Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “representation 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 613 records · Page 34Linked to original sources
PubMed1993

Right hemisphere representation of autonomic conditioning to facial emotional expressions.

In this experiment, a lateralized right hemisphere effect was found for electrodermal associative learning to facial emotional expressions. Sixty-two subjects were presented simultaneously with a slide of a happy face in the right or left visual half field (VHF) and a slide of an angry face in the opposite VHF. Four groups were formed by the combination of the two VHF positions of angry/happy faces and the administration/omission of shock unconditioned stimuli. The results showed that simultaneous presentation of the angry face to the right hemisphere and the happy face to the left hemisphere, together with shock, resulted in a strong conditioned association with the angry face and a relatively weak association with the happy face. Furthermore, simultaneous presentation of the angry face to the left hemisphere and the happy face to the right hemisphere, together with shock, resulted in a relatively weak association with both stimuli. No significant differences were found for the no-shock control groups. The present results confirm previous findings of a right hemisphere advantage for representation of associative learning.

Adolescent↗
PubMed2003

Hippocampal map realignment and spatial learning.

The spatial selectivity of hippocampal neurons suggests that they contribute to an internal representation of current location. The activity of hippocampal pyramidal cells was recorded while adult (10-13 months old) and aged (24-28 months old) rats performed a task in which two spatial reference frames were put in conflict. Rats attempted to find an unmarked goal whose position was fixed relative to only one of the two reference frames. The ability of a rat's hippocampus to adjust to the conflicting information and use the 'correct' position estimate (hippocampal map 'realignment') was correlated with the rat's ability to find the hidden goal. In addition, aged rats were impaired relative to adult rats in both goal-finding accuracy and map realignment. Thus, changes in the effectiveness with which the hippocampal spatial representation is updated on the basis of external cues may contribute to both within-age-group spatial learning variability and age-related spatial learning deficits.

Adaptation, Physiological↗
PubMed2006

Relationship between spatial abilities, mental rotation and functional anatomy learning.

This study investigated the relationship between visuo-spatial representation, mental rotation (MR) and functional anatomy examination results. A total of 184 students completed the Group Embedded Figures Test (GEFT), Mental Rotation Test (MRT) and Gordon Test of Visual Imagery Control. The time spent on personal assignment was also considered. Men were found to score better than women on both GEFT and MRT, but the gender effect was limited to the interaction with MRT ability in the anatomy learning process. Significant correlations were found between visuo-spatial, MR abilities, and anatomy examination results. Data resulting from the best students' analyzes underscore the effect of high MR ability which may be considered reliable predictor of success in learning anatomy. The use of specific tests during learning sessions may facilitate the acquisition of anatomical knowledge.

Adolescent↗
PubMed2006

Effect of sparse basis selection on ultrasonic signal representation.

Recently, adaptive sparse representations of ultrasonic signals have been utilized to improve the performance of scanning acoustic microscopy (SAM), a common nondestructive tool for failure analysis of microelectronic packages. The adaptive sparse representation of an ultrasonic signal is generated by decomposing it in a learned overcomplete dictionary using a sparse basis selection algorithm. Detection and location of ultrasonic echoes are then performed on the basis of the resulting redundant representation. This paper investigates the effect of sparse basis selection algorithms on ultrasonic signal representation. The overcomplete independent component analysis, focal underdetermined system solver (FOCUSS), and sparse Bayesian learning algorithms are examined. Numerical simulations are performed to quantitatively analyze the efficiency of ultrasonic signal representations. Experiments with ultrasonic A-scans acquired from flip-chip packages are also carried out in the study. The efficiency of ultrasonic signal representations are evaluated in terms of the different criteria that can be used to measure its performance for different SAM applications, such as waveform estimation, echo detection, echo location and C-scan imaging. The results show that the FOCUSS algorithm performs best overall.

Algorithms↗
PubMed2005

Classification of non-coding RNA using graph representations of secondary structure.

Some genes produce transcripts that function directly in regulatory, catalytic, or structural roles in the cell. These non-coding RNAs are prevalent in all living organisms, and methods that aid the understanding of their functional roles are essential. RNA secondary structure, the pattern of base-pairing, contains the critical information for determining the three dimensional structure and function of the molecule. In this work we examine whether the basic geometric and topological properties of secondary structure are sufficient to distinguish between RNA families in a learning framework. First, we develop a labeled dual graph representation of RNA secondary structure by adding biologically meaningful labels to the dual graphs proposed by Gan et al [1]. Next, we define a similarity measure directly on the labeled dual graphs using the recently developed marginalized kernels [2]. Using this similarity measure, we were able to train Support Vector Machine classifiers to distinguish RNAs of known families from random RNAs with similar statistics. For 22 of the 25 families tested, the classifier achieved better than 70% accuracy, with much higher accuracy rates for some families. Training a set of classifiers to automatically assign family labels to RNAs using a one vs. all multi-class scheme also yielded encouraging results. From these initial learning experiments, we suggest that the labeled dual graph representation, together with kernel machine methods, has potential for use in automated analysis and classification of uncharacterized RNA molecules or efficient genome-wide screens for RNA molecules from existing families.

