Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Probability Learning”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 541 records · Page 30Linked to original sources
PubMed2004

A new criterion using information gain for action selection strategy in reinforcement learning.

In this paper, we regard the sequence of returns as outputs from a parametric compound source. Utilizing the fact that the coding rate of the source shows the amount of information about the return, we describe l-learning algorithms based on the predictive coding idea for estimating an expected information gain concerning future information and give a convergence proof of the information gain. Using the information gain, we propose the ratio w of return loss to information gain as a new criterion to be used in probabilistic action-selection strategies. In experimental results, we found that our w-based strategy performs well compared with the conventional Q-based strategy.

Algorithms↗
PubMed1989

On the status of nonconscious processes in human cognition: comment on Reber.

Two issues are discussed. The first one pertains to the generality of the nonconscious learning processes and their somewhat paradoxical status in cognitive psychology. We argue that the ability of the human cognitive system to nonconsciously acquire complex knowledge structures is one of its elementary and indispensable properties. Moreover, the existence of this ability constitutes one of the necessary metatheoretical assumptions of contemporary cognitive psychology. Nevertheless, the contemporary cognitive psychology literature often implies that it is only one of many controversial and unusual phenomena. The second issue pertains to the distinction between the so-called primitive unconscious and the sophisticated unconscious as proposed by Reber (1989).

Concept Formation↗
PubMed2004

Multi-layered greedy network-growing algorithm: extension of greedy network-growing algorithm to multi-layered networks.

In this paper, we extend our greedy network-growing algorithm to multi-layered networks. With multi-layered networks, we can solve many complex problems that single-layered networks fail to solve. In addition, the network-growing algorithm is used in conjunction with teacher-directed learning that produces appropriate outputs without computing errors between targets and outputs. Thus, the present algorithm is a very efficient network-growing algorithm. The new algorithm was applied to three problems: the famous vertical-horizontal lines detection problem, a medical data problem and a road classification problem. In all these cases, experimental results confirmed that the method could solve problems that single-layered networks failed to. In addition, information maximization makes it possible to extract salient features in input patterns.

Algorithms↗
PubMed2006

Variation of bar-press duration: where do new responses come from?

Instrumental learning involves both variation and selection: variation of what the animal does, and selection by reward from among the variation. Four experiments with rats suggested a rule about how variation is controlled by recent events. Experiment 1 used the peak procedure. Measurements of bar-press durations showed a sharp increase in mean duration after the time that food was sometimes given. The increase was triggered by the omission of expected food. Our first explanation of the increase was that it was a frustration effect. Experiment 2 tested this explanation with a procedure in which the first response of a trial usually produced food, ending the trial. In Experiment 2, unlike Experiment 1, omission of expected food did not produce a large increase in bar-press duration, which cast doubt on the frustration explanation. Experiments 3 and 4 tested an alternative explanation: a decrease in expectation of reward increases variation. Both used two signals associated with different probabilities of reward. Bar presses were more variable in duration during the signal with the lower probability of reward, supporting this alternative. These experiments show how variation can be studied with ordinary equipment and responses.

Animals↗
PubMedDate not supplied

Applications of the self-organising map to reinforcement learning.

This article is concerned with the representation and generalisation of continuous action spaces in reinforcement learning (RL) problems. A model is proposed based on the self-organising map (SOM) of Kohonen [Self Organisation and Associative Memory, 1987] which allows either the one-to-one, many-to-one or one-to-many structure of the desired state-action mapping to be captured. Although presented here for tasks involving immediate reward, the approach is easily extended to delayed reward. We conclude that the SOM is a useful tool for providing real-time, on-line generalisation in RL problems in which the latent dimensionalities of the state and action spaces are small. Scalability issues are also discussed.

Neural Networks, Computer↗
PubMed2006

A mechanistic account of striatal dopamine function in human cognition: psychopharmacological studies with cabergoline and haloperidol.

