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 235 records · Page 13Linked to original sources
PubMed2002

Integration and representation in rats' serial pattern learning in the T-maze.

Rats were exposed to three-trial series consisting of reinforced (R) trials and one nonreinforced (N) trial in a fixed order, RRN and RNR (Experiments 1 and 2) or NRR and RRN (Experiment 3), on extended visually distinct runways in a T-maze. When initially presented with the same sequence on each series in a session (separate presentations) with the same runway on all trials within a series (Experiments 1 and 3), all the rats developed slower running speeds on N than on R trials. When a runway was sometimes changed between the first and next two trials during separate presentations training (Experiment 2) or both sequences were later intermixed within each session in each experiment, only rats exposed to each sequence on a specific runway maintained these serial running patterns. Rats displayed serial running patterns on a test RNN sequence similar to that on the RNR sequence (Experiment 2), as would be predicted by an intertrial association model of serial pattern learning (Capaldi & Molina, 1979), but responded on test RRR and NRN sequences (Experiment 3) as would be predicted by an ordinal-trial-tag/intratrial association model (Burns, Wiley, & Payne, 1986). Results from test series of free-choice trials in Experiments 1 and 2 failed to support a prediction of the intratrial association model that these rats would integrate RRN and RNR sequences. Rather than always selecting a baited runway on both the second and the third free-choice trials, the rats only selected a baited runway on the third trial on the basis of their choice on the second trial, as would be predicted by the intertrial association model. Only after experiencing all possible outcome sequences during forced-choice training in Experiment 3 did these rats predominantly select a baited runway on every free-choice trial.

Animals↗
PubMedDate not supplied

[The neural bases of taste aversion learning: the formation of acquired hedonic taste representations].

INTRODUCTION: Taste aversion conditioning is a form of associative learning in which certain qualities of a food (mainly its taste) are associated to specific negative visceral consequences that derive from eating it. Establishing this learning depends on gustatory-visceral integration processes carried out in the central nervous system. DEVELOPMENT: In this manuscript our aim is to offer a global view of the centres and connections that play the most significant roles in the formation of taste aversion learning (TAL). CONCLUSIONS: Many researchers consider that the initial level of integration is situated within the parabrachial nuclei. A priori and given the basic vital nature of TAL, its formation and completion could be thought to take place at this brain stem level, without requiring the intervention of the higher processing structures. Nevertheless, in the literature on TAL there is a large body of both neuroanatomical and neurobehavioural evidence that seems to indicate that the formation of TAL requires complex interactions between the parabrachial nuclei and certain prosencephalic structures, such as the insular cortex or the amygdala, among others.

Amygdala↗
PubMed1997

A neural model of multimodal adaptive saccadic eye movement control by superior colliculus.

How does the saccadic movement system select a target when visual, auditory, and planned movement commands differ? How do retinal, head-centered, and motor error coordinates interact during the selection process? Recent data on superior colliculus (SC) reveal a spreading wave of activation across buildup cells the peak activity of which covaries with the current gaze error. In contrast, the locus of peak activity remains constant at burst cells, whereas their activity level decays with residual gaze error. A neural model answers these questions and simulates burst and buildup responses in visual, overlap, memory, and gap tasks. The model also simulates data on multimodal enhancement and suppression of activity in the deeper SC layers and suggests a functional role for NMDA receptors in this region. In particular, the model suggests how auditory and planned saccadic target positions become aligned and compete with visually reactive target positions to select a movement command. For this to occur, a transformation between auditory and planned head-centered representations and a retinotopic target representation is learned. Burst cells in the model generate teaching signals to the spreading wave layer. Spreading waves are produced by corollary discharges that render planned and visually reactive targets dimensionally consistent and enable them to compete for attention to generate a movement command in motor error coordinates. The attentional selection process also helps to stabilize the map-learning process. The model functionally interprets cells in the superior colliculus, frontal eye field, parietal cortex, mesencephalic reticular formation, paramedian pontine reticular formation, and substantia nigra pars reticulata.

Animals↗
PubMed2004

Infants can rapidly form new categorical representations.

Young infants learn common categorical distinctions, such as animals versus vehicles. But can they, like adults, rapidly form new categories, such as black-and-white animals? To answer this question, 6-, 10-, and 13-month-old infants were familiarized with four land animals that were black and white in coloring (e.g., a zebra and a black-and-white tiger) and then were tested with novel animals and a truck. The infants responded to an exclusive category that apparently included only black-and-white animals, suggesting that they formed a new categorical representation during familiarization. A comparison group of infants familiarized with a set of land animals that were more variable in coloring (e.g., a pale yellow horse and a yellow-and-brown tiger) formed a very general categorical representation that included many different kinds of animals, regardless of coloring. Therefore, like adults, infants rapidly form new categorical representations in response to the context.