Base Sequence↗
PubMed2006

Enhanced context-dependency of object recognition in rats with hippocampal lesions.

Object recognition memory was assessed on a novel-object preference (NOP) task in rats with lesions of the hippocampal formation (HPC). The learning and test phases of NOP trials occurred in either the same context or in different contexts. When the learning and test contexts were the same, rats with HPC lesions performed like control rats, displaying a significant tendency to investigate a novel object more than a familiar sample object. When the test occurred in a context that was familiar but different from the learning context, performance was unaffected in control rats, but rats with HPC lesions no longer discriminated between the objects, and therefore showed no evidence of recognizing the sample object. When the test context was unfamiliar, novel-object preference in control rats was attenuated but still above chance levels, whereas rats with HPC lesions did not show a preference. The data suggest that the HPC is not critical for encoding or retrieving conjunctive representations of the context in which incidental learning occurs, whereas it plays an essential role in recognition of objects that are subsequently encountered in different contexts.

Animals↗
PubMed2005

Retrospective revaluation as simple associative learning.

Backward blocking, unovershadowing, and backward conditioned inhibition are examples of retrospective revaluation phenomena that have been suggested to involve more than simple associative learning. Models of these phenomena have thus used additional concepts, for example, appealing to attentional effects or more elaborate learning mechanisms. The author shows that a suitable representation of stimuli, paired with a careful analysis of the discriminations faced by animals, leads to an account of these and other phenomena in terms of a simple elemental model of associative learning, with essentially the same learning mechanism as the R. A. Rescorla and A. R. Wagner (1972) model. The author concludes with a discussion of some implications for theories of learning.

Animals↗
PubMed1999

Apraxia in corticobasal degeneration.

Corticobasal degeneration (CBD) is a degenerative disease that often presents with an asymmetric progressive ideomotor limb apraxia. Some apraxic subjects may fail to perform skilled purposive movements on command because they have lost the memories or representations that specify how these movements should be performed (representational deficit). In contrast, other apraxic subjects may have the movement representations but are unable to utilize the information contained in them to execute skilled purposive movements (production-execution deficit). To learn if the apraxic deficit in CBD is induced by a representational or a production-execution deficit, we tested three nondemented subjects with CBD on tasks requiring production of meaningful or meaningless gestures to command, gesture imitation, gesture discrimination, and novel gesture learning. A fourth subject with incomplete data also is presented. The results suggest that the apraxia associated with CBD is initially induced by a production-execution defect with relative sparing of the movement representations.

Aged↗
PubMed1996

Natural image statistics and efficient coding.

Natural images contain characteristic statistical regularities that set them apart from purely random images. Understanding what these regularities are can enable natural images to be coded more efficiently. In this paper, we describe some of the forms of structure that are contained in natural images, and we show how these are related to the response properties of neurons at early stages of the visual system. Many of the important forms of structure require higher-order (i.e. more than linear, pairwise) statistics to characterize, which makes models based on linear Hebbian learning, or principal components analysis, inappropriate for finding efficient codes for natural images. We suggest that a good objective for an efficient coding of natural scenes is to maximize the sparseness of the representation, and we show that a network that learns sparse codes of natural scenes succeeds in developing localized, oriented, bandpass receptive fields similar to those in the mammalian striate cortex.

Journal Article↗
PubMed1998

Can connectionism save constructivism?

Constructivism is the Piagetian notion that learning leads the child to develop new types of representations. For example, on the Piagetian view, a child is born without knowing that objects persist in time even when they are occluded; through a process of learning, the child comes to know that objects persist in time. The trouble with this view has always been the lack of a concrete, computational account of how a learning mechanism could lead to such a change. Recently, however, in a book entitled Rethinking Innateness. Elman et al. (Elman, J.L., Bates, E., Johnson, M.H., Karmiloff-Smith, A., Parisi, D., Plunkett, K., 1996. Rethinking Innateness: A Connectionist Perspective on Development. Cambridge, MA: MIT Press) have claimed that connectionist models might provide an account of the development of new kinds of representations that would not depend on the existence of innate representations. I show that the models described in Rethinking Innateness depend on innately assumed representations and that they do not offer a genuine alternative to nativism. Moreover, I present simulation results which show that these models are incapable of deriving genuine abstract representations that are not presupposed. I then give a formal account of why the models fail to generalize in the ways that humans do. Thus, connectionism, at least in its current form, does not provide any support for constructivism. I conclude by sketching a possible alternative.

