This article reviews a number of different areas in the foundations of formal learning theory. After outlining the general framework for formal models of learning, the Bayesian approach to learning is summarized. This leads to a discussion of Solomonoff's Universal Prior Distribution for Bayesian learning. Gold's model of identification in the limit is also outlined. We next discuss a number of aspects of learning theory raised in contributed papers, related to both (...) computational and representational complexity. The article concludes with a description of how semi-supervised learning can be applied to the study of cognitive learning models. Throughout this overview, the specific points raised by our contributing authors are connected to the models and methods under review. (shrink)

15년 전부터 등장한 베이지안 확률론적 추론 모형은 통계학, 과학철학, 심리학, 인지과학, 컴퓨터과학, 신경과학 등에서 학계의 연구를 지배하는 강력한 핵심논제가 된 후로 이젠 교육학과 심지어는 논리학 분야에까지 큰 반향을 일으키고 있다. 이러한 다양화로 인해, 굿에 따르면, 베이즈주의 유형은 46,656 가지로 분류된다. 이러한 다양한 모형들 중에서 베이지안 확률모형을 학습이론에 적용하려는 학자들은 학습자의 학습 과정이 정확히 베이지안 확률추론 과정을 항상 따른다고 주장한다. 그런데 이러한 확률론적 모형은 피아제의 구성주의적 전망을 계승한 것은 분명하다. 왜냐하면, 이 모델을 지지하는 학자들 자신이 자신들의 학적 운동을 “새로운 합리적 구성주의”라고 (...) 명시적으로 명명하기 때문이다. 그럼에도 불구하고, 이러한 확률론적 학습모형은 “지식은 인과성을 중심으로 구조화된다.”라는 가설을 확고하게 지지하면서, 피아제의 명제 논리적 모형에 의한 학습이론을 강하게 비판한다. 그러나 베이지안 학습 이론가들은 경험을 통한 새로운 항들의 연결로서의 학습, 그리고 학습에서 모순 및 인과성의 역할 등에 대한 피아제의 세부적인 연구 성과들을 놓치고 있다. 다른 한편으로 베이지안 학습이론은 여러 가지 점에서 온전한 발달 이론을 제공하기에는 아직 미흡하다. 이런 점에서, 필자는 현상 없이는 구조를 이해할 수 없듯이, 구조 없이는 현상을 이해할 수 없다는 사실을 강조하면서, 인과성 추론을 핵심으로 하는 베이지안 학습 모델과 인과적 물리적 인식뿐만 아니라, 동시에 논리-수학적 인식 모델을 강조하는 피아제의 학습 모델은 새로이 종합되어야 한다는 것을 논증하고자 한다. (shrink)

We investigate the general properties of general Bayesian learning, where “general Bayesian learning” means inferring a state from another that is regarded as evidence, and where the inference is conditionalizing the evidence using the conditional expectation determined by a reference probability measure representing the background subjective degrees of belief of a Bayesian Agent performing the inference. States are linear functionals that encode probability measures by assigning expectation values to random variables via integrating them with respect (...) to the probability measure. If a state can be learned from another this way, then it is said to be Bayes accessible from the evidence. It is shown that the Bayes accessibility relation is reflexive, antisymmetric and non-transitive. If every state is Bayes accessible from some other defined on the same set of random variables, then the set of states is called weakly Bayes connected. It is shown that the set of states is not weakly Bayes connected if the probability space is standard. The set of states is called weakly Bayes connectable if, given any state, the probability space can be extended in such a way that the given state becomes Bayes accessible from some other state in the extended space. It is shown that probability spaces are weakly Bayes connectable. Since conditioning using the theory of conditional expectations includes both Bayes’ rule and Jeffrey conditionalization as special cases, the results presented generalize substantially some results obtained earlier for Jeffrey conditionalization. (shrink)

We investigate the general properties of general Bayesian learning, where ``general Bayesian learning'' means inferring a state from another that is regarded as evidence, and where the inference is conditionalizing the evidence using the conditional expectation determined by a reference probability measure representing the background subjective degrees of belief of a Bayesian Agent performing the inference. States are linear functionals that encode probability measures by assigning expectation values to random variables via integrating them with respect (...) to the probability measure. If a state can be learned from another this way, then it is said to be Bayes accessible from the evidence. It is shown that the Bayes accessibility relation is reflexive, antisymmetric and non-transitive. If every state is Bayes accessible from some other defined on the same set of random variables, then the set of states is called weakly Bayes connected. It is shown that the set of states is not weakly Bayes connected if the probability space is standard. The set of states is called weakly Bayes connectable if, given any state, the probability space can be extended in such a way that the given state becomes Bayes accessible from some other state in the extended space. It is shown that probability spaces are weakly Bayes connectable. Since conditioning using the theory of conditional expectations includes both Bayes' rule and Jeffrey conditionalization as special cases, the results presented generalize substantially some results obtained earlier for Jeffrey conditionalization. (shrink)