Child Development↗
PubMed2006

A computational theory of hippocampal function, and empirical tests of the theory.

The main aim of the paper is to present an up-to-date computational theory of hippocampal function and the predictions it makes about the different subregions (dentate gyrus, CA3 and CA1), and to examine behavioral and electrophysiological data that address the functions of the hippocampus and particularly its subregions. Based on the computational proposal that the dentate gyrus produces sparse representations by competitive learning and via the mossy fiber pathway forces new representations on the CA3 during learning (encoding), it has been shown behaviorally that the dentate gyrus supports spatial pattern separation during learning. Based on the computational proposal that CA3-CA3 autoassociative networks are important for episodic memory, it has been shown behaviorally that the CA3 supports spatial rapid one-trial learning, learning of arbitrary associations where space is a component, pattern completion, spatial short-term memory, and sequence learning by associations formed between successive items. The concept that the CA1 recodes information from CA3 and sets up associatively learned backprojections to neocortex to allow subsequent retrieval of information to neocortex, is consistent with findings on consolidation. Behaviorally, the CA1 is implicated in processing temporal information as shown by investigations requiring temporal order pattern separation and associations across time; computationally this could involve temporal decay memory, and temporal sequence memory which might also require CA3. The perforant path input to DG is implicated in learning, to CA3 in retrieval from CA3, and to CA1 in retrieval after longer time intervals ("intermediate-term memory").

Animals↗
PubMedDate not supplied

Neural substrates of response-based sequence learning using fMRI.

Representation of sequential structure can occur with respect to the order of perceptual events or the order in which actions are linked. Neural correlates of sequence retrieval associated with the order of motor responses were identified in a variant of the serial reaction time task in which training occurred with a spatially incompatible mapping between stimuli and finger responses. After transfer to a spatially compatible version of the task, performance enhancements indicative of learning were only present in subjects required to make finger movements in the same order used during training. In contrast, a second group of subjects performed the compatible task using an identical sequence of stimuli (and different order of finger movements) as in training. They demonstrated no performance benefit, indicating that learning was response based. Analysis was restricted to subjects demonstrating low recall of the sequence structure to rule out effects of explicit awareness. The interaction of group (motor vs. perceptual transfer) with sequence retrieval (sequencing vs. rest) revealed significantly greater activation in the bilateral supplementary motor area, cingulate motor area, ventral premotor cortex, left caudate, and inferior parietal lobule for subjects in the motor group (illustrating successful sequence retrieval at the response level). Retrieval of sequential responses occurs within mesial motor areas and related motor planning areas.

Adolescent↗
PubMed2000

Effects of repetitive motor training on movement representations in adult squirrel monkeys: role of use versus learning.

Current evidence indicates that repetitive motor behavior during motor learning paradigms can produce changes in representational organization in motor cortex. In a previous study, we trained adult squirrel monkeys on a repetitive motor task that required the retrieval of food pellets from a small-diameter well. It was found that training produced consistent task-related changes in movement representations in primary motor cortex (M1) in conjunction with the acquisition of a new motor skill. In the present study, we trained adult squirrel monkeys on a similar motor task that required pellet retrievals from a much larger diameter well. This large-well retrieval task was designed to produce repetitive use of a limited set of distal forelimb movements in the absence of motor skill acquisition. Motor activity levels, estimated by the total number of finger flexions performed during training, were matched between the two training groups. This experiment was intended to evaluate whether simple, repetitive motor activity alone is sufficient to produce representational plasticity in cortical motor maps. Detailed analysis of the motor behavior of the monkeys indicates that their retrieval behavior was highly successful and stereotypical throughout the training period, suggesting that no new motor skills were learned during the performance of the large-well retrieval task. Comparisons between pretraining and posttraining maps of M1 movement representations revealed no task-related changes in the cortical area devoted to individual distal forelimb movement representations. We conclude that repetitive motor activity alone does not produce functional reorganization of cortical maps. Instead, we propose that motor skill acquisition, or motor learning, is a prerequisite factor in driving representational plasticity in M1.

Animals↗
PubMed2006

Statistical learning within and between modalities: pitting abstract against stimulus-specific representations.

When learners encode sequential patterns and generalize their knowledge to novel instances, are they relying on abstract or stimulus-specific representations? Research on artificial grammar learning (AGL) has shown transfer of learning from one stimulus set to another, and such findings have encouraged the view that statistical learning is mediated by abstract representations that are independent of the sense modality or perceptual features of the stimuli. Using a novel modification of the standard AGL paradigm, we obtained data to the contrary. These experiments pitted abstract processing against stimulus-specific learning. The findings show that statistical learning results in knowledge that is stimulus-specific rather than abstract. They show furthermore that learning can proceed in parallel for multiple input streams along separate perceptual dimensions or sense modalities. We conclude that learning sequential structure and generalizing to novel stimuli inherently involve learning mechanisms that are closely tied to the perceptual characteristics of the input.

