs about finitely many sentences to suitable probabilities over all sentences, and in particular least dogmatic or least biased ones. We conclude with a brief outlook on how the developed theory might be used and approximated in autonomous reasoning agents. Our theory is a step towards a globally consistent and empirically satisfactory unification of probability and logic. (shrink)

Someone acts in a morally worthy way when they deserve credit for doing the morally right thing. But when and why do agents deserve credit for the success involved in doing the right thing? It is tempting to seek an answer to that question by drawing an analogy with creditworthy success in other domains of human agency, especially in sports, arts, and crafts. Accordingly, some authors have recently argued that, just like creditworthy success in, say, chess, playing the piano, or (...) archery, creditworthy moral success is a matter of getting things right by way of manifesting a relevant skill. My main aim in this paper is to bring out an important structural difference between moral creditworthiness and creditworthiness in sports, arts, and craft, undermining attempts to use examples of the latter as a model for understanding the former. As an alternative, I propose an account of morally creditworthy action, according to which such action is a matter of manifesting virtue, not skill—a claim that's based on an important, but underappreciated, difference between the sorts of excellences constituting virtues and skills. The paper thus contributes to a more nuanced picture of normative achievements across different domains of human agency, highlighting largely overlooked structural dissimilarities among them. (shrink)

In this paper, we provide a detailed review of previous classifications of 2 × 2 games and suggest a mathematically simple way to classify the symmetric 2 × 2 games based on a decomposition of the payoff matrix into a cooperative and a zero-sum part. We argue that differences in the interaction between the parts is what makes games interesting in different ways. Our claim is supported by evolutionary computer experiments and findings in previous literature. In addition, we provide a (...) method for using a stereographic projection to create a compact 2-d representation of the game space. (shrink)

Inspired by recent progress in multi-agent Reinforcement Learning (RL), in this work we examine the collective intelligent behaviour of theoretical universal agents by introducing a weighted mixture operation. Given a weighted set of agents, their weighted mixture is a new agent whose expected total reward in any environment is the corresponding weighted average of the original agents' expected total rewards in that environment. Thus, if RL agent intelligence is quantified in terms of performance across environments, the weighted mixture's intelligence is (...) the weighted average of the original agents' intelligences. This operation enables various interesting new theorems that shed light on the geometry of RL agent intelligence, namely: results about symmetries, convex agent-sets, and local extrema. We also show that any RL agent intelligence measure based on average performance across environments, subject to certain weak technical conditions, is identical (up to a constant factor) to performance within a single environment dependent on said intelligence measure. (shrink)

Die Übernahme des KI-Unternehmens DeepMind durch Google für rund eine halbe Milliarde US-Dollar signalisierte vor einem Jahr, dass von der KI-Forschung vielversprechende Ergebnisse erwartet werden. Spätestens seit bekannte Wissenschaftler wie Stephen Hawking und Unternehmer wie Elon Musk oder Bill Gates davor warnen, dass künstliche Intelligenz eine Bedrohung für die Menschheit darstellt, schlägt das KI-Thema hohe Wellen. Die Stiftung für Effektiven Altruismus (EAS, vormals GBS Schweiz) hat mit der Unterstützung von Experten/innen aus Informatik und KI ein umfassendes Diskussionspapier zu den Chancen (...) und Risiken der künstlichen Intelligenz verfasst und geht darin auf aktuelle, mittel- und langfristige Herausforderungen im Bereich der KI-Entwicklung ein. (shrink)

Are the real numbers rich enough to measure intelligence? We generalize a result of Alexander and Hutter about the so-called Legg-Hutter intelligence measures of reinforcement learning agents. Using the generalized result, we exhibit a paradox: in one particular version of the Legg-Hutter intelligence measure, certain agents all have intelligence 0, even though in a certain sense some of them outperform others. We show that this paradox disappears if we vary the Legg-Hutter intelligence measure to be hyperreal-valued (...) rather than real-valued. (shrink)

In their thought-provoking paper, Legg and Hutter consider a certain abstrac- tion of an intelligent agent, and define a universal intelligence measure, which assigns every such agent a numerical intelligence rating. We will briefly summarize Legg and Hutter’s paper, and then give a tongue-in-cheek argument that if one’s goal is to become more intelligent by cultivating music appreciation, then it is bet- ter to use classical music (such as Bach, Mozart, and Beethoven) than to use more recent pop (...) music. The same argument could be adapted to other media: books beat films, card games beat first-person shooters, parables beat dissertations, etc. We leave it to the reader to decide whether this argument tells us something about classical music, something about Legg-Hutter intelligence, or something about both. (shrink)

Legg and Hutter, as well as subsequent authors, considered intelligent agents through the lens of interaction with reward-giving environments, attempting to assign numeric intelligence measures to such agents, with the guiding principle that a more intelligent agent should gain higher rewards from environments in some aggregate sense. In this paper, we consider a related question: rather than measure numeric intelligence of one Legg- Hutter agent, how can we compare the relative intelligence of two Legg-Hutter agents? We propose (...) an elegant answer based on the following insight: we can view Legg-Hutter agents as candidates in an election, whose voters are environments, letting each environment vote (via its rewards) which agent (if either) is more intelligent. This leads to an abstract family of comparators simple enough that we can prove some structural theorems about them. It is an open question whether these structural theorems apply to more practical intelligence measures. (shrink)

This paper proposes a comparison between the myth of Prometheus and Pliny the Elder’s account of the origins of painting to assess the technical characterof images. Such a character is seen as a precondition for understanding their epistemological value, as well as for supporting the need for hermeneutics at any level of artificial intervention. Finally, part of Horst Bredekamp’s work will be presented as an example of a methodology that combines hermeneutics and iconology based on such a technical character.