It is usually taken as given that consciousness involves superior or more elaborate forms of information processing. Contemporary models equate consciousness with global processing, system complexity, or depth or stability of computation. This is in stark contrast with the powerful philosophical intuition that being conscious is more than just having the ability to compute. I argue that it is also incompatible with current empirical findings. I present a model that is free from the strong assumption that consciousness predicts superior performance. (...) The model is based on Bayesian decision theory, of which signal detection theory is a special case. It reflects the fact that the capacity for perceptual decisions is fundamentally limited by the presence and amount of noise in the system. To optimize performance, one therefore needs to set decision criteria that are based on the behaviour, i.e. the probability distributions, of the internal signals. One important realization is that the knowledge of how our internal signals behave statistically has to be learned over time. Essentially, we are doing statistics on our own brain. This ‘higherorder’ learning, however, may err, and this impairs our ability to set and maintain optimal criteria for perceptual decisions, which I argue is central to perception consciousness. I outline three possibilities of how conscious perception might be affected by failures of ‘higher-order’ representation. These all imply that one can have a dissociation between consciousness and performance. This model readily explains blindsight and hallucinations in formal terms, and is beginning to receive direct empirical support. I end by discussing some philosophical implications of the model. (shrink)

There is currently no viable alternative to the Bayesian analysis of scientific inference, yet the available versions of Bayesianism fail to do justice to several aspects of the testing and confirmation of scientific hypotheses. Bayes or Bust? provides the first balanced treatment of the complex set of issues involved in this nagging conundrum in the philosophy of science. Both Bayesians and anti-Bayesians will find a wealth of new insights on topics ranging from Bayes’s original paper to contemporary formal (...) class='Hi'>learning theory.In a paper published posthumously in 1763, the Reverend Thomas Bayes made a seminal contribution to the understanding of "analogical or inductive reasoning." Building on his insights, modem Bayesians have developed an account of scientific inference that has attracted numerous champions as well as numerous detractors. Earman argues that Bayesianism provides the best hope for a comprehensive and unified account of scientific inference, yet the presently available versions of Bayesianisin fail to do justice to several aspects of the testing and confirming of scientific theories and hypotheses. By focusing on the need for a resolution to this impasse, Earman sharpens the issues on which a resolution turns. John Earman is Professor of History and Philosophy of Science at the University of Pittsburgh. (shrink)

본 연구의 목적은 최근 20여 년 동안 진행되어 온 학습이론에 대한 피아제의 명제논리학적 학습이론과 베이즈주의의 확률론적 학습이론의 융합에 근거하는 새로운 융합교수학습모형을 개발하는 것이다. 연구자는 이 새로운 교수학습모델을 “베이지안 구조구성주의 교수학습모형”(Bayesian structure-constructivist Model of Teaching-learning: 이하 약칭 BMT)이라 명명한다. 본고는 역사-비판적 관점 및 형식화적 관점에서 피아제의 명제논리학적 학습모형에서 해석된 학습이론과 베이즈주의의 확률론적 추론모형에서 해석된 학습이론을 일차적으로 분석하고, 논문의 후반부에서는 이를 근거로 교수법의 관점에서 양자의 학습이론을 통합하는 새로운 교수학습모델, 즉 BMT의 중요한 특성들을 세부적으로 제시한다. 몇 가지 핵심만 언급하면, 첫째로, BMT는 (...) 개념 학습이든 이론 학습이든 모든 학습은 표상과 샘플링의 상호작용임을 강조한다. 매우 이른 나이에서부터 아동은 학습하기 이전에 자기에게 주어진 예시들이 약한 샘플링인지 강한 샘플링인지, 특히 교육학적 샘플링인지 비교육학적 샘플링인지를 구분한다. 둘째로, 따라서 학습을 촉진시킬 수 있는 샘플을 선택하기 위해서 교사는 교육학적 추론을 위한 베이지안 모델을 활용하면서 직접적인 강의식 수업과 발견학습 수업을 병행해야 한다. 셋째로, 아동은 때때로 구체적인 규칙들보다는 추상적인 규칙들을 먼저 학습하기도 하지만, 대부분의 영역에서 추상적인 지식의 학습은 구체적인 지식의 학습에 의존한다. 따라서 성공적인 학습을 위해서는 피아제가 강조한 보편적 언어로서 논리적 문법과 그들 사이의 위계성을 고려하는 샘플을 개발하고 제시해야 한다. (shrink)

Two key research issues in the field of causal learning are how people acquire causal knowledge when observing data that are presented sequentially, and the level of abstraction at which learning takes place. Does sequential causal learning solely involve the acquisition of specific cause-effect links, or do learners also acquire knowledge about abstract causal constraints? Recent empirical studies have revealed that experience with one set of causal cues can dramatically alter subsequent learning and performance with entirely (...) different cues, suggesting that learning involves abstract transfer, and such transfer effects involve sequential presentation of distinct sets of causal cues. It has been demonstrated that pre-training can modulate classic causal learning phenomena such as forward and backward blocking. To account for these effects, we propose a Bayesian theory of sequential causal learning. The theory assumes that humans are able to consider and use several alternative causal generative models, each instantiating a different causal integration rule. Model selection is used to decide which integration rule to use in a given learning environment in order to infer causal knowledge from sequential data. Detailed computer simulations demonstrate that humans rely on the abstract characteristics of outcome variables to select a causal integration rule, which in turn alters causal learning in a variety of blocking and overshadowing paradigms. When the nature of the outcome variable is ambiguous, humans select the model that yields the best fit with the recent environment, and then apply it to subsequent learning tasks. Based on sequential patterns of cue-outcome co-occurrence, the theory can account for a range of phenomena in sequential causal learning, including various blocking effects, primacy effects in some experimental conditions, and apparently abstract transfer of causal knowledge. (shrink)

