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 145 records · Page 8Linked to original sources
PubMedDate not supplied

A spike based learning rule for generation of invariant representations.

For biological realism, models of learning in neuronal networks often assume that synaptic plasticity solely depends on locally available signals, in particular only on the activity of the pre- and post-synaptic cells. As a consequence, synapses influence the plasticity of other synapses exclusively via the post-synaptic activity. Inspired by recent research on the properties of apical dendrites it has been suggested, that a second integration site in the apical dendrite may mediate specific global information. Here we explore this issue considering the example of learning invariant responses by examining a network of spiking neurones with two sites of synaptic integration. We demonstrate that results obtained in networks of units with continuous outputs transfer to the more realistic neuronal model. This allows a number of more specific experimental predictions, and is a necessary step to unified description of learning rules exploiting timing of action potentials.

Action Potentials↗
PubMed1994

Spatial planning deficits in limb apraxia.

Geschwind (1975) proposed a disconnection model in which an apraxic subject is unable to carry out movements to command because the left hemisphere that comprehended the verbal command is disconnected from the right premotor and motor areas which controls the left hand. An alternate model, however, proposes that apraxia results from destruction of spatiotemporal representations of learned movement stored in the left hemisphere (Heilman, 1979). The disconnection hypothesis would predict that apraxic subjects should be able to correctly imitate gestures and correctly use actual tools since these tasks do not require language. The movement representation model predicts that imitation and actual tool use would also be impaired. Motion analyses were performed on the trajectories of repetitive 'slicing' gestures made in a series of conditions in which contextual cues were introduced in a graded fashion. Four cue conditions were presented: no cues (verbal command), object present, tool present and both object and tool present. Positions of the hand, wrist, elbow, and shoulder were digitized from neighbouring views, reconstructed in three dimensions and analysed with respect to specific spatiotemporal features of the trajectories. Three subjects with limb apraxia, who had lesions that included left parietal cortex, and four neurologically intact subjects participated. The apraxic subjects showed disturbances in planning the movement of the hand in space across the cue conditions. For example, they showed deficits in the plane of motion, the shape of the trajectory and in the coupling of hand speed and trajectory shape even when given full contextual cues. These data support the hypothesis that apraxia can result from the destruction of spatiotemporal representations of learned movement, rather than from a disconnection between the receptive language areas in the left hemisphere and the contralateral motor cortices.

Aged↗
PubMed2005

Amount of training effects in representation-mediated food aversion learning: no evidence of a role for associability changes.

Rats acquired aversions to food pellets when a previously trained signal for that food was paired with a toxin, but only after minimal signal-food training. After extensive signal-food training, signal-toxin pairings had no effect on food consumption even after manipulations that enhanced the associability of the signal. By contrast, conditioned responding to the signal retained its sensitivity to devaluation of the food reinforcer by food-toxin pairings after extensive training. These results suggest that the nature of associatively activated event representations changes over the course of training.

Animals↗
PubMedDate not supplied

Misconceptions in Van den Broeck's representation of misconceptions about learning disability research.

The methods and conclusions of Van den Broeck (in this issue) are evaluated from two perspectives: (a) statistical considerations and (b) theoretical models of IQ and achievement, specifically reading achievement. We consider the statistical model proposed by Van den Broeck for the regression-based discrepancy model (RDM) to be either irrelevant or conceptually inconsistent with current models of IQ and achievement. The resulting simulation produced exemplar cases that are not realistic in terms of practice. The theoretical representations of IQ and achievement were, in our understanding, inconsistent with contemporary models of either. We suggest that acceptable models support the use of the RDM as it has been proposed by us and by others as one component of the determination of the presence or absence of a learning disability.

Child↗
PubMed2005

Learning viewpoint invariant object representations using a temporal coherence principle.

Invariant object recognition is arguably one of the major challenges for contemporary machine vision systems. In contrast, the mammalian visual system performs this task virtually effortlessly. How can we exploit our knowledge on the biological system to improve artificial systems? Our understanding of the mammalian early visual system has been augmented by the discovery that general coding principles could explain many aspects of neuronal response properties. How can such schemes be transferred to system level performance? In the present study we train cells on a particular variant of the general principle of temporal coherence, the "stability" objective. These cells are trained on unlabeled real-world images without a teaching signal. We show that after training, the cells form a representation that is largely independent of the viewpoint from which the stimulus is looked at. This finding includes generalization to previously unseen viewpoints. The achieved representation is better suited for view-point invariant object classification than the cells' input patterns. This property to facilitate view-point invariant classification is maintained even if training and classification take place in the presence of an--also unlabeled--distractor object. In summary, here we show that unsupervised learning using a general coding principle facilitates the classification of real-world objects, that are not segmented from the background and undergo complex, non-isomorphic, transformations.

