SEARCH · PubMed Health

Results for “GENERALIZATION LEARNING”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 181 records · Page 10Linked to original sources

PubMed2004

Learning eigenfunctions links spectral embedding and kernel PCA.

In this letter, we show a direct relation between spectral embedding methods and kernel principal components analysis and how both are special cases of a more general learning problem: learning the principal eigenfunctions of an operator defined from a kernel and the unknown data-generating density. Whereas spectral embedding methods provided only coordinates for the training points, the analysis justifies a simple extension to out-of-sample examples (the Nyström formula) for multidimensional scaling (MDS), spectral clustering, Laplacian eigenmaps, locally linear embedding (LLE), and Isomap. The analysis provides, for all such spectral embedding methods, the definition of a loss function, whose empirical average is minimized by the traditional algorithms. The asymptotic expected value of that loss defines a generalization performance and clarifies what these algorithms are trying to learn. Experiments with LLE, Isomap, spectral clustering, and MDS show that this out-of-sample embedding formula generalizes well, with a level of error comparable to the effect of small perturbations of the training set on the embedding.

Algorithms↗

PubMed2000

Universal learning network and its application to chaos control.

Universal Learning Networks (ULNs) are proposed and their application to chaos control is discussed. ULNs provide a generalized framework to model and control complex systems. They consist of a number of inter-connected nodes where the nodes may have any continuously differentiable nonlinear functions in them and each pair of nodes can be connected by multiple branches with arbitrary time delays. Therefore, physical systems, which can be described by differential or difference equations and also their controllers, can be modeled in a unified way, and so ULNs may form a super set of neural networks and fuzzy neural networks. In order to optimize the ULNs, a generalized learning algorithm is derived, in which both the first order derivatives (gradients) and the higher order derivatives are incorporated. The derivatives are calculated by using forward or backward propagation schemes. These algorithms for calculating the derivatives are extended versions of Back Propagation Through Time (BPTT) and Real Time Recurrent Learning (RTRL) of Williams in the sense that generalized node functions, generalized network connections with multi-branch of arbitrary time delays, generalized criterion functions and higher order derivatives can be deal with. As an application of ULNs, a chaos control method using maximum Lyapunov exponent of ULNs is proposed. Maximum Lyapunov exponent of ULNs can be formulated by using higher order derivatives of ULNs, and the parameters of ULNs can be adjusted so that the maximum Lyapunov exponent approaches the target value. From the simulation results, it has been shown that a fully connected ULN with three nodes is able to display chaotic behaviors.

Artificial Intelligence↗

PubMed2000

Incremental active learning for optimal generalization.

The problem of designing input signals for optimal generalization is called active learning. In this article, we give a two-stage sampling scheme for reducing both the bias and variance, and based on this scheme, we propose two active learning methods. One is the multipoint search method applicable to arbitrary models. The effectiveness of this method is shown through computer simulations. The other is the optimal sampling method in trigonometric polynomial models. This method precisely specifies the optimal sampling locations.

Algorithms↗

PubMed2003

Temporally massed CS presentations generate more fear extinction than spaced presentations.

Rodent fear conditioning models both excitatory learning and the pathogenesis of human anxiety, whereas extinction of conditional fear is a paradigm of inhibitory learning and the explicit model for behavior therapy. Many studies support a general learning rule for acquisition: Temporally spaced training is more effective than massed training. The authors asked whether this rule applies to extinction of conditional fear in mice. The authors find that both short- and long-term fear extinction are greater with temporally massed presentations of the conditional stimulus (CS). The data also indicate that once CS presentations are sufficiently massed to initiate, or "induce," extinction learning, further CS presentations are more effective when spaced.

Animals↗

PubMed1997

Learning in the development of infant locomotion.

Infants master crawling and walking in an environment filled with varied and unfamiliar surfaces. At the same time, infants' bodies and skills continually change. The changing demands of everyday locomotion require infants to adapt locomotion to the properties of the terrain and to their own physical abilities. This Monograph examines how infants acquire adaptive locomotion in a novel task--going up and down slopes. Infants were tested longitudinally from their first week of crawling until several weeks after they began walking. Everyday locomotor experience played a central role in adaptive responding. Over weeks of crawling, infants' judgments became increasingly accurate, and exploration became increasingly efficient. There was no transfer over the transition from crawling to walking. Instead, infants learned, all over again, how to cope with slopes from an upright position. Findings indicate that learning generalized from everyday experience traveling over flat surfaces at home but that learning was specific to infants' typical method of locomotion and vantage point. Moreover, learning was not the result of simple associations between a particular locomotor response and a particular slope. Rather, infants learned to gauge their abilities on-line as they encountered each hill at the start of the trial. Change in locomotor responses and exploratory movements revealed a process of differentiation and selection spurred by changes in infants' everyday experience, body dimensions, and locomotor proficiency on flat ground.