본고는 딥러닝의 이미지 인식 원리와 유아의 이미지 인식 원리를 종합하면서, 이미지-개념 학습을 위한 새로운 교수학습모델, 즉 “베이지안 구조구성주의 교수학습모델”(Bayesian Structure-constructivist Teaching-learning Model: BSTM)을 제안한다. 달리 말하면, 기계학습 원리와 인간학습 원리를 비교함으로써 얻게 되는 시너지 효과를 바탕으로, 유아들의 이미지-개념 학습을 위한 새로운 교수 모델을 구성하는 것을 목표로 한다. 이런 맥락에서 본고는 전체적으로 3가지 차원에서 논의된다. 첫째, 아동의 이미지 학습에 대한 역사적 중요 이론인 “대상 전체론적 가설”, “분류학적 가설”, “배타적 가설”, “기본 수준 범주 가설” 등을 역사 비판적 관점에서 검토한다. 둘째, (...) 컴퓨터 공학에서 전개된 머신러닝의 이미지 학습 원리 및 그 구조의 역사적 변형들에 대해서 비판적으로 분석하면서, 특히 딥러닝의 교육 인식론적 의미를 도출한다. 셋째, 선행하는 논의들을 토대로 이미지와 개념을 연결시키는 학습, 즉 단어학습을 위한 필자의 고유한 이론인 구조구성주의 개념교수모델을 제안한다. 이때 필자는 피아제의 명제논리학적 학습 가설, 즉 논리-수학적 구조를 중시하는 기존의 피아제 학파의 학습 가설과 최근의 베이지안 학습 이론에서 강조된 확률론적 추론 가설, 특히 “교육학적 추론 가설”을 동시에 포함할 수 있는 교수학습모델의 구조 및 주요 지향점들을 제안한다. (shrink)

Recent studies suggest that learners who are asked to predict the outcome of an event learn more than learners who are asked to evaluate it retrospectively or not at all. One possible explanation for this “prediction boost” is that it helps learners engage metacognitive reasoning skills that may not be spontaneously leveraged, especially for individuals with still-developing executive functions. In this paper, we combined multiple analytic approaches to investigate the potential role of executive functions in elementary school-aged children's science (...) class='Hi'>learning. We performed an experiment that investigates children's science learning during a water displacement task where a “prediction boost” had previously been observed—children either made an explicit prediction or evaluated an event post hoc (i.e., postdiction). We then considered the relation of executive function measures and learning, which were collected following the main experiment. Via mixed effects regression models, we found that stronger executive function skills (i.e., stronger inhibition and switching scores) were associated with higher accuracy in Postdiction but not in the Prediction Condition. Using a theory-based Bayesian model, we simulated children's individual performance on the learning task (capturing “belief flexibility”), and compared this “flexibility” to the other measures to understand the relationship between belief revision, executive function, and prediction. Children in the Prediction Condition showed near-ceiling “belief flexibility” scores, which were significantly higher than among children in the Postdiction Condition. We also found a significant correlation between children's executive function measures to our “belief flexibility” parameter, but only for children in the Postdiction Condition. These results indicate that when children provided responses post hoc, they may have required stronger executive function capacities to navigate the learning task. Additionally, these results suggest that the “prediction boost” in children's science learning could be explained by increased metacognitive flexibility in the belief revision process. (shrink)

Bayesian learning theory and evolutionary theory both formalize adaptive competition dynamics in possibly high‐dimensional, varying, and noisy environments. What do they have in common and how do they differ? In this paper, we discuss structural and dynamical analogies and their limits, both at a computational and an algorithmic‐mechanical level. We point out mathematical equivalences between their basic dynamical equations, generalizing the isomorphism between Bayesian update and replicator dynamics. We discuss how these mechanisms provide analogous answers (...) to the challenge of adapting to stochastically changing environments at multiple timescales. We elucidate an algorithmic equivalence between a sampling approximation, particle filters, and the Wright‐Fisher model of population genetics. These equivalences suggest that the frequency distribution of types in replicator populations optimally encodes regularities of a stochastic environment to predict future environments, without invoking the known mechanisms of multilevel selection and evolvability. A unified view of the theories of learning and evolution comes in sight. (shrink)

Conditional learning, where agents learn a conditional sentence ‘If A, then B,’ is difficult to incorporate into existing Bayesian models of learning. This is because conditional learning is not uniform: in some cases, learning a conditional requires decreasing the probability of the antecedent, while in other cases, the antecedent probability stays constant or increases. I argue that how one learns a conditional depends on the causal structure relating the antecedent and the consequent, leading to a (...) causal model of conditional learning. This model extends traditional Bayesian learning by incorporating causal models into agents’ epistemic states. On this theory, conditional learning proceeds in two steps. First, an agent learns a new causal model with the appropriate relationship between the antecedent and the consequent. Then, the agent narrows down the set of possible worlds to include only those which make the conditional proposition true. This model of learning can incorporate both standard cases of Bayesian learning and the non-uniform learning required to learn conditional information. (shrink)

Hierarchical Bayesian models (HBMs) provide an account of Bayesian inference in a hierarchically structured hypothesis space. Scientific theories are plausibly regarded as organized into hierarchies in many cases, with higher levels sometimes called ‘paradigms’ and lower levels encoding more specific or concrete hypotheses. Therefore, HBMs provide a useful model for scientific theory change, showing how higher‐level theory change may be driven by the impact of evidence on lower levels. HBMs capture features described in the Kuhnian tradition, (...) particularly the idea that higher‐level theories guide learning at lower levels. In addition, they help resolve certain issues for Bayesians, such as scientific preference for simplicity and the problem of new theories. *Received July 2009; revised October 2009. †To contact the authors, please write to: Leah Henderson, Massachusetts Institute of Technology, 77 Massachusetts Avenue, 32D‐808, Cambridge, MA 02139; e‐mail: [email protected]. (shrink)