Action Potentials↗
PubMed1989

Further studies of hippocampal representation during odor discrimination learning.

The contribution of hippocampal and nonhippocampal memory processing to simultaneous-cue odor discrimination learning was assessed. In this task, rats with hippocampal system damage consequent to fornix lesions (fornix rats) were severely and persistently impaired in discrimination learning, acquisition of learning set, and concurrent discrimination, although they occasionally solved some problems at a normal rate. By using those problems on which fornix rats succeeded, to permit comparisons of performance strategies with normal rats, differences between groups were shown on response latency measures and on probe trials involving the novel pairing of familiar odors. Normal rats had a bimodal distribution of response latencies, and their latency depended on where the S+ was presented. Fornix rats had short response latencies and responded equally quickly wherever the S+ was presented. Furthermore, when the representation of familiar S+ and S- odor pairs was challenged in probe trials, normal rats responded appropriately to the correct stimulus, whereas fornix rats behaved as if presented with a new odor pair. These data provide support for the view that the hippocampus participates in the representation of relations among odor (and other) stimuli and among other experiences and that it permits the flexible use of these representations in new contexts. In contrast, memory processing outside the hippocampal system can represent only the significance of individual stimuli and can be revealed only in a repetition of the original learning event.

Animals↗
PubMed2002

Statistical learning of new visual feature combinations by infants.

The ability of humans to recognize a nearly unlimited number of unique visual objects must be based on a robust and efficient learning mechanism that extracts complex visual features from the environment. To determine whether statistically optimal representations of scenes are formed during early development, we used a habituation paradigm with 9-month-old infants and found that, by mere observation of multielement scenes, they become sensitive to the underlying statistical structure of those scenes. After exposure to a large number of scenes, infants paid more attention not only to element pairs that cooccurred more often as embedded elements in the scenes than other pairs, but also to pairs that had higher predictability (conditional probability) between the elements of the pair. These findings suggest that, similar to lower-level visual representations, infants learn higher-order visual features based on the statistical coherence of elements within the scenes, thereby allowing them to develop an efficient representation for further associative learning.

Association Learning↗
PubMed1994

Reversible inactivation of the lateral dorsal thalamus disrupts hippocampal place representation and impairs spatial learning.

Place-specific discharge of hippocampal cells was monitored while rats performed daily 15 trials of a spatial memory task. During the intertrial interval between trials 5 and 6, the lateral dorsal nucleus of the thalamus (LDN) was reversibly inactivated. Choice accuracy on the maze became impaired, and many hippocampal place fields became disrupted. These data support the proposition that the LDN passes onto hippocampus important (spatial) information that is used for accurate maze navigation.

Animals↗
PubMed2005

Encoding of learned importance of sound by magnitude of representational area in primary auditory cortex.

We hypothesized that learning-induced representational expansion in the primary auditory cortex (AI) directly encodes the degree of behavioral importance of a sound. Rats trained on an operant auditory conditioning task were variably motivated to the conditioned stimulus (CS) through different levels of water deprivation. Mean performance values correlated with deprivation level, validating them as a measure of the overall control and, therefore, behavioral importance of the CS. Electrophysiological mapping revealed expanded representations of the CS, compared with other frequencies in experimental subjects, but not in naive or visually trained controls that received noncontingent CS tones. Importantly, representational area showed a significant positive correlation with mean performance levels for only the CS band, with significant effects for relative area in contrast to only modest changes in absolute area. CS representational expansion was asymmetric into high-frequency zones, thus performance level also was significantly correlated with the relative anterior-posterior location of the enlarged representation. An increased representation of low frequencies, related to the acoustic spectrum of the reward delivery equipment, also was discovered in both experimental and control trained subjects, supporting the conclusion that behaviorally important sounds gain representational area. Furthermore, there was a surprising reduction in total AI area for the experimental and control groups, compared with untrained naive subjects, indicating that the functional dimensions of AI are not fixed. Overall, the findings support the encoding of acquired stimulus importance based on representational size in AI.

Acoustic Stimulation↗
PubMed1994

Adaptive representation of dynamics during learning of a motor task.

