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 343 records · Page 19Linked to original sources
PubMed2005

Intermanual transfer effects in sequential tactuomotor learning: evidence for effector independent coding.

Results from our earlier brain imaging studies regarding motor learning have shown different areas activated during naive and practiced performance. When right handed participants moved a pen either with the dominant or non-dominant hand continuously through a cut-out maze as quickly and accurately as possible, practice resulted in decreased brain activity in right premotor and parietal areas as well as left cerebellum, while increased activity was found in the supplementary motor area (SMA). These lateralized practiced-related changes in brain activation suggest effector-independent abstract coding of information. To test this hypothesis more extensively, intermanual transfer of learning was examined in 24 male and female participants (12 right- and 12 left-handed) using the same maze-learning task. It was hypothesized that if an abstract representation of the movement is learned and stored, intermanual transfer effects should be more pronounced when participants transferred to a same maze as opposed to a mirror image of the maze. Errors and velocity were measured during the following conditions: initial naive performance (Naive); after practice on the maze (Prac); during intermanual transfer to the same maze (Transfer Identical); and to the mirror maze (Transfer Mirror). Transfer direction was tested from the dominant to non-dominant hand and vice versa. No significant differences were found between right- and left-handed participants, males and females, and transfer directions. However, intermanual transfer of learning was significantly greater to the identical maze as opposed to the mirror maze. These results showed that learning was indeed taking place at an abstract effector independent level.

Adolescent↗
PubMed2006

The tempotron: a neuron that learns spike timing-based decisions.

The timing of action potentials in sensory neurons contains substantial information about the eliciting stimuli. Although the computational advantages of spike timing-based neuronal codes have long been recognized, it is unclear whether, and if so how, neurons can learn to read out such representations. We propose a new, biologically plausible supervised synaptic learning rule that enables neurons to efficiently learn a broad range of decision rules, even when information is embedded in the spatiotemporal structure of spike patterns rather than in mean firing rates. The number of categorizations of random spatiotemporal patterns that a neuron can implement is several times larger than the number of its synapses. The underlying nonlinear temporal computation allows neurons to access information beyond single-neuron statistics and to discriminate between inputs on the basis of multineuronal spike statistics. Our work demonstrates the high capacity of neural systems to learn to decode information embedded in distributed patterns of spike synchrony.

Action Potentials↗
PubMed2003

Independent processing of the temporal and ordinal structure of movement sequences.

We investigated if the temporal and ordinal structures of sequences can be represented and learned independently. In Experiment 1, subjects learned three rhythmic sequences of key presses with the right index finger: Combined consisted of nine key presses with a corresponding temporal structure of eight intervals; Temporal had the temporal structure of Combined but was performed on one key; Ordinal had the ordinal structure of Combined but an isochronous rhythm. Subjects were divided into two groups. Group 1 first learned Combined, then Temporal and Ordinal; Group 2 first learned Temporal and Ordinal, then Combined. Strong transfer effects were seen in both groups. In Group 1, having learned combined facilitated the learning of the temporal (Temporal) or ordinal (Ordinal) sequence alone; in Group 2, having learned Temporal and Ordinal facilitated the learning of Combined, where the two are combined. This supports that subjects had formed independent temporal and ordinal representations. In Experiment 2, we investigated if these can be learned independently. Subjects repeatedly reproduced sequences with fixed temporal and random ordinal structure; random temporal and fixed ordinal structure; and random temporal and ordinal structures. Temporal and ordinal learning was seen only in the first and second sequences, respectively. In summary, we provide evidence for the existence of independent systems for learning and representation of ordinal and temporal sequences and for implicit learning of temporal sequences. This may be important for fast learning and flexibility in motor control.

Adult↗
PubMed1998

The development of features in object concepts.

According to one productive and influential approach to cognition, categorization, object recognition, and higher level cognitive processes operate on a set of fixed features, which are the output of lower level perceptual processes. In many situations, however, it is the higher level cognitive process being executed that influences the lower level features that are created. Rather than viewing the repertoire of features as being fixed by low-level processes, we present a theory in which people create features to subserve the representation and categorization of objects. Two types of category learning should be distinguished. Fixed space category learning occurs when new categorizations are representable with the available feature set. Flexible space category learning occurs when new categorizations cannot be represented with the features available. Whether fixed or flexible, learning depends on the featural contrasts and similarities between the new category to be represented and the individuals existing concepts. Fixed feature approaches face one of two problems with tasks that call for new features: If the fixed features are fairly high level and directly useful for categorization, then they will not be flexible enough to represent all objects that might be relevant for a new task. If the fixed features are small, subsymbolic fragments (such as pixels), then regularities at the level of the functional features required to accomplish categorizations will not be captured by these primitives. We present evidence of flexible perceptual changes arising from category learning and theoretical arguments for the importance of this flexibility. We describe conditions that promote feature creation and argue against interpreting them in terms of fixed features. Finally, we discuss the implications of functional features for object categorization, conceptual development, chunking, constructive induction, and formal models of dimensionality reduction.