Child Development↗
PubMed1999

Early learning and the development of filial preferences in the chick.

Newly hatched domestic chicks (Gallus gallus domesticus) rapidly form a social preference for a conspicuous stimulus to which they are exposed. The learning process involved is known as filial imprinting. When chicks are exposed to an audio-visual compound stimulus, both auditory and visual learning are enhanced. The enhancement of visual imprinting is virtually abolished when chicks are exposed separately to the auditory element, either before or after training with the audio-visual compound. Simultaneous exposure to the two elements of the compound is superior to sequential exposure in achieving the enhancement of visual learning. These results are unlike Pavlovian conditioning, but are consistent with an interpretation of imprinting as a form of within-event learning, where links are formed between the representations of the elements of the compound, that can be weakened by separate exposure to an element. Apart from imprinting, chicks may show a developing predisposition to approach stimuli resembling conspecifics. The predisposition emerges in dark-reared chicks given some non-specific experience during a sensitive period, and is expressed as a relatively general preference for stimuli with a head and neck region. In the natural situation, the animal's response may be biased by the predisposition, and through imprinting it then learns the characteristics of individuals.

Animals↗
PubMed2005

Modeling individual differences in cognition.

Many evaluations of cognitive models rely on data that have been averaged or aggregated across all experimental subjects, and so fail to consider the possibility of important individual differences between subjects. Other evaluations are done at the single-subject level, and so fail to benefit from the reduction of noise that data averaging or aggregation potentially provides. To overcome these weaknesses, we have developed a general approach to modeling individual differences using families of cognitive models in which different groups of subjects are identified as having different psychological behavior. Separate models with separate parameterizations are applied to each group of subjects, and Bayesian model selection is used to determine the appropriate number of groups. We evaluate this individual differences approach in a simulation study and show that it is superior in terms of the key modeling goals of prediction and understanding. We also provide two practical demonstrations of the approach, one using the ALCOVE model of category learning with data from four previously analyzed category learning experiments, the other using multidimensional scaling representational models with previously analyzed similarity data for colors. In both demonstrations, meaningful individual differences are found and the psychological models are able to account for this variation through interpretable differences in parameterization. The results highlight the potential of extending cognitive models to consider individual differences.

Cognition↗
PubMed1996

Theoretical foundations of cognitive-behavior therapy for anxiety and depression.

Cognitive-behavior therapy (CBT) involves a highly diverse set of terms and procedures. In this review, the origins of CBT are briefly considered, and an integrative theoretical framework is proposed that (a) distinguishes therapy interventions targeted at circumscribed disorders from those targeted at generalized disorders and (b) distinguishes interventions aimed at modifying conscious beliefs and representations from those aimed at modifying unconscious representations in memory. Interventions aimed at altering consciously accessible beliefs are related to their theoretical bases in appraisal theories of emotion and cognitive theories of emotion and motivation. Interventions aimed at modifying unconscious representations are related to their theoretical bases in learning theory and findings from experimental cognitive psychology. In the review, different formulations of CBT for anxiety disorders and depression are analyzed in terms of this framework, and theoretical issues relating to self-representations in memory and to emotional processing are considered.

Anxiety Disorders↗
PubMedDate not supplied

Predictive regulation of associative learning in a neural network by reinforcement and attentive feedback.

At least four types of learning processes are relevant in the present paper: learning of conditioned reinforcement, incentive motivation, sensory expectancy, and motor command. These several types of learning processes, which operate on a slow time scale, regulate and are regulated by rapidly fluctuating limited capacity STM representations of sensory events. The theory suggest how nonlinear feedback interactions among these fast information processing mechanisms and slow learning mechanisms participate in different conditioning paradigms, and actively regulate learning and memory to generate predictive internal representations of external environmental contingencies.

Animals↗
PubMed1988

Models of memory: information processing.