The authors test a neurocomputational model of dopamine function in cognition by administering to healthy participants low doses of D2 agents cabergoline and haloperidol. The model suggests that DA dynamically modulates the balance of Go and No-Go basal ganglia pathways during cognitive learning and performance. Cabergoline impaired, while haloperidol enhanced, Go learning from positive reinforcement, consistent with presynaptic drug effects. Cabergoline also caused an overall bias toward Go responding, consistent with postsynaptic action. These same effects extended to working memory and attentional domains, supporting the idea that the basal ganglia/dopamine system modulates the updating of prefrontal representations. Drug effects interacted with baseline working memory span in all tasks. Taken together, the results support a unified account of the role of dopamine in modulating cognitive processes that depend on the basal ganglia.

Adolescent↗
PubMed2000

Unsupervised learning of binary vectors: a Gaussian scenario.

We study a model of unsupervised learning where the real-valued data vectors are isotropically distributed, except for a single symmetry-breaking binary direction Bin¿-1,+1¿(N), onto which the projections have a Gaussian distribution. We show that a candidate vector J undergoing Gibbs learning in this discrete space, approaches the perfect match J=B exponentially. In addition to the second-order "retarded learning" phase transition for unbiased distributions, we show that first-order transitions can also occur. Extending the known result that the center of mass of the Gibbs ensemble has Bayes-optimal performance, we show that taking the sign of the components of this vector (clipping) leads to the vector with optimal performance in the binary space. These upper bounds are shown generally not to be saturated with the technique of transforming the components of a special continuous vector, except in asymptotic limits and in a special linear case. Simulations are presented which are in excellent agreement with the theoretical results.

Bayes Theorem↗
PubMedDate not supplied

Effects of emotional arousal on multiple memory systems: evidence from declarative and procedural learning.

Extensive evidence documents emotional modulation of hippocampus-dependent declarative memory in humans. However, little is known about the emotional modulation of striatum-dependent procedural memory. To address how emotional arousal influences declarative and procedural memory, the current study utilized (1) a picture recognition and (2) a weather prediction (WP) task (a probabilistic classification learning task), which have been shown to rely on hippocampal- and striatum-based memory systems, respectively. Observers viewed arousing or neutral pictures after (Experiment 1) or during (Experiment 2) WP training trials. A 1-wk delayed picture recognition memory test revealed enhanced declarative memory for arousing compared with neutral pictures. Arousal during encoding impaired initial WP acquisition but did not influence retention when tested after a 1-wk delay. Data from a subsequent 3-mo delayed test, however, suggested that arousal during acquisition may enhance remote WP retention. These results suggest a potential dissociation between how readily emotional arousal influences hippocampus-dependent and striatum-dependent memory systems in humans.

Adolescent↗
PubMed2004

Human midbrain sensitivity to cognitive feedback and uncertainty during classification learning.

Mesencephalic dopaminergic system (MDS) neurons may participate in learning by providing a prediction error signal to their targets, which include ventral striatal, orbital, and medial frontal regions, as well as by showing sensitivity to the degree of uncertainty associated with individual stimuli. We investigated the mechanisms of probabilistic classification learning in humans using functional magnetic resonance imaging to examine the effects of feedback and uncertainty. The design was optimized for separating neural responses to stimulus, delay, and negative and positive feedback components. Compared with fixation, stimulus and feedback activated brain regions consistent with the MDS, whereas the delay period did not. Midbrain activity was significantly different for negative versus positive feedback (consistent with coding of the "prediction error") and was reliably correlated with the degree of uncertainty as well as with activity in MDS target regions. Purely cognitive feedback apparently engages the same regions as rewarding stimuli, consistent with a broader characterization of this network.

Adult↗
PubMedDate not supplied

Generalized relevance learning vector quantization.

We propose a new scheme for enlarging generalized learning vector quantization (GLVQ) with weighting factors for the input dimensions. The factors allow an appropriate scaling of the input dimensions according to their relevance. They are adapted automatically during training according to the specific classification task whereby training can be interpreted as stochastic gradient descent on an appropriate error function. This method leads to a more powerful classifier and to an adaptive metric with little extra cost compared to standard GLVQ. Moreover, the size of the weighting factors indicates the relevance of the input dimensions. This proposes a scheme for automatically pruning irrelevant input dimensions. The algorithm is verified on artificial data sets and the iris data from the UCI repository. Afterwards, the method is compared to several well known algorithms which determine the intrinsic data dimension on real world satellite image data.