We design a double-or-quits game to compare the speed of learning one’s specific ability with the speed of rising confidence as the task gets increasingly difficult. We find that people on average learn to be overconfident faster than they learn their true ability and we present an intuitive-Bayesian model of confidence which integrates confidence biases and learning. Uncertainty about one’s true ability to perform a task in isolation can be responsible for large and stable confidence biases, namely (...) limited discrimination, the hard–easy effect, the Dunning–Kruger effect, conservative learning from experience and the overprecision phenomenon if subjects act as Bayesian learners who rely only on sequentially perceived performance cues and contrarian illusory signals induced by doubt. Moreover, these biases are likely to persist since the Bayesian aggregation of past information consolidates the accumulation of errors and the perception of contrarian illusory signals generates conservatism and under-reaction to events. Taken together, these two features may explain why intuitive Bayesians make systematically wrong predictions of their own performance. (shrink)

Modularity theorists have challenged that there are, or could be, general learning mechanisms that explain theory-of-mind development. In response, supporters of the ‘scientific theory-theory’ account of theory-of-mind development have appealed to children's use of auxiliary hypotheses and probabilistic causal modeling. This article argues that these general learning mechanisms are not sufficient to meet the modularist's challenge. The article then explores an alternative domain-general learning mechanism by proposing that children grasp the concept belief through (...) the progressive alignment of relational structure that occurs as a result of structural-comparison. The article also explores the implications of the proposed account for Fodor's puzzle of conceptual learning. (shrink)

Complex constraints like conditionals ('If A, then B') and probabilistic constraints ('The probability that A is p') pose problems for Bayesian theories of learning. Since these propositions do not express constraints on outcomes, agents cannot simply conditionalize on the new information. Furthermore, a natural extension of conditionalization, relative information minimization, leads to many counterintuitive predictions, evidenced by the sundowners problem and the Judy Benjamin problem. Building on the notion of a `paradigm shift' and empirical research in psychology and (...) economics, I argue that the model of hypothesis testing can explain how people learn complex, theory-laden propositions like conditionals and probability constraints. Theories are formalized as probability distributions over a set of possible outcomes and theory change is triggered by a constraint which is incompatible with the initial theory. This leads agents to consult a higher order probability function, or a 'prior over priors,' to choose the most likely alternative theory which satisfies the constraint. The hypothesis testing model is applied to three examples: learning a simple probabilistic constraint involving coin bias, the sundowners problem for conditional learning, and the Judy Benjamin problem for learning conditional probability constraints. (shrink)

Children acquire complex concepts like DOG earlier than simple concepts like BROWN, even though our best neuroscientific theories suggest that learning the former is harder than learning the latter and, thus, should take more time (Werning 2010). This is the Complex- First Paradox. We present a novel solution to the Complex-First Paradox. Our solution builds on a generalization of Xu and Tenenbaum’s (2007) Bayesian model of word learning. By focusing on a rational theory of concept (...) learning, we show that it is easier to infer the meaning of complex concepts than that of simple concepts. (shrink)

According to the Bayesian paradigm in the psychology of reasoning, the norms by which everyday human cognition is best evaluated are probabilistic rather than logical in character. Recently, the Bayesian paradigm has been applied to the domain of argumentation, where the fundamental norms are traditionally assumed to be logical. Here, we present a major generalisation of extant Bayesian approaches to argumentation that (i)utilizes a new class of Bayesian learning methods that are better suited to modelling (...) dynamic and conditional inferences than standard Bayesian conditionalization, (ii) is able to characterise the special value of logically valid argument schemes in uncertain reasoning contexts, (iii) greatly extends the range of inferences and argumentative phenomena that can be adequately described in a Bayesian framework, and (iv) undermines some influential theoretical motivations for dual function models of human cognition. We conclude that the probabilistic norms given by the Bayesian approach to rationality are not necessarily at odds with the norms given by classical logic. Rather, the Bayesian theory of argumentation can be seen as justifying and enriching the argumentative norms of classical logic. (shrink)

Whether pattern-parsing mechanisms are specific to language or apply across multiple cognitive domains remains unresolved. Formal language theory provides a mathematical framework for classifying pattern-generating rule sets (or “grammars”) according to complexity. This framework applies to patterns at any level of complexity, stretching from simple sequences, to highly complex tree-like or net-like structures, to any Turing-computable set of strings. Here, we explored human pattern-processing capabilities in the visual domain by generating abstract visual sequences made up of abstract tiles differing (...) in form and color. We constructed different sets of sequences, using artificial “grammars” (rule sets) at three key complexity levels. Because human linguistic syntax is classed as “mildly context-sensitive,” we specifically included a visual grammar at this complexity level. Acquisition of these three grammars was tested in an artificial grammar-learning paradigm: after exposure to a set of well-formed strings, participants were asked to discriminate novel grammatical patterns from non-grammatical patterns. Participants successfully acquired all three grammars after only minutes of exposure, correctly generalizing to novel stimuli and to novel stimulus lengths. A Bayesian analysis excluded multiple alternative hypotheses and shows that the success in rule acquisition applies both at the group level and for most participants analyzed individually. These experimental results demonstrate rapid pattern learning for abstract visual patterns, extending to the mildly context-sensitive level characterizing language. We suggest that a formal equivalence of processing at the mildly context sensitive level in the visual and linguistic domains implies that cognitive mechanisms with the computational power to process linguistic syntax are not specific to the domain of language, but extend to abstract visual patterns with no meaning. (shrink)