Since the noun phrase `artificial intelligence' (AI) was coined, it has been debated whether humans are able to create intelligence using technology. We shed new light on this question from the point of view of themodynamics and mathematics. First, we define what it is to be an agent (device) that could be the bearer of AI. Then we show that the mainstream definitions of `intelligence' proposed by Hutter and others and still accepted by the AI community are too weak (...) even to capture what is involved when we ascribe intelligence to an insect. We then summarise the highly useful definition of basic (arthropod) intelligence proposed by Rodney Brooks, and we identify the properties that an AI agent would need to possess in order to be the bearer of intelligence by this definition. Finally, we show that, from the perspective of the disciplines needed to create such an agent, namely mathematics and physics, these properties are realisable by neither implicit nor explicit mathematical design nor by setting up an environment in which an AI could evolve spontaneously. (shrink)

Zuse proposed that the universe is being computed by some sort of cellular automaton or other discrete computing machinery, challenging the long-held view that some physical laws are continuous by nature. Calculating Space is the title of MIT's English translation of Konrad Zuse's 1969 Rechnender Raum, the first work on digital physics. This is the LaTeX edition by A. German and H. Zenil based on the MIT's English translation with permission from the MIT and Konrad Zuse's son Horst Zuse. (...) Followed by an afterword by A. German and H. Zenil. (shrink)

BAIXE O PDF COMPLETO COM 58 FIGURAS E 50 PÁGINAS!!! -/- INTRODUÇÃO Conhecer a anatomia de um aparelho representa a base para seu estudo amplo e profundo, além de facilitar a compreensão de seu funcionamento. Por isso é importante o conhecimento da anatomia funcional do aparelho reprodutor, já que sua análise permite fixar sólidas bases para o estudo do fenômeno pelo qual se perpetuam as espécies: a reprodução. Nos subtemas compreendidos referentes ao macho e a fêmea, descreve-se cada um dos (...) órgãos que formam o aparelho reprodutor e as suas diferenças entre espécies, assinalando-se em cada um deles a sua função. Obviamente, a parte que trata do funcionamento não é abordada de uma forma muito ampla. Para facilitar o estudo do aparelho reprodutor do macho, seus órgãos podem ser classificados de acordo com sua localização anatômica, em órgãos genitais internos (que compreendem os ductos deferentes, as glândulas acessórias e a uretra pélvica) e órgãos genitais externos (testículos com seu epidídimo). O testículo também foi considerado como órgão sexual primário, enquanto os ductos excretores, glândulas acessórias, pênis e prepúcio são considerados órgãos sexuais secundários. O aparelho reprodutor da fêmea é dividido em ovários e órgãos tubulares (útero, oviduto, cérvix, vagina e vulva). 2.2 APARELHO REPRODUTOR DO MACHO Para avaliar a capacidade reprodutiva dos machos, é primordial conhecer a anatomia e a fisiologia do aparelho reprodutor. Os órgãos genitais incluem testículos, epidídimos, canais deferentes, ampolas deferentes, uretra, glândulas acessórias (próstata, glândulas seminais e glândulas bulbouretrais), pênis e prepúcio, que apresentam diferenças entre as espécies domésticas (tabela 1). Sabe-se que o sistema genital masculino compreende os órgãos envolvidos no desenvolvimento, no amadurecimento, no transporte e no armazenamento dos gametas masculinos (espermatozoides). Ele consiste em um par de testículos, o ducto contorcido do epidídimo, o ducto deferente, a uretra e as glândulas genitais acessórias. Os testículos produzem esperma e hormônios. O epidídimo armazena espermatozoides durante seu amadurecimento antes de passarem para o ducto deferente e pela uretra. As glândulas acessórias também liberam suas secreções na uretra e contribuem para o volume do sêmen. A parte distal da uretra forma o caminho combinado para a passagem tanto da urina como do sêmen. O pênis é o órgão copulador masculino e deposita sêmen no trato reprodutor feminino. As figuras 1-5 apresentam as diferenciações quanto ao sistema genital das espécies domésticas. Conhecer a anatomia dos órgãos reprodutivos é essencial para poder identificar as patologias que possam afetar a capacidade reprodutiva, bem como para entender as diferenças de manejo ao empregar algumas biotecnologias reprodutivas aplicadas aos animais domésticos. Ao longo do texto, a espécie bovina será utilizada como modelo e serão feitas as observações correspondentes às diferenças com as outras espécies. Os órgãos genitais da fêmea incluem os órgãos genitais internos (ovários, ovidutos, útero, cérvix, vestíbulo e vagina) e os órgãos genitais externos (lábios vulvares e clitóris) (tabela 3). Os ovários produzem tanto gametas femininos quanto hormônios. As tubas uterinas pares capturam os oócitos liberados pelos ovários e os transportam para o útero, onde o ovo fertilizado é mantido. A vagina serve como órgão copulatório e, juntamente com sua continuação, o vestíbulo, como canal de parto e passagem para excreção urinária. -/- REFERÊNCIAS BIBLIOGRÁFICAS BOETA, M. et al. Fisiología reproductiva de los animales domésticos. México: FMVZ – UNAM, 2018. DERIVAUX, Jules; BARNABÉ, Renato Campanarut. Reprodução dos animais domésticos. Zaragoza: Acribia, 1980. DUKES, Henry Hugh; SWENSON, Melvin J.; REECE, William O. Dukes fisiologia dos animais domésticos. Rio de Janeiro: Editora Guanabara Koogan, 1996. DYCE, Keith M.; WENSING, Cornelius JG; SACK, Wolfganf O. Tratado de anatomia veterinária. Elsevier Brasil, 2004. HAFEZ, Elsayed Saad Eldin; HAFEZ, B. Reprodução animal. São Paulo: Manole, 2004. HOLY, Lubos; MARTÍNEZ JÚSTIZ, G. Colab. Biología de la reproducción bovina. Havana: Revolucionária, 1975. KÖNIG, Horst Erich; LIEBICH, Hans-Georg. Anatomia dos Animais Domésticos-: Texto e Atlas Colorido. Porto Alegre: Artmed Editora, 2016. MIES FILHO, Antonio. Reprodução dos animais. Porto Alegre: Sulina, 1987. (shrink)