Child Development↗

PubMed2003

Rapid motor learning in the translational vestibulo-ocular reflex.

Motor learning was induced in the translational vestibulo-ocular reflex (TVOR) when monkeys were repeatedly subjected to a brief (0.5 sec) head translation while they tried to maintain binocular fixation on a visual target for juice rewards. If the target was world-fixed, the initial eye speed of the TVOR gradually increased; if the target was head-fixed, the initial eye speed of the TVOR gradually decreased. The rate of learning acquisition was very rapid, with a time constant of approximately 100 trials, which was equivalent to <1 min of accumulated stimulation. These learned changes were consolidated over >or=1 d without any reinforcement, indicating induction of long-term synaptic plasticity. Although the learning generalized to targets with different viewing distances and to head translations with different accelerations, it was highly specific for the particular combination of head motion and evoked eye movement associated with the training. For example, it was specific to the modality of the stimulus (translation vs rotation) and the direction of the evoked eye movement in the training. Furthermore, when one eye was aligned with the heading direction so that it remained motionless during training, learning was not expressed in this eye, but only in the other nonaligned eye. These specificities show that the learning sites are neither in the sensory nor the motor limb of the reflex but in the sensory-motor transformation stage of the reflex. The dependence of the learning on both head motion and evoked eye movement suggests that Hebbian learning may be one of the underlying cellular mechanisms.

Animals↗

PubMed1996

Modelling techniques and their application for monitoring in high dependency environments--learning models.

This paper reviews the use of learning models including Bayesian classifiers and artificial neural networks in monitoring and interpreting biosignals. Generally learning models applied for analysis of biosignals are "black-box' types trained on the basis of measured signals. It is illustrated that the training and application of learning models more or less follow the same sequences. The main focus is the interpretation of electrical signals from the brain (electroencephalogram (EEG) and evoked potentials (EP)). Current analysis of these signals often reveals sudden changes in the EEG or evoked potentials to be the earliest discernible signs of inadequate perfusion of the brain. They may reflect problems such as systemic arterial oxygen desaturation or hypotension arising from other body system failures during critical illness. It is suggested that these brain signals should be recorded in the critical care unit, and that they should form part of the annotated database of biosignals established during the IMPROVE project. This would allow for the development of new methods for on-line warning of impending damage to the central nervous system, such that corrective actions could be taken before permanent damage occurred.

Auscultation↗

PubMed1981

Reflections on the treatment process by nonabstinent alcoholics.

This study reports on the subjective evaluations of treatment by 162 patients who had been treated and discharged with staff approval at Hazelden Rehabilitation Center. At 2 years following treatment, these patients were asked to comment on what they considered the most important aspects of their treatment experience. Four principal groupings of comments emerged as being most important: (1) Formal Hazelden activities, i.e., lectures, counseling; (2) Inner-directed learning, i.e., understanding, insight; (3) Informal Hazelden activities, i.e., personal friendships, fellowship; and (4) General learning, i.e., effect of alcohol on mind and body. The implications of these findings are discussed in terms of the utility of evaluating treatment effectiveness through the subjective judgments of the treated patients.

Adolescent↗

PubMed2006

The effect of L-stepholidine, a novel extract of Chinese herb, on the acquisition, expression, maintenance, and re-acquisition of morphine conditioned place preference in rats.

The effect of L-stepholidine (SPD), a novel alkaloid extract of the Chinese herb Stephania with partial dopamine D1 receptor agonistic and D2 receptor antagonistic dual actions, on morphine conditioned place preference (CPP) was studied. Daily injection of morphine (10 mg/kg, i.p.) for 6 days induced CPP in rats, and daily treatment with SPD at 10 or 20 mg/kg before morphine injection dose-dependently attenuated morphine-induced CPP. On the day following acquisition of morphine CPP, a single administration of SPD at 10 or 20 mg/kg failed to block the expression of CPP. However, daily administration of SPD at 20 mg/kg for 7 days attenuated the maintenance of CPP. Morphine-induced CPP extinguished after a 21-day saline training and then a single injection of morphine (3 mg/kg, i.p.) induced re-acquisition of morphine CPP; however, pretreatment with SPD at 10 or 20 mg/kg 30 min before morphine injection dose-dependently blocked morphine (3 mg/kg, i.p.)-induced re-acquisition of morphine CPP. Furthermore, our data indicate that SPD had no effect on food-induced CPP or state-dependent learning, suggesting that the observed effect of SPD does not result from an inhibition of general learning ability. These results demonstrate that SPD can inhibit acquisition, maintenance, and re-acquisition of morphine conditioned place preference and suggest its potential for treatment of opioid addiction.