My dissertation focuses on problems that arise when a group makes decisions that are in reasonable ways connected to the beliefs and values of the group members. These situations are represented by models of decision-theoretic aggregation: Suppose a model of individual rationality in decision-making applies to each of a group of agents. Suppose this model also applies to the group as a whole, and that this group model is aggregated from the individual models. Two questions arise. First, what sets of (...) reasonable conditions can we consistently impose on the aggregation? Second, what can we learn from insights gained in such models, in particular for ethics and political philosophy? The dissertation has three parts, each using a different model of individual rationality. Each part contains essays addressing questions of either the first or the second kind with regard to its underlying model. Part I uses a social-choice-theoretic model, in which agents are only assumed to rank alternatives. Arrow's "Impossibility Theorem" shows that certain conditions that it would be natural to impose on all aggregation cannot be so imposed, on pain of contradiction. I discuss the impact of this result on the evaluation of majority rule and on democratic theory. I also discuss Sen's "Liberal Paradox", which addresses the possibility of incorporating rights into a social-choice theoretic model. Part II uses the von Neumann/Morgenstern theory of expected utility. A result under this model has been regarded as providing an argument for utilitarianism. I locate a place within utilitarianism where the theorem can do some work. To prepare the ground, I interpret the von Neumann/Morgenstern theory as part of a theory of deliberation. Thereby, this version of expected utility theory can be embedded into the philosophy of practical reasoning. The underlying model in Part III is Bayesian decision theory. Whereas the other two inquiries focus on the philosophical assessment of formal results, the third makes a formal contribution. One essay surveys the area, and a second explores a curious phenomenon that occurs in determining outcomes under one type of Bayesian aggregation. (shrink)

According to certain normative theories in epistemology, rationality requires us to be logically omniscient. Yet this prescription clashes with our ordinary judgments of rationality. How should we resolve this tension? In this paper, I focus particularly on the logical omniscience requirement in Bayesian epistemology. Building on a key insight by Hacking :311–325, 1967), I develop a version of Bayesianism that permits logical ignorance. This includes: an account of the synchronic norms that govern a logically ignorant individual at any given (...) time; an account of how we reduce our logical ignorance by learning logical facts and how we should update our credences in response to such evidence; and an account of when logical ignorance is irrational and when it isn’t. At the end, I explain why the requirement of logical omniscience remains true of ideal agents with no computational, processing, or storage limitations. (shrink)

When making decisions, people naturally face uncertainty about the potential consequences of their actions due in part to limits in their capacity to represent, evaluate or deliberate. Nonetheless, they aim to make the best decisions possible. In Decision Theory with a Human Face, Richard Bradley develops new theories of agency and rational decision-making, offering guidance on how 'real' agents who are aware of their bounds should represent the uncertainty they face, how they should revise their opinions as a result (...) of experience and how they should make decisions when lacking full awareness of, or precise opinions on relevant contingencies. He engages with the strengths and flaws of Bayesian reasoning, and presents clear and comprehensive explorations of key issues in decision theory, from belief and desire to semantics and learning. His book draws on philosophy, economics, decision science and psychology, and will appeal to readers in all of these disciplines. (shrink)

Nowadays, the development of human society and daily life are inseparable from the power supply. Therefore, people also put forward higher requirements for the reliability of distribution network, but power companies can only passively deal with distribution network failures, which is a bottleneck for the improvement of distribution network reliability. The Internet of Things is the best solution for online equipment status monitoring and basic data sharing for large, widely distributed, relatively fixed, and large numbers of equipment. The construction of (...) the IoT for power distribution equipment faces many important problems, including the selection of networking, equipment selection, and interaction standards. When researching the implementation plan, research on the distribution of IoT market was carried out. Based on the grid, the idea of optimizing the investment selection plan of the power distribution using IoT was discussed, and a result verification model was established. After the completion of the theoretical part, a case study of medium-voltage distribution grid equipment management and medium-voltage distribution network equipment management based on the grid was carried out by taking a real enterprise application situation as an example. Realizing fault diagnosis of distribution network will not only provide decision support for operation and maintenance of distribution network for power companies, but also have great economic and social benefits. Aiming at the shortcomings of single data mining method in distribution network fault diagnosis, hybrid data mining method is proposed. First, rough set theory is used to reduce the original fault data and form a simplified rule set. Because of the non-linearity of distribution network fault and the strong learning ability, adaptability, and robustness of Bayesian network, Bayesian network can be used to classify distribution network faults. Therefore, a simplified fault diagnosis system is established in this paper, and its correctness is confirmed. Then, the learning and training are carried out by using Bayesian network to call the simplest rule set, which has the characteristics of short learning and training time and high diagnostic accuracy. (shrink)

According to classical arguments, language learning is both facilitated and constrained by cognitive biases. These biases are reflected in linguistic typology—the distribution of linguistic patterns across the world's languages—and can be probed with artificial grammar experiments on child and adult learners. Beginning with a widely successful approach to typology (Optimality Theory), and adapting techniques from computational approaches to statistical learning, we develop a Bayesian model of cognitive biases and show that it accounts for the detailed pattern (...) of results of artificial grammar experiments on noun-phrase word order (Culbertson, Smolensky, & Legendre, 2012). Our proposal has several novel properties that distinguish it from prior work in the domains of linguistic theory, computational cognitive science, and machine learning. This study illustrates how ideas from these domains can be synthesized into a model of language learning in which biases range in strength from hard (absolute) to soft (statistical), and in which language-specific and domain-general biases combine to account for data from the macro-level scale of typological distribution to the micro-level scale of learning by individuals. (shrink)