Child↗
PubMed2002

Concurrent learning of temporal and spatial sequences.

In a serial reaction time task, stimulus events simultaneously defined spatial and temporal sequences. Responses were based on the spatial dimension. The temporal sequence was incidental to the task, defined by the response-to-stimulus intervals in Experiment 1 and stimulus onset asynchronies in Experiment 2. The two sequences were either of equal length and correlated or of unequal length. In both experiments, spatial learning occurred regardless of sequence length condition. In contrast, temporal learning occurred only in the correlated condition. These results suggest that timing is an integrated part of action representations and that incidental learning for a temporal pattern does not occur independently from the action. Interestingly, sequence learning was enhanced in the correlated condition, reflecting the integration of spatial-temporal information.

Analysis of Variance↗
PubMed2003

Learning about cues in their absence: evidence from flavour preferences and aversions.

A simple behavioural task may involve the presentation of two or more stimuli. Any learning that takes place in such a situation may be analysed in terms of the formation of an association between the central representations of those stimuli. Presumably performance based on this learning can occur because presentation of one stimulus will then activate the representations of other stimuli that were previously presented with it. To examine the role that these representations play in learning in and of themselves requires that the stimuli themselves are absent. A review of a number of flavour preference and aversion studies indicates that an associatively activated stimulus representation can support learning that is both similar to and the opposite of that maintained by the presentation of the stimulus itself. Which occurs is dependent upon the nature of the reinforcer and the temporal relationships between the training events. Although this pattern of results appears suggestive of separate learning rules, a reanalysis raises the possibility of an explanation in terms of a single associative system.

Association Learning↗
PubMed1999

The link between brain learning, attention, and consciousness.

The processes whereby our brains continue to learn about a changing world in a stable fashion throughout life are proposed to lead to conscious experiences. These processes include the learning of top-down expectations, the matching of these expectations against bottom-up data, the focusing of attention upon the expected clusters of information, and the development of resonant states between bottom-up and top-down processes as they reach an attentive consensus between what is expected and what is there in the outside world. It is suggested that all conscious states in the brain are resonant states and that these resonant states trigger learning of sensory and cognitive representations. The models which summarize these concepts are therefore called Adaptive Resonance Theory, or ART, models. Psychophysical and neurobiological data in support of ART are presented from early vision, visual object recognition, auditory streaming, variable-rate speech perception, somatosensory perception, and cognitive-emotional interactions, among others. It is noted that ART mechanisms seem to be operative at all levels of the visual system, and it is proposed how these mechanisms are realized by known laminar circuits of visual cortex. It is predicted that the same circuit realization of ART mechanisms will be found in the laminar circuits of all sensory and cognitive neocortex. Concepts and data are summarized concerning how some visual percepts may be visibly, or modally, perceived, whereas amodal percepts may be consciously recognized even though they are perceptually invisible. It is also suggested that sensory and cognitive processing in the What processing stream of the brain obey top-down matching and learning laws that are often complementary to those used for spatial and motor processing in the brain's Where processing stream. This enables our sensory and cognitive representations to maintain their stability as we learn more about the world, while allowing spatial and motor representations to forget learned maps and gains that are no longer appropriate as our bodies develop and grow from infanthood to adulthood. Procedural memories are proposed to be unconscious because the inhibitory matching process that supports these spatial and motor processes cannot lead to resonance.

Attention↗
PubMed1989

The development of representation in young children.

The research summarized here shows that young children undergo an abrupt transition in their ability to understand the relation between a scale model and a larger space. Virtually none of our 2.5-year-old subjects seemed to understand the relationship between the model and the room; almost all of the 3-year-olds did understand it. The difference seems to be that the 2.5-year-olds do not respond to the model both as a real thing and as a representation of something else. Its status as a complex, meaningful real object prevents their apprehension of its abstract relation to the room. The resistance to instruction, abrupt developmental shift, and negligible individual differences in the model task suggest the possibility of a strong maturational underpinning (Espenschade & Eckert, 1967). Further research, including a longitudinal study and cross-cultural comparisons, will be addressed to the issue of the role of experience in the development of mastery of the model task. The primary contribution of this research lies in the revelation of a hitherto undocumented abrupt developmental shift in very young children's representational flexibility--in their ability to form and coordinate multiple representations. The ability to think of one thing in two ways is an important aspect of early symbolic development. In Western cultures, where so much of a child's learning occurs via various representational media, this is a crucial step. More generally, the cognitive advance that occurs between 2.5 and 3 years of age provides a foundation for further developments in the understanding of multiple representations.