Algorithms↗
PubMed2003

From covariation to causation: a test of the assumption of causal power.

How humans infer causation from covariation has been the subject of a vigorous debate, most recently between the computational causal power account (P. W. Cheng, 1997) and associative learning theorists (e.g., K. Lober & D. R. Shanks, 2000). Whereas most researchers in the subject area agree that causal power as computed by the power PC theory offers a normative account of the inductive process. Lober and Shanks, among others, have questioned the empirical validity of the theory. This article offers a full report and additional analyses of the original study featured in Lober and Shanks's critique (M. J. Buehner & P. W. Cheng, 1997) and reports tests of Lober and Shanks's and other explanations of the pattern of causal judgments. Deviations from normativity, including the outcome-density bias, were found to be misperceptions of the input or other artifacts of the experimental procedures rather than inherent to the process of causal induction.

Analysis of Variance↗
PubMed2003

Relationship between priming and recognition in deterministic and probabilistic sequence learning.

Exposure to a repeating sequence of target stimuli in a speeded localization task can support both priming of sequence-consistent responses and recognition of sequence components. In 3 experiments with both deterministic and probabilistic sequences, the authors used a novel procedure in which measures or priming and recognition were taken concurrently and asked whether these measures can be dissociated. In all of these experiments, both measures were above chance at the group level and no evidence of dissociation was found. Item-level analyses of the data in Experiment 3 did reveal dissociations in that (a) recognition judgments were affected by response speed independently of old-new status and (b) items that were not discriminated in recognition nonetheless showed priming. However, the authors show that these data, together with the group-level results, are compatible with a formal model in which priming and recognition are based on a single common memory variable.

Adolescent↗
PubMed2003

Effector-independent and effector-dependent learning in the discrete sequence production task.

This study examined whether skill in the discrete sequence production task involves, apart from the typical effector-independent component, an effector-dependent component. To that end, 12 participants practiced two 5-key sequences, each for 1,060 trials. One group practiced with three fingers of one hand, the other group with three fingers of two hands. In a subsequent test phase, participants in both groups executed the same sequences and two new sequences with the hand configuration they had used during practice, and with the hand configuration of the other group. The results provide support for an effector-dependent component in that both groups performed the practiced sequences faster with the hand configuration they had used during practice than with the hand configuration that was new to them. In addition, the unpracticed hand configuration performed the practiced sequences faster than the new sequence, which demonstrated the effector-independent component.

Adult↗
PubMedDate not supplied

[Instruments for evaluation of memory abilities in children].

OBJECTIVES: This is a review of the more relevant instruments to measure memory abilities of children and adolescents. The strengths, limitations and future developments of these instruments are discussed. CONTENTS: Memory disturbances are present in developing individuals as a consequence of perinatal damage, brain trauma, neuroinfections, epilepsy etc. Also mnesic deficits may be observed as an element of language and learning disorders. Probably all types of amnesic syndromes well known in adults do exist in children, but there are only a few observations documented in the literature. Here are described the available batteries (TOMAL, WRAML, CMS, BEM 144) and specific tests (RBMT C,VADS, CAVLT 2, etc.) that are useful for the diagnostic measurement of verbal and visual memory abilities in children and adolescents; a Spanish version and validation is available for many of these instruments. The use of tests must be preceded by structured clinical interviews and questionnaires for parents and teachers. CONCLUSIONS: In spite of all these instruments, the diagnosis must be based on clinical criteria; new developments of instruments adapted to children aged under five years have to be done. All aspects of memory functions shall be considered procedural, explicit (semantic an autobiographical), prospective, etc. , parallel forms and typifications on more large samples are claimed.

Adolescent↗
PubMed2003

Typical and atypical antipsychotic medications differentially affect two nondeclarative memory tasks in schizophrenic patients: a double dissociation.