This work provides a conceptual foundation for a Bayesian approach to artificial inference and learning. I argue that Bayesian confirmation theory provides a general normative theory of inductive learning and therefore should have a role in any artificially intelligent system that is to learn inductively about its world. I modify the usual Bayesian theory in three ways directly pertinent to an eventual research program in artificial intelligence. First, I construe Bayesian inference (...) rules as defeasible, allowing them to be overridden in certain contexts and therefore allowing them to play a part in a hybrid system, coexisting with non-Bayesian modes of inference. I take seriously the need to find meaningful prior probabilities for hypotheses, and elaborate means for supplying an artificial intelligence with such. And I address the computational complexity of Bayesian inference by reference to simplifications using causal networks and by allowing the probabilistic acceptance of hypotheses and subsequent qualitative inference to supplement Bayesian reasoning. The result is a Pragmatic Bayesian model of induction. (shrink)

According to Bayesian epistemology, rational learning from experience is consistent learning, that is learning should incorporate new information consistently into one's old system of beliefs. Simon M. Huttegger argues that this core idea can be transferred to situations where the learner's informational inputs are much more limited than Bayesianism assumes, thereby significantly expanding the reach of a Bayesian type of epistemology. What results from this is a unified account of probabilistic learning in the tradition (...) of Richard Jeffrey's 'radical probabilism'. Along the way, Huttegger addresses a number of debates in epistemology and the philosophy of science, including the status of prior probabilities, whether Bayes' rule is the only legitimate form of learning from experience, and whether rational agents can have sustained disagreements. His book will be of interest to students and scholars of epistemology, of game and decision theory, and of cognitive, economic, and computer sciences. (shrink)

Shepard has argued that a universal law should govern generalization across different domains of perception and cognition, as well as across organisms from different species or even different planets. Starting with some basic assumptions about natural kinds, he derived an exponential decay function as the form of the universal generalization gradient, which accords strikingly well with a wide range of empirical data. However, his original formulation applied only to the ideal case of generalization from a single encountered stimulus to a (...) single novel stimulus, and for stimuli that can be represented as points in a continuous metric psychological space. Here we recast Shepard's theory in a more general Bayesian framework and show how this naturally extends his approach to the more realistic situation of generalizing from multiple consequential stimuli with arbitrary representational structure. Our framework also subsumes a version of Tversky's set-theoretic model of similarity, which is conventionally thought of as the primary alternative to Shepard's continuous metric space model of similarity and generalization. This unification allows us not only to draw deep parallels between the set-theoretic and spatial approaches, but also to significantly advance the explanatory power of set-theoretic models. Key Words: additive clustering; Bayesian inference; categorization; concept learning; contrast model; features; generalization; psychological space; similarity. (shrink)

We evaluate the asymptotic performance of boundedly-rational strategies in multi-armed bandit problems, where performance is measured in terms of the tendency (in the limit) to play optimal actions in either (i) isolation or (ii) networks of other learners. We show that, for many strategies commonly employed in economics, psychology, and machine learning, performance in isolation and performance in networks are essentially unrelated. Our results suggest that the appropriateness of various, common boundedly-rational strategies depends crucially upon the social context (if (...) any) in which such strategies are to be employed. (shrink)

Event segmentation theory posits that people segment continuous experience into discrete events and that event boundaries occur when there are large transient increases in prediction error. Here, we set out to test this theory in the context of story listening, by using a deep learning language model (GPT‐2) to compute the predicted probability distribution of the next word, at each point in the story. For three stories, we used the probability distributions generated by GPT‐2 to compute the (...) time series of prediction error. We also asked participants to listen to these stories while marking event boundaries. We used regression models to relate the GPT‐2 measures to the human segmentation data. We found that event boundaries are associated with transient increases in Bayesian surprise but not with a simpler measure of prediction error (surprisal) that tracks, for each word in the story, how strongly that word was predicted at the previous time point. These results support the hypothesis that prediction error serves as a control mechanism governing event segmentation and point to important differences between operational definitions of prediction error. (shrink)

Collaboration requires agents to coordinate their behavior on the fly, sometimes cooperating to solve a single task together and other times dividing it up into sub‐tasks to work on in parallel. Underlying the human ability to collaborate is theory‐of‐mind (ToM), the ability to infer the hidden mental states that drive others to act. Here, we develop Bayesian Delegation, a decentralized multi‐agent learning mechanism with these abilities. Bayesian Delegation enables agents to rapidly infer the hidden intentions of (...) others by inverse planning. We test Bayesian Delegation in a suite of multi‐agent Markov decision processes inspired by cooking problems. On these tasks, agents with Bayesian Delegation coordinate both their high‐level plans (e.g., what sub‐task they should work on) and their low‐level actions (e.g., avoiding getting in each other's way). When matched with partners that act using the same algorithm, Bayesian Delegation outperforms alternatives. Bayesian Delegation is also a capable ad hoc collaborator and successfully coordinates with other agent types even in the absence of prior experience. Finally, in a behavioral experiment, we show that Bayesian Delegation makes inferences similar to human observers about the intent of others. Together, these results argue for the centrality of ToM for successful decentralized multi‐agent collaboration. (shrink)