There are (at least) three approaches to quantifying information. The first, algorithmic information or Kolmogorov complexity, takes events as strings and, given a universal Turing machine, quantifies the information content of a string as the length of the shortest program producing it [1]. The second, Shannon information, takes events as belonging to ensembles and quantifies the information resulting from observing the given event in terms of the number of alternate events that have been ruled out [2]. The third, statistical learning (...) theory, has introduced measures of capacity that control (in part) the expected risk of classifiers [3]. These capacities quantify the expectations regarding future data that learning algorithms embed into classifiers. Solomonoff and Hutter have applied algorithmic information to prove remarkable results on universal induction. Shannon information provides the mathematical foundation for communication and coding theory. However, both approaches have shortcomings. Algorithmic information is not computable, severely limiting its practical usefulness. Shannon information refers to ensembles rather than actual events: it makes no sense to compute the Shannon information of a single string – or rather, there are many answers to this question depending on how a related ensemble is constructed. Although there are asymptotic results linking algorithmic and Shannon information, it is unsatisfying that there is such a large gap – a difference in kind – between the two measures. This note describes a new method of quantifying information, effective information, that links algorithmic information to Shannon information, and also links both to capacities arising in statistical learning theory [4, 5]. After introducing the measure, we show that it provides a non-universal analog of Kolmogorov complexity. We then apply it to derive basic capacities in statistical learning theory: empirical VC-entropy and empirical Rademacher complexity. A nice byproduct of our approach is an interpretation of the explanatory power of a learning algorithm in terms of the number of hypotheses it falsifies [6], counted in two different ways for the two capacities. We also discuss how effective information relates to information gain, Shannon and mutual information. (shrink)

The concept of supermatrix for social scientists was first introduced by Paul Horst. The main purpose of his book was to introduce this concept to social scientists, students, teachers and research workers who lacked mathematical training. This book introduces the concept of fuzzy super matrices and operations on them. The author has provided only those operations on fuzzy supermatrices that are essential for developing super fuzzy multi expert models. This book will be highly useful to social scientists who wish (...) to work with multi expert models. (shrink)

In this book, the authors introduce the notion of Super linear algebra and super vector spaces using the definition of super matrices defined by Horst (1963). Many theorems on super linear algebra and its properties are proved. Some theorems are left as exercises for the reader. These new class of super linear algebras which can be thought of as a set of linear algebras, following a stipulated condition, will find applications in several fields using computers. The authors feel that (...) such a paradigm shift is essential in this computerized world. (shrink)

Since the beginning of the 20th Century to the present day, it has rarely been doubted that whenever formal aesthetic methods meet their iconological counterparts, the two approaches appear to be mutually exclusive. In reality, though, an ahistorical concept is challenging a historical analysis of art. It is especially Susanne K. Langer´s long-overlooked system of analogies between perceptions of the world and of artistic creations that are dependent on feelings which today allows a rapprochement of these positions. Krois’s insistence on (...) a similar point supports this analysis. - I - Unbestritten bis heute gilt, formwissenschaftliche und ikonologische Methoden scheinen sich grundsätzlich auszuschließen, da die ersteren auf ahistorischen und die letzteren auf historischen Grundlagen aufbauen. Dem entgegen soll mit diesem Beitrag gezeigt werden, wie insbesondere die Forschungen Susanne K. Langers und ergänzend diejenigen von John M. Krois eine Annäherung beider Positionen ermöglichen. (shrink)