We investigated how the CNS learns to control movements in different dynamical conditions, and how this learned behavior is represented. In particular, we considered the task of making reaching movements in the presence of externally imposed forces from a mechanical environment. This environment was a force field produced by a robot manipulandum, and the subjects made reaching movements while holding the end-effector of this manipulandum. Since the force field significantly changed the dynamics of the task, subjects' initial movements in the force field were grossly distorted compared to their movements in free space. However, with practice, hand trajectories in the force field converged to a path very similar to that observed in free space. This indicated that for reaching movements, there was a kinematic plan independent of dynamical conditions. The recovery of performance within the changed mechanical environment is motor adaptation. In order to investigate the mechanism underlying this adaptation, we considered the response to the sudden removal of the field after a training phase. The resulting trajectories, named aftereffects, were approximately mirror images of those that were observed when the subjects were initially exposed to the field. This suggested that the motor controller was gradually composing a model of the force field, a model that the nervous system used to predict and compensate for the forces imposed by the environment. In order to explore the structure of the model, we investigated whether adaptation to a force field, as presented in a small region, led to aftereffects in other regions of the workspace. We found that indeed there were aftereffects in workspace regions where no exposure to the field had taken place; that is, there was transfer beyond the boundary of the training data. This observation rules out the hypothesis that the subject's model of the force field was constructed as a narrow association between visited states and experienced forces; that is, adaptation was not via composition of a look-up table. In contrast, subjects modeled the force field by a combination of computational elements whose output was broadly tuned across the motor state space. These elements formed a model that extrapolated to outside the training region in a coordinate system similar to that of the joints and muscles rather than end-point forces. This geometric property suggests that the elements of the adaptive process represent dynamics of a motor task in terms of the intrinsic coordinate system of the sensors and actuators.

Adaptation, Physiological↗
PubMed2006

Elemental representations of stimuli in associative learning.

This article reviews evidence and theories concerning the nature of stimulus representations in Pavlovian conditioning. It focuses on the elemental approach developed in stimulus sampling theory (R. C. Atkinson & W. K. Estes, 1963; R. R. Bush & F. Mosteller, 1951b) and extended by I. P. L. McLaren and N. J. Mackintosh (2000, 2002) and contrasts this with models that invoke notions of configural representations that uniquely code for different patterns of stimulus inputs (e.g., J. M. Pearce, 1987, 1994; R. A. Rescorla & A. R. Wagner, 1972; A. R. Wagner & S. E. Brandon, 2001). The article then presents a new elemental model that emphasizes interactions between stimulus elements. This model is shown to explain a range of behavioral findings, including those (e.g., negative patterning and biconditional discriminations) traditionally thought to be beyond the explanatory capabilities of elemental models. Moreover, the model offers a ready explanation for recent findings reported by R. A. Rescorla (2000, 2001, 2002b) concerning the way that stimuli with different conditioning histories acquire associative strength when conditioned in compound.

Association Learning↗
PubMed1990

Forming sparse representations by local anti-Hebbian learning.

How does the brain form a useful representation of its environment? It is shown here that a layer of simple Hebbian units connected by modifiable anti-Hebbian feed-back connections can learn to code a set of patterns in such a way that statistical dependency between the elements of the representation is reduced, while information is preserved. The resulting code is sparse, which is favourable if it is to be used as input to a subsequent supervised associative layer. The operation of the network is demonstrated on two simple problems.

Brain↗
PubMed1998

Learning pop-out detection: building representations for conflicting target-distractor relationships.

Studies of perceptual learning consistently found that improvement is stimulus specific. These findings were interpreted as indicating an early cortical learning site. In line with this interpretation, we consider two alternative hypotheses: the 'earliest modification' and the 'output-level modification' assumptions, which respectively assume that learning occurs within the earliest representation which is selective for the trained stimuli, or at cortical levels receiving its output. We studied performance in a pop-out task using light bar distractor elements of one orientation, and a target element rotated by 30 degrees (or 90 degrees). We tested the alternative hypotheses by examining pop-out learning through an initial training phase, a subsequent learning stage with swapped target and distracted orientations, and a final re-test with the originally trained stimuli. We found learning does not transfer across orientation swapping. However, following training with swapped orientations, a similar performance level is reached as with original orientations. That is, learning neither facilitates nor interferes to a substantial degree with subsequent performance with altered stimuli. Furthermore, this re-training does not hamper performance with the original trained stimuli. If training changed the earliest orientation selective representation (specializing it for performance of the particular performed task) it would necessarily affect performance with swapped orientations, as well. The co-existence of similar asymptotes for apparently conflicting stimulus sets refutes the 'earliest modification' hypothesis, supporting the alternative 'output level modification' hypothesis. We conclude that secondary cortical processing levels use outputs from the earliest orientation representation to compute higher order structures, promoting and improving successful task performance.

Adult↗
PubMed1993

Constructive induction and protein tertiary structure prediction.