Attention↗
PubMed1999

Effects of lengthened formant transition duration on discrimination and neural representation of synthetic CV syllables by normal and learning-disabled children.

In order to investigate the precise acoustic features of stop consonants that pose perceptual difficulties for some children with learning problems, discrimination thresholds along two separate synthetic /da-ga/ continua were compared in a group of children with learning problems (LP) and a group of normal children. The continua differed only in the duration of the formant transitions. Results showed that simply lengthening the formant transition duration from 40 to 80 ms did not result in improved discrimination thresholds for the LP group relative to the normal group. Consistent with previous findings, an electrophysiologic response that is known to reflect the brain's representation of a change from one auditory stimulus to another--the mismatch negativity (MMN)--indicated diminished responses in the LP group relative to the normal group to /da/ versus /ga/ when the transition duration was 40 ms. In the lengthened transition duration condition the MMN responses from the LP group were more similar to those from the normal group, and were enhanced relative to the short transition duration condition. These data suggest that extending the duration of the critical portion of the acoustic stimulus can result in enhanced encoding at a preattentive neural level; however, this stimulus manipulation on its own is not a sufficient acoustic enhancement to facilitate increased perceptual discrimination of this place-of-articulation contrast.

Child↗
PubMed2004

Limits on reacquisition of song in adult zebra finches exposed to white noise.

Zebra finches (Taeniopygia guttata) learn a specific song pattern during a sensitive period of development, after which song changes little or not at all. However, recent studies have demonstrated substantial behavioral plasticity in song behavior during adulthood under a range of conditions. The current experiment examined song behavior of adult zebra finches temporarily deprived of auditory feedback by chronic exposure to loud white noise (WN). Long-term exposure to continuous WN resulted in disruption of song similar to that observed after deafening. When auditory feedback was restored by discontinuing WN, birds were either tutored using tape-recorded playback or housed with adult conspecific tutors. No evidence of learning new tutor syllables was observed, and recovery of pre-WN song patterns was very limited after restoration of hearing. However, many birds did reacquire some aspects of their pretreatment song, suggesting an adult form of learning that may retain some of the initial aspects of sensorimotor acquisition of song in which vocalizations are shaped to match a stored template representation. The failure to learn novel song elements and the modest degree of recovery observed overall suggest a limit on plasticity in adult birds that have acquired species-typical song patterns and may reflect an important species difference between zebra finches and Bengalese finches.

Age Factors↗
PubMed2002

Thalamocortical dynamics of the McCollough effect: boundary-surface alignment through perceptual learning.

This article further develops the FACADE neural model of 3-D vision and figure-ground perception to quantitatively explain properties of the McCollough effect (ME). The model proposes that many ME data result from visual system mechanisms whose primary function is to adaptively align, through learning, boundary and surface representations that are positionally shifted due to the process of binocular fusion. For example, binocular boundary representations are shifted by binocular fusion relative to monocular surface representations, yet the boundaries must become positionally aligned with the surfaces to control binocular surface capture and filling-in. The model also includes perceptual reset mechanisms that use habituative transmitters in opponent processing circuits. Thus the model shows how ME data may arise from a combination of mechanisms that have a clear functional role in biological vision. Simulation results with a single set of parameters quantitatively fit data from 13 experiments that probe the nature of achromatic/chromatic and monocular/binocular interactions during induction of the ME. The model proposes how perceptual learning, opponent processing, and habituation at both monocular and binocular surface representations are involved, including early thalamocortical sites. In particular, it explains the anomalous ME utilizing these multiple processing sites. Alternative models of the ME are also summarized and compared with the present model.

Adaptation, Ocular↗
PubMed2003

Dynamics of the learning of a morphokinetic movement sequence in humans.

We studied the dynamics of the process through which subjects build a visuomotor representation in memory as they learn a morphokinetic movement sequence presented on videotape. It was hypothesized that recall accuracy is closely tied to the dynamics of the movement-representation process. When analyzed in the form of time series, recall performance should exhibit a deterministic inter-trial relationship reflecting the memorization activity. Twelve subjects divided into two groups recalled a sequence of arm movements 60 times. The model was presented to Group 1 (G1) on all 60 trials, whereas Group 2 only saw the model on the first 30 trials. During these first 30 trials, the analysis of accuracy time series pointed out the short-term historicity of the deterministic inter-trials relationship. For the last 30 trials, this deterministic relationship disappeared in both experimental conditions, except for two subjects of G1, and was replaced by a white-noise process suggesting that a stable representation of the sequence has been built in long-term memory.

Analysis of Variance↗
PubMed2005

A new image representation algorithm inspired by image submodality models, redundancy reduction, and learning in biological vision.