Nondeclarative memory (NDM) has subtypes associated with different brain regions; learning of a probabilistic classification task is impaired by striatal damage and learning of a gambling task is impaired by ventromedial prefrontocortical damage. Typical and atypical antipsychotic medications differentially affect immediate early gene expression in the striatum and frontal cortex in normal rats. This suggested the hypothesis that schizophrenic patients treated with typical antipsychotics will have impaired probabilistic classification learning (PCL) and that similar patients treated with atypical antipsychotics will have impaired learning of the gambling task. Groups of schizophrenia patients treated with typical or atypical antipsychotics did not differ from each other on the Brief Psychiatric Rating Scale (BPRS), Mini Mental State Exam (MMSE) or a number of indexes of the Wisconsin Card Sorting Task (WCST) but performed worse than normal controls on these instruments. In the first study, patients treated with typicals (n=20) but not atypicals (n=20) or normal controls (n=32) were impaired in probabilistic classification. In the second study, those treated with atypicals (n=18) but not typicals (n=18) or normal controls (n=18) were impaired in the gambling task. Results suggest that typical and atypical antipsychotics differentially affect nondeclarative memory mediated by different brain regions.

Adult↗
PubMed2004

Advanced search algorithms for information-theoretic learning with kernel-based estimators.

Recent publications have proposed various information-theoretic learning (ITL) criteria based on Renyi's quadratic entropy with nonparametric kernel-based density estimation as alternative performance metrics for both supervised and unsupervised adaptive system training. These metrics, based on entropy and mutual information, take into account higher order statistics unlike the mean-square error (MSE) criterion. The drawback of these information-based metrics is the increased computational complexity, which underscores the importance of efficient training algorithms. In this paper, we examine familiar advanced-parameter search algorithms and propose modifications to allow training of systems with these ITL criteria. The well known algorithms tailored here for ITL include various improved gradient-descent methods, conjugate gradient approaches, and the Levenberg-Marquardt (LM) algorithm. Sample problems and metrics are presented to illustrate the computational efficiency attained by employing the proposed algorithms.

Algorithms↗
PubMedDate not supplied

A self-organising network that grows when required.

The ability to grow extra nodes is a potentially useful facility for a self-organising neural network. A network that can add nodes into its map space can approximate the input space more accurately, and often more parsimoniously, than a network with predefined structure and size, such as the Self-Organising Map. In addition, a growing network can deal with dynamic input distributions. Most of the growing networks that have been proposed in the literature add new nodes to support the node that has accumulated the highest error during previous iterations or to support topological structures. This usually means that new nodes are added only when the number of iterations is an integer multiple of some pre-defined constant, A. This paper suggests a way in which the learning algorithm can add nodes whenever the network in its current state does not sufficiently match the input. In this way the network grows very quickly when new data is presented, but stops growing once the network has matched the data. This is particularly important when we consider dynamic data sets, where the distribution of inputs can change to a new regime after some time. We also demonstrate the preservation of neighbourhood relations in the data by the network. The new network is compared to an existing growing network, the Growing Neural Gas (GNG), on a artificial dataset, showing how the network deals with a change in input distribution after some time. Finally, the new network is applied to several novelty detection tasks and is compared with both the GNG and an unsupervised form of the Reduced Coulomb Energy network on a robotic inspection task and with a Support Vector Machine on two benchmark novelty detection tasks.

Algorithms↗
PubMed2006

An incremental training method for the probabilistic RBF network.

The probabilistic radial basis function (PRBF) network constitutes a probabilistic version of the RBF network for classification that extends the typical mixture model approach to classification by allowing the sharing of mixture components among all classes. The typical learning method of PRBF for a classification task employs the expectation-maximization (EM) algorithm and depends strongly on the initial parameter values. In this paper, we propose a technique for incremental training of the PRBF network for classification. The proposed algorithm starts with a single component and incrementally adds more components at appropriate positions in the data space. The addition of a new component is based on criteria for detecting a region in the data space that is crucial for the classification task. After the addition of all components, the algorithm splits every component of the network into subcomponents, each one corresponding to a different class. Experimental results using several well-known classification data sets indicate that the incremental method provides solutions of superior classification performance compared to the hierarchical PRBF training method. We also conducted comparative experiments with the support vector machines method and present the obtained results along with a qualitative comparison of the two approaches.

Algorithms↗