Analgesics, Opioid↗

PubMed2006

Application of generalized dynamic neural networks to biomedical data.

This paper reviews the application of continuous recurrent neural networks with time-varying weights to pattern recognition tasks in medicine. A general learning algorithm based on Pontryagin's maximum principle is recapitulated, and possibilities of improving the generalization capabilities of these networks are given. The effectiveness of the methods is demonstrated by three different real-world examples taken from the fields of anesthesiology, orthopedics, and radiology.

Algorithms↗

PubMed1986

Lack of correlation between cortical levels of choline acetyltransferase and learning scores in rats with globus pallidus lesions.

Previous work has shown that rats with lesions of the globus pallidus (GP) exhibit a generalized learning impairment. Data are presented suggesting that this impairment is not due to inadvertent damage to the nucleus basalis magnocellularis. Rats with GP lesions evidenced a significant visual discrimination learning loss and a significant reduction in cortical choline acetyltransferase (ChAT) activity. However, there was no significant correlation between the severity of the learning loss and the amount of reduction of cortical ChAT activity.

Animals↗

PubMed2003

Average-KR schedule benefits generalized motor program learning.

The main purpose of this study was to examine the effects of average Knowledge of Results (KR) on generalized motor program learning and parameter learning. Two groups of participants (n = 15 per group) performed 80 acquisition trials of sequential timing tasks. All participants were asked to depress sequentially four keys (2, 4, 8, and 6) on the numeric pad portion of the computer keyboard with the index finger of the right hand. The author presented average feedback on timing errors based on 5-trial blocks and compared this feedback schedule with every-trial feedback. Analysis of the delayed no-feedback retention test indicated a strong advantage for the average KR compared with the every-trial condition in both generalized motor program learning and parameter learning. The current results suggest that the average KR schedule may have positive effects on generalized motor program learning and parameter learning.

Adult↗

PubMed2005

Intact visual latent inhibition in schizophrenia patients in a within-subject paradigm.

People are normally slower to learn a CS-UCS association if they first experience the CS without the UCS. This normal slowing, termed "latent inhibition" (LI), is reported by some to be absent in schizophrenia patients. Our previous studies detected generalized learning deficits but not LI deficits in schizophrenia patients, using between-subject auditory and visual LI paradigms. To understand our divergent results, we developed a within-subject visual LI paradigm that detects LI in normal male subjects that we previously reported to be disrupted by acute treatment with dopamine agonists. In the present study, we verified the ability of this dopamine-sensitive within-subject LI paradigm to detect LI among both male and female normal control subjects, and then used this paradigm to assess LI in schizophrenia patients. Among normals, LI exhibited no sex differences or menstrual cyclicity. Compared to normals, schizophrenia patients exhibited learning deficits with both preexposed (PE) and non-preexposed (NPE) stimuli. Despite these generalized deficits, both acutely hospitalized patients and stable outpatients with schizophrenia exhibited robust LI, as evidenced by significantly faster learning with NPE than PE stimuli. LI deficits in schizophrenia may be paradigm-specific and are not detected by a paradigm that we previously reported to be sensitive to disruption by dopamine agonists.

Adolescent↗

PubMed1998

An intervention programme for children with moderate learning difficulties.

BACKGROUND: Children with learning difficulties typically demonstrate problems generalising what they have learned in specific contexts to new situations. Intervention programmes that teach children regulatory skills have been shown to overcome these problems for children with reading difficulties. This study applies the principles that underlie such interventions to the design of a programme for children with more general learning difficulties. AIM: The aim of the study reported here was to design and evaluate an intervention programme which would facilitate the transfer of skills learned over the course of the intervention to different types of task. SAMPLE: A total of 41 children (aged 13-16 yrs; 14 females, 27 males) with moderate learning difficulties participated in the intervention programme. METHOD: The children participated in a twelve-week intervention programme designed to promote regulatory strategies over a range of activities. The children worked in pairs and were guided by an adult who provided explicit instruction and modelled appropriate strategies. The children were assessed before and after the programme to determine whether improvements in strategic behaviour were evident and whether this generalised to new situations. RESULTS: Gains made in regulatory skills over the course of the intervention were accompanied by improvements in other performance measures such as reading and IQ. CONCLUSIONS: Even after a relatively short intervention, which focused on the development of regulatory skills, significant improvements on a range of tasks were observed for a group of children with moderate learning difficulties.