How should we revise our beliefs in response to the expressed probabilistic opinions of experts on some proposition when these experts are in disagreement? In this paper I examine the suggestion that in such circumstances we should adopt a linear average of the experts’ opinions and consider whether such a belief revision policy is compatible with Bayesian conditionalisation. By looking at situations in which full or partial deference to the expressed opinions of others is required by Bayesianism I show (...) that only in trivial circumstances are the requirements imposed by linear averaging compatible with it. (shrink)

Constructing an intuitive theory from data confronts learners with a “chicken‐and‐egg” problem: The laws can only be expressed in terms of the theory's core concepts, but these concepts are only meaningful in terms of the role they play in the theory's laws; how can a learner discover appropriate concepts and laws simultaneously, knowing neither to begin with? We explore how children can solve this chicken‐and‐egg problem in the domain of magnetism, drawing on perspectives from computational modeling and (...) behavioral experiments. We present 4‐ and 5‐year‐olds with two different simplified magnet‐learning tasks. Children appropriately constrain their beliefs to two hypotheses following ambiguous but informative evidence. Following a critical intervention, they learn the correct theory. In the second study, children infer the correct number of categories given no information about the possible causal laws. Children's hypotheses in these tasks are explained as rational inferences within a Bayesian computational framework. (shrink)

The predictive processing literature contains at least two different versions of the framework with different theoretical resources at their disposal. One version appeals to so-called optimistic priors to explain agents’ motivation to act (call this optimistic predictive processing). A more recent version appeals to expected free energy minimization to explain how agents can decide between different action policies (call this preference predictive processing). The difference between the two versions has not been properly appreciated, and they are not sufficiently separated in (...) the literature. They constitute two different theories with strikingly different accounts of motivation and action. By reducing all desire-like constructs to belief-like constructs, optimistic predictive processing entails a substantial revision of standard accounts of motivation and action in philosophy and cognitive science. By contrast, preference predictive processing introduces desire-like constructs that play Humean motivational roles in the explanation of action. In this Humean iteration, predictive processing resembles other prominent computational frameworks implementing a distinction between beliefs and desires, such as reinforcement learning and Bayesian decision theory. Ultimately, predictive processing faces a dilemma between parsimony of mental constructs and completeness of its explanations of agency and the mind. (shrink)

This paper is dedicated to the analysis of the maximin principle, which is one of the key theoretical concepts of John Rawls’s theory of justice, and the problem that this principle creates for any attempt to provide a naturalistic interpretation of Rawls’s concept of fairness . Analysis shows that maximin principle is, in fact, incompatible with the Bayesian decision theory. This paper is intended to show that recent breakthroughs in evolutionary game theory could help to reconcile (...) the maximin principle with a certain naturally emerging and evolutionary stable pattern of human social behaviour.Keywords: Maximin principle, Nash bargaining game, Evolutionary Game Theory, Evolutionary Stable Strategy, learning rules. (shrink)

The coherentist theory of justification provides a response to the sceptical challenge: even though the independent processes by which we gather information about the world may be of dubious quality, the internal coherence of the information provides the justification for our empirical beliefs. This central canon of the coherence theory of justification is tested within the framework of Bayesian networks, which is a theory of probabilistic reasoning in artificial intelligence. We interpret the independence of the information (...) gathering processes (IGPs) in terms of conditional independences, construct a minimal sufficient condition for a coherence ranking of information sets and assess whether the confidence boost that results from receiving information through independent IGPs is indeed a positive function of the coherence of the information set. There are multiple interpretations of what constitute IGPs of dubious quality. Do we know our IGPs to be no better than randomization processes? Or, do we know them to be better than randomization processes but not quite fully reliable, and if so, what is the nature of this lack of full reliability? Or, do we not know whether they are fully reliable or not? Within the latter interpretation, does learning something about the quality of some IGPs teach us anything about the quality of the other IGPs? The Bayesian-network models demonstrate that the success of the coherentist canon is contingent on what interpretation one endorses of the claim that our IGPs are of dubious quality. (shrink)

This paper explores how the Bayesian program benefits from allowing for objective chance as well as subjective degree of belief. It applies David Lewis’s Principal Principle and David Christensen’s principle of informed preference to defend Howard Raiffa’s appeal to preferences between reference lotteries and scaling lotteries to represent degrees of belief. It goes on to outline the role of objective lotteries in an application of rationality axioms equivalent to the existence of a utility assignment to represent preferences in Savage’s (...) famous omelet example of a rational choice problem. An example motivating causal decision theory illustrates the need for representing subjunctive dependencies to do justice to intuitive examples where epistemic and causal independence come apart. We argue to extend Lewis’s account of chance as a guide to epistemic probability to include De Finetti’s convergence results. We explore Diachronic Dutch book arguments as illustrating commitments for treating transitions as learning experiences. Finally, we explore implications for Martingale convergence results for motivating commitment to objective chances. (shrink)

How do teachers learn about what learners already know? How do learners aid teachers by providing them with information about their background knowledge and what they find confusing? We formalize this collaborative reasoning process using a hierarchical Bayesian model of pedagogy. We then evaluate this model in two online behavioral experiments (N = 312 adults). In Experiment 1, we show that teachers select examples that account for learners' background knowledge, and adjust their examples based on learners' feedback. In Experiment (...) 2, we show that learners strategically provide more feedback when teachers' examples deviate from their background knowledge. These findings provide a foundation for extending computational accounts of pedagogy to richer interactive settings. (shrink)

People draw subtle distinctions in the normative domain. But it remains unclear exactly what gives rise to such distinctions. On one prominent approach, emotion systems trigger non-utilitarian judgments. The main alternative, inspired by Chomskyan linguistics, suggests that moral distinctions derive from an innate moral grammar. In this article, we draw on Bayesian learning theory to develop a rational learning account. We argue that the ‘size principle’, which is implicated in word learning, can also explain how (...) children would use scant and equivocal evidence to interpret candidate rules as applying more narrowly than utilitarian rules. (shrink)

My research program proposes that children have representations and learning mechanisms that can be characterized as causal models of the world Bayesian Fundamentalism.”.