A complete understanding of human memory will necessarily involve consideration of the active processes involved at the time of learning and of the organization and nature of representation of information in long-term memory. In addition to process and structure, it is important for theory to indicate the ways in which stimulus-driven and conceptually driven processes interact with each other in the learning situation. Not surprisingly, no existent theory provides a detailed specification of all of these factors. However, there are a number of more specific theories which are successful in illuminating some of the component structures and processes. The working memory model proposed by Baddeley and Hitch (1974) and modified subsequently has shown how the earlier theoretical construct of the short-term store should be replaced with the notion of working memory. In essence, working memory is a system which is used both to process information and to permit the transient storage of information. It comprises a number of conceptually distinct, but functionally interdependent components. So far as long-term memory is concerned, there is evidence of a number of different kinds of representation. Of particular importance is the distinction between declarative knowledge and procedural knowledge, a distinction which has received support from the study of amnesic patients. Kosslyn has argued for a distinction between literal representation and propositional representation, whereas Tulving has distinguished between episodic and semantic memories. While Tulving's distinction is perhaps the best known, there is increasing evidence that episodic and semantic memory differ primarily in content rather than in process, and so the distinction may be of less theoretical value than was originally believed.(ABSTRACT TRUNCATED AT 250 WORDS)

Humans↗
PubMed2005

The dynamics of perceptual learning: an incremental reweighting model.

The mechanisms of perceptual learning are analyzed theoretically, probed in an orientation-discrimination experiment involving a novel nonstationary context manipulation, and instantiated in a detailed computational model. Two hypotheses are examined: modification of early cortical representations versus task-specific selective reweighting. Representation modification seems neither functionally necessary nor implied by the available psychophysical and physiological evidence. Computer simulations and mathematical analyses demonstrate the functional and empirical adequacy of selective reweighting as a perceptual learning mechanism. The stimulus images are processed by standard orientation- and frequency-tuned representational units, divisively normalized. Learning occurs only in the "read-out" connections to a decision unit; the stimulus representations never change. An incremental Hebbian rule tracks the task-dependent predictive value of each unit, thereby improving the signal-to-noise ratio of their weighted combination. Each abrupt change in the environmental statistics induces a switch cost in the learning curves as the system temporarily works with suboptimal weights.

Brain↗
PubMed1998

Is there a neural code?

Rate coding and temporal coding are two extremes of the neural coding process. The concept of a stationary state corresponds to the information processing approach that views the brain as a decision maker, adopts rate coding as its main strategy and endorses the single- or few neuron approach. If information derived from sensory stimulation is used to continuously update the brain's internal representation of the world, then neural codes may change with time through learning. As a consequence, the same spike sequence may be interpreted differently (or evoke a different behavior) later in the day. This non-stationary viewpoint is embodied in the representational model of brain function that stresses learning and plasticity and employs temporal coding in neural assemblies. We argue that the switching between quasi-stable brain states as a result of learning is more relevant than the neuronal patterns, and the correlations between them, that are found during stationary states. The neural code likely resides in the activity patterns that cause this state-switching.

Animals↗
PubMed2004

From plasticity to complexity: a new diagnostic method for psychiatry.

There is growing dissatisfaction regarding the available diagnostic systems for psychiatric disorders (DSM, ICD). Psychiatrists acknowledge that though mental disease reflects brain disorders, the descriptive and symptom based nature of psychiatric diagnosis bears no relation to brain functions. According to Helmut's article published in the October 2003 issue of Science, in the coming decade researchers and psychiatrists will be called upon to propose a basis for the psychiatric diagnostic system of the future. I propose a new etiology-oriented diagnostic system for psychiatry by integrating two recently emerging bodies of knowledge, one regarding plasticity and the other involving complex systems. Plasticity refers to all brain processes involved in dynamic alterations within communicating neuronal ensembles or networks, in the brain. Complexity refers to certain formulations from system theories relevant to brain dynamics and plasticity. It is proposed to divide plasticity processes into three types based on time domains: (1) "developmental plasticity", (2) "tuning plasticity" and (3) "fast stabilizing plasticity". Each type of plasticity is related to different complexity models achieved by the brain, developmental plasticity is life-long brain organization, it is related to state-space configurations molded into brain representations internalized via processes such as Hebbian learning. Tuning plasticity is related to "matching complexity" a measure of adaptability between internal configurations in the brain-system and externally originating event stimuli. Fast stabilizing plasticity is related to "neural complexity" a measure of neural network integration in the brain. Neural complexity meets the mental requirement to extract important features from different sensory inputs and to simultaneously generate coherent perceptual and cognitive states, thus balancing specialized segregated brain processes with coherent globally integrated whole brain activity. Mental disorders can be reconceptualized as disorders of plasticity resulting in disturbances of state-space brain configurations, matching and neural complexities. Personality disorders result from altered internal representations of the psychosocial environment. Depression and anxiety have been recently linked to alterations of adaptive neuronal plasticity thus reconceptualized as disorders of matching complexity. Finally, psychoses, including schizophrenia spectrum disorders, are reconceptualized as disturbances of neural complexity resulting in altered fast stabilizing plasticity. The new diagnostic system generates testable predictions regarding diagnosis and treatments of mental disorders which may be the future of psychiatry.

Cognition↗