Adolescent↗

PubMed1998

Perturbed dentate gyrus function in serotonin 5-HT2C receptor mutant mice.

Serotonin systems have been implicated in the regulation of hippocampal function. Serotonin 5-HT2C receptors are widely expressed throughout the hippocampal formation, and these receptors have been proposed to modulate synaptic plasticity in the visual cortex. To assess the contribution of 5-HT2C receptors to the serotonergic regulation of hippocampal function, mice with a targeted 5-HT2C-receptor gene mutation were examined. An examination of long-term potentiation at each of four principal regions of the hippocampal formation revealed a selective impairment restricted to medial perforant path-dentate gyrus synapses of mutant mice. This deficit was accompanied by abnormal performance in behavioral assays associated with dentate gyrus function. 5-HT2C receptor mutants exhibited abnormal performance in the Morris water maze assay of spatial learning and reduced aversion to a novel environment. These deficits were selective and were not associated with a generalized learning deficit or with an impairment in the discrimination of spatial context. These results indicate that a genetic perturbation of serotonin receptor function can modulate dentate gyrus plasticity and that plasticity in this structure may contribute to neural mechanisms underlying hippocampus-dependent behaviors.

Animals↗

PubMed2004

Context-independent directional cue learning by hippocampal place cells.

In a symmetrical environment possessing no other polarizing visual cues, the spatially localized firing of hippocampal place cells can be primarily orientated by a reliable distal visual stimulus, such as a white cue card. However, if such a directional cue is made unreliable by being frequently moved in full view of the rat, the rat's internal sense of direction comes, over the course of a few days, to control the orientation of place fields instead. We investigated whether this simple form of 'cue-instability' learning would transfer to a new context, in which the firing patterns of the place cells become reorganized and in which a new spatial representation is thus active. We found that after cue-instability learning, the 'remapped' place field representation in the new environment was also orientated by the internal sense of direction of the rat rather than by the cue card, showing that the cue learning generalized from one context (and hence spatial representation) to another. This contrasts with another kind of place cell learning, in which the cells can acquire the ability to discriminate two spatial locations in one context but do not transfer this discrimination to a new context. We discuss the different effects of context changes on learned place cell activity in terms of the possible architecture of the inputs to place cells.

Animals↗

PubMed1992

Active and passive P3 latency and psychometric performance: influence of age and individual differences.

The relationship of P3 latency of the event-related potential (ERP) to psychometric performance was investigated in 41 subjects who ranged in age from 20 to 88 years. P3 responses were recorded from subjects using an auditory oddball paradigm with and without task-demands. Subjects also received psychometric tests of verbal performance, visuospatial performance, concentration, and immediate, recent and remote memory. Factor analysis was used to reduce the set of psychometric measures to four factors (Verbal learning, general intelligence, narrative recall/fluency, and concentration). Both passive and active P3 latency showed a linear increase with age. Age was inversely correlated with verbal learning performance. After accounting for the influence of age, passive P3 latency correlated with the psychometric factor associated with narrative recall and verbal fluency. Active P3 latency was correlated with factors reflecting general intelligence and concentration. These findings suggest that cognitive processing speed contributes to psychometric performance in adults. The psychological or biological basis for this relationship remains to be identified.

Adult↗

PubMed1997

Performance of PTSD patients on standard tests of memory. Implications for trauma.

Mental health professionals have employed a variety of clinical and experimental neuropsychological tests for exploring purported memory alterations in PTSD. Protocols range from standard tests of immediate and delayed learning, recall, and recognition to elaborate paradigms using experimental stimuli for assessment of information-processing skills. Whereas the former have typically focused on general learning and memory capabilities, experimental paradigms have examined the role of trauma-related cues and their impact on remembering. Findings to date suggest that memory abilities in PTSD patients range from intact to mildly impaired on general tests of verbal or visual memory. At the same time, memory tests involving trauma-specific stimuli point to alterations in cognitive information processing, specifically, an attentional bias manifested by changes in speed, accuracy, and depth of processing. The role of a semantic information network involving enhanced specificity for trauma cues is discussed along with possible implications for brain structures and theories of PTSD.

Humans↗