The Bayesian model is used in psychology as the reference for the study of dynamic probability judgment. The main limit induced by this model is that it confines the study of revision of degrees of belief to the sole situations of revision in which the universe is static (revising situations). However, it may happen that individuals have to revise their degrees of belief when the message they learn specifies a change of direction in the universe, which is considered as (...) changing with time (updating situations). We analyze the main results of the experimental literature with regard to elementary qualitative properties of these two situations of revision. First, the order effect phenomenon is confronted with the commutative property. Second, an apparent new phenomenon is presented: the redundancy effect that is confronted with the idempotence property. Finally, results obtained in this kind of experimental situations are reinterpreted in the light of pragmatic analysis. (shrink)

Recent decades have witnessed a sea change in thinking about emotion, which has gone from being seen as a disruptive force in human thought and action to being seen as an important source of situation- and goal-relevant information and evaluation, continuous with perception and cognition. Here I argue on philosophical and empirical grounds that the role of emotion in contributing to our ability to respond to reasons for action runs deeper still: The affective system is at the core of the (...) process of evaluatively modeling situations, actions, and outcomes, which is the foundation upon which rational deliberation and action can be built. Taking up this perspective affords new approaches to long-standing problems in the theory of reason-based action. (shrink)

Zuidema et al. illustrate how empirical AGL studies can benefit from computational models and techniques. Computational models can help clarifying theories, and thus in delineating research questions, but also in facilitating experimental design, stimulus generation, and data analysis. The authors show, with a series of examples, how computational modeling can be integrated with empirical AGL approaches, and how model selection techniques can indicate the most likely model to explain experimental outcomes.

본 연구는 “교육학적 추론을 위한 베이지언 모델”의 형식적 타당성 및 이 모델이 갖는 교육학적 함의와 인식론적 함의에 대해 비판적으로 검토한다.BR 베이즈주의 학습이론가들에 따르면, 교육학적 목표를 가장 잘 성취하기 위해서는 “정확한 가설”(h)에 대한 학습자의 믿음을 최대화하는 “데이터”(d)를 교사가 선택해야 한다. 달리 말하면, 학생이 추측하는 문제의 가설(개념)이 교사가 목표로 하는 바로 그 가설(개념)에 최대로 가까워지게 하는 예시를 교사가 학생에게 제공해야 한다. 이를 위해서는 교사가 생산하는 “데이터의 분포”(p teacher (d｜h))가 “가설(h)에 대한 학습자의 사후 믿음”(p learner (h｜d))을 최대화하는 데이터들을 중심으로 균등하게 분포되어야 할 것이다. (...) 학생은 처음부터 문제의 개념을 소유하고 있지 않다. 학생은 교사가 제공하는 예시를 통해서 비로소 개념을 구성해야 하는 위치에 있다. 이를 위해서 우선 교사는 학생의 인지 상태를 알고 있어야 한다. 즉 학생이 어떤 데이터를 접하게 될 때, 어떤 개념을 상상할 것이라고 교사는 예측해야 한다. 결국 교사는 “p teacher (d｜h)” 와 “p learner (h｜d)”, 양자 모두를 알아야 한다. 이러한 맥락에서 교사가 활용할 수 있는 모델이 아래와 같은 방정식이다.BR 이것이 베이지언 학습이론가들이 지지하는 “교육학적 추론 모델”이다. 그들에 따르면, 이 방정식의 해는 문제의 개념(가설)을 가장 잘 전달하게 할 예시 선택을 가능하게 하는 대안이다. 그런데 이 모델이 구체적 교육 현장에서 어떻게 적용될 수 있는가를 보기 위해서, 또한 이 모델이 가지는 논리-형식적 타당성 및 교육-인식론적 함의를 이해하기 위해서, 필자는 베이지언 이론가들이 제시하는 많은 예시들 중에서 가장 단순한 예시인 “선 게임 이론”(theory of line game)을 분석하고자 한다.BR. (shrink)

This chapter explores a way in which visual processing may involve innate constraints and attempts to show how such processing overcomes one enduring challenge to nativism. In particular, many challenges to nativist theories in other areas of cognitive psychology have focused on the later development of such abilities, and have argued that such development is in conflict with innate origins. Innateness, in these contexts, is seen as antidevelopmental, associated instead with static processes and principles. In contrast, certain perceptual models demonstrate (...) how the very same mental processes can both be innately specified and yet develop richly in response to experience with the environment. This process is entirely unmysterious, as shown in certain formal theories of visual perception, including those that appeal to spontaneous endogenous stimulation and those based on Bayesian inference. (shrink)

The relationship between machine learning and the philosophy of science can be classed as a dynamic interaction: a mutually beneficial connection between two autonomous fields that changes direction over time. I discuss the nature of this interaction and give a case study highlighting interactions between research on Bayesian networks in machine learning and research on causality and probability in the philosophy of science.