We develop a new biologically motivated algorithm for representing natural images using successive projections into complementary subspaces. An image is first projected into an edge subspace spanned using an ICA basis adapted to natural images which captures the sharp features of an image like edges and curves. The residual image obtained after extraction of the sharp image features is approximated using a mixture of probabilistic principal component analyzers (MPPCA) model. The model is consistent with cellular, functional, information theoretic, and learning paradigms in visual pathway modeling. We demonstrate the efficiency of our model for representing different attributes of natural images like color and luminance. We compare the performance of our model in terms of quality of representation against commonly used basis, like the discrete cosine transform (DCT), independent component analysis (ICA), and principal components analysis (PCA), based on their entropies. Chrominance and luminance components of images are represented using codes having lower entropy than DCT, ICA, or PCA for similar visual quality. The model attains considerable simplification for learning from images by using a sparse independent code for representing edges and explicitly evaluating probabilities in the residual subspace.

Algorithms↗
PubMedDate not supplied

Lexicalization of idioms in urban fifth graders: a reaction time study.

Idioms are an important aspect of language that comprises a sizeable portion of our vocabulary. However, lexical access research has largely been limited to understanding how literal words are accessed and processed in the mental lexicon. Adult data show that idioms are quickly accessed from the lexicon and are likely processed as long words [J. Verbal Learn. Verbal Behav. 18 (1979) 523]. The purpose of this study was to examine lexicalization of idioms in a group of school-aged children. Using a phrase classification design, this research tested the Lexical Representation Hypothesis [J. Verbal Learn. Verbal Behav. 18 (1979) 523] in 19 urban fifth graders (5 boys, 14 girls; M age=10.16 years). On a computer, the students classified 54 phrases, including 24 idioms (high, moderate, and low familiarity [J. Speech Hear. Res. 36 (1993) 728]), 24 grammatical control word strings and 6 unrelated foils as either idioms or nonidioms. The idioms were identified with 62% accuracy. Unexpectedly, the boys tended to show higher rates of accuracy than the girls. Response latencies were shorter on the idioms compared to the controls and high familiarity idioms were processed faster than moderate and low familiarity idioms. These findings provide developmental data for lexicalization of idioms and the relationship between lexicalization and familiarity. At the cocnlusion of this articel, the reader will be able to: (1). discuss the various theories of idiom access and processing, (2). discuss how lexical access relates to idiom knowledge in school-aged children, and (3). consider the association between lexical access and familiarity relative to idiom comprehension in school-aged children.

Child↗
PubMed2003

Orientational minimal redundancy wavelets: from edge detection to perception.

Natural images are complex but very structured objects and, in spite of its complexity, the sensory areas in the neocortex in mammals are able to devise learned strategies to encode them efficiently. How is this goal achieved? In this paper, we will discuss the multiscaling approach, which has been recently used to derive a redundancy reducing wavelet basis. This kind of representation can be statistically learned from the data and is optimally adapted for image coding; besides, it presents some remarkable features found in the visual pathway. We will show that the introduction of oriented wavelets is necessary to provide a complete description, which stresses the role of the wavelets as edge detectors.

Computational Biology↗
PubMed2003

Lesions of the Basal forebrain cholinergic system impair task acquisition and abolish cortical plasticity associated with motor skill learning.

The contribution of the basal forebrain cholinergic system in mediating plasticity of cortical sensorimotor representations was examined in the context of normal learning. The effects of specific basal forebrain cholinergic lesions upon cortical reorganization associated with learning a skilled motor task were investigated, addressing, for the first time, the functional consequences of blocking cortical map plasticity. Results demonstrate that disrupting basal forebrain cholinergic function disrupts cortical map reorganization and impairs motor learning. Cholinergic lesions do not impair associative fear learning or overall sensorimotor function. These results support the hypothesis that the basal forebrain cholinergic system may be specifically implicated in forms of learning requiring plasticity of cortical representations.

Acetylcholine↗
PubMed2005

Robust self-localisation and navigation based on hippocampal place cells.

A computational model of the hippocampal function in spatial learning is presented. A spatial representation is incrementally acquired during exploration. Visual and self-motion information is fed into a network of rate-coded neurons. A consistent and stable place code emerges by unsupervised Hebbian learning between place- and head direction cells. Based on this representation, goal-oriented navigation is learnt by applying a reward-based learning mechanism between the hippocampus and nucleus accumbens. The model, validated on a real and simulated robot, successfully localises itself by recalibrating its path integrator using visual input. A navigation map is learnt after about 20 trials, comparable to rats in the water maze. In contrast to previous works, this system processes realistic visual input. No compass is needed for localisation and the reward-based learning mechanism extends discrete navigation models to continuous space. The model reproduces experimental findings and suggests several neurophysiological and behavioural predictions in the rat.

Animals↗