To date, the only methods that have been used successfully to predict protein structures have been based on identifying homologous proteins whose structures are known. However, such methods are limited by the fact that some proteins have similar structure but no significant sequence homology. We consider two ways of applying machine learning to facilitate protein structure prediction. We argue that a straightforward approach will not be able to improve the accuracy of classification achieved by clustering by alignment scores alone. In contrast, we present a novel constructive induction approach that learns better representations of amino acid sequences in terms of physical and chemical properties. Our learning method combines knowledge and search to shift the representation of sequences so that semantic similarity is more easily recognized by syntactic matching. Our approach promises not only to find new structural relationships among protein sequences, but also expands our understanding of the roles knowledge can play in learning via experience in this challenging domain.

Artificial Intelligence↗
PubMed2003

Distinct brain representations for early and late learned words.

Recently there has been a renewed interest in cognitive psychology on the effects of the age of word acquisition (AoA) on lexical processing. In particular, it is currently unclear whether AoA or word frequency are better predictors of word recognition. To date no study has investigated the neural bases of the AoA effect or attempted to dissociate it from word frequency. We report a visual and an auditory event-related fMRI experiment investigating the influence of AoA and word frequency on neural activity, and show that AoA modulates brain areas that are not influenced by word frequency. The precuneus was activated for early learned words across auditory and visual presentation modalities. Additional activity in the auditory cortex was observed specifically for the reading of early acquired words. Late learned words, in contrast, led to a selective activation increase in lateral inferior frontal areas. These findings support models that suggest that early and late learned words are represented differently in the brain. They further allow to specify the nature of the representational differences, namely that early learned words are represented in the brain in a more sensory manner than late learned words.

Adolescent↗
PubMed2003

Learned changes in the sensitivity of stimulus representations: associative and nonassociative mechanisms.

Central to associative learning theory is the proposal that the concurrent activation of a pair of event representations will establish or strengthen a link between them. Associative theorists have devoted much energy to establishing what representations are involved in any given learning paradigm and the rules that determine the degree to which the link is strengthened. They have paid less attention to the question of what determines that a representation will be activated, assuming, for the case of classical conditioning, that presentation of an appropriately intense stimulus from an appropriate modality will be enough. But this assumption is unjustified. I present the results of experiments on the effects of stimulus exposure in rats that suggest that mere exposure to a stimulus can influence its perceptual effectiveness -- that the ability of a stimulus to activate its representation can be changed by experience. This conclusion is of interest for two reasons. First, it supplies a direct explanation for the phenomenon of perceptual learning -- the enhancement of stimulus discriminability produced by some forms of stimulus exposure. Second, it poses a theoretical challenge in that it seems to require the existence of a learning mechanism outside the scope of those envisaged by current formal theories of associative learning. I offer some speculations as to how this mechanism might be incorporated into such theories.

Animals↗
PubMedDate not supplied

A layered architecture for computer-based simulation supporting skills learning: an X-ray imaging paradigm.

Simulation is characterized by strong learning potential, providing the basis for a new category of systems, the simulation-based learning systems. To strengthen the learning potential of these systems, models are needed not only of the actual system being imitated, but also of the operational expertise required to carry out manipulations of the simulated system, inherently linked to learning. In this paper, an architecture is reported aimed at supporting the organization of multimodal simulation resources to induce skills learning. This architecture is based on distinct layers, allowing independent representation of learning and simulation components. Its applicability has been demonstrated by means of a paradigm, including simulation of X-ray imaging procedure, as well as authoring of learning scenarios pertaining to such procedures.

Algorithms↗
PubMed2026

A quantitative coordinate system for developmental dynamics.

Quantitative comparison of morphogenesis across individuals remains a fundamental challenge, as developing embryos vary in shape, orientation and developmental tempo. Moreover, real-time three-dimensional imaging generates large, heterogeneous four-dimensional datasets that are difficult to directly align. As a result, developmental variability is typically described qualitatively rather than measured. Here we introduce STERN, a quantitative framework that learns continuous spatiotemporal representations of morphogenesis directly from in vivo 4D imaging data. By embedding embryos into a shared spatiotemporal space, STERN defines a quantitative developmental coordinate system that enables direct comparison of developmental trajectories across individuals without requiring explicit registration or staging. Applied to mouse embryogenesis, STERN reveals that embryos follow conserved developmental trajectories while progressing at distinct temporal rates, providing a quantitative measure of developmental heterochrony. Extending this framework to zebrafish neural crest light-sheet timelapse imaging, we further show that developmental order is preserved across distinct imaging views even with altered anatomical coverage, supporting the generality of the learned representation across vertebrate imaging contexts. Finally, in developing mouse hearts, where morphogenesis proceeds through subtle and continuously evolving structural changes, STERN resolves fine-scale developmental dynamics at minute-scale temporal resolution that are difficult to localize reproducibly using human experts or general-purpose multimodal AI. Together, these results establish a shared quantitative coordinate system for morphogenesis, in which developmental trajectories become directly comparable across individuals and developmental variability becomes a measurable property.

Journal Article↗