Association Learning↗
PubMed2002

Preintegration lateral inhibition enhances unsupervised learning.

A large and influential class of neural network architectures uses postintegration lateral inhibition as a mechanism for competition. We argue that these algorithms are computationally deficient in that they fail to generate, or learn, appropriate perceptual representations under certain circumstances. An alternative neural network architecture is presented here in which nodes compete for the right to receive inputs rather than for the right to generate outputs. This form of competition, implemented through preintegration lateral inhibition, does provide appropriate coding properties and can be used to learn such representations efficiently. Furthermore, this architecture is consistent with both neuroanatomical and neurophysiological data. We thus argue that preintegration lateral inhibition has computational advantages over conventional neural network architectures while remaining equally biologically plausible.

Algorithms↗
PubMed1997

Trading spaces: computation, representation, and the limits of uninformed learning.

Some regularities enjoy only an attenuated existence in a body of training data. These are regularities whose statistical visibility depends on some systematic recoding of the data. The space of possible recoding is, however, infinitely large--it is the space of applicable Turing machines. As a result, mappings that pivot on such attenuated regularities cannot, in general, be found by brute-force search. The class of problems that present such mappings we call the class of "type-2 problems." Type-1 problems, by contrast, present tractable problems of search insofar as the relevant regularities can be found by sampling the input data as originally coded. Type-2 problems, we suggest, present neither rare nor pathological cases. They are rife in biologically realistic settings and in domains ranging from simple animat (simulated animal or autonomous robot) behaviors to language acquisition. Not only are such problems rife--they are standardly solved! This presents a puzzle. How, given the statistical intractability of these type-2 cases, does nature turn the trick? One answer, which we do not pursue, is to suppose that evolution gifts us with exactly the right set of recording biases so as to reduce specific type-2 problems to (tractable) type-1 mappings. Such a heavy-duty nativism is no doubt sometimes plausible. But we believe there are other, more general mechanisms also at work. Such mechanisms provide general (not task-specific) strategies for managing problems of type-2 complexity. Several such mechanisms are investigated. At the heart of each is a fundamental poly--namely, the maximal exploitation of states of representation already achieved by prior, simpler (type-1) learning so as to reduce the amount of subsequent computational search. Such exploitation both characterizes and helps make unitary sense of a diverse range of mechanisms. These include simple incremental learning (Elman 1993), modular connectionism (Jacobs et al. 1991), and the developmental hypothesis of "representational redescription" (Karmiloff-Smith 1979; 1992). In addition, the most distinctive features of human cognition--language and culture--may themselves be viewed as adaptations enabling this representation/computation trade-off to be pursued on an even grander scale.

Humans↗
PubMed2001

Sensitivity of cortical movement representations to motor experience: evidence that skill learning but not strength training induces cortical reorganization.

The topography of forelimb movement representations within the rat motor cortex was examined following forelimb strength training. Adult male rats were allocated to either a Power Reaching, Control Reaching or Non-Reaching Condition. Power Reaching rats were trained to grasp and break progressively larger bundles of dried pasta strands with their preferred forelimb. Control Reaching animals were trained to break a single pasta strand and Non-Reaching animals were not trained. Power Reaching animals exhibited a progressive increase in the maximal size of the pasta bundle that could be retrieved during a 30-day training period. Kinematic analyses showed that this improvement was not due to a change in reaching strategy. Intracortical microelectrode stimulation was used to derive maps of forelimb movement representations within the motor cortex of all animals following training. In comparison to Non-Reaching animals, both Power Reaching and Control Reaching animals exhibited a significant increase in the proportion of motor cortex occupied by distal forelimb movement representations (wrist/digit) and a decrease in the proportion of proximal representations (elbow/shoulder). These results demonstrate that the development of skilled forelimb movements, but not increased forelimb strength, was associated with a reorganization of forelimb movement representations within motor cortex.

Animals↗
PubMed2004

Cognitive navigation based on nonuniform Gabor space sampling, unsupervised growing networks, and reinforcement learning.

We study spatial learning and navigation for autonomous agents. A state space representation is constructed by unsupervised Hebbian learning during exploration. As a result of learning, a representation of the continuous two-dimensional (2-D) manifold in the high-dimensional input space is found. The representation consists of a population of localized overlapping place fields covering the 2-D space densely and uniformly. This space coding is comparable to the representation provided by hippocampal place cells in rats. Place fields are learned by extracting spatio-temporal properties of the environment from sensory inputs. The visual scene is modeled using the responses of modified Gabor filters placed at the nodes of a sparse Log-polar graph. Visual sensory aliasing is eliminated by taking into account self-motion signals via path integration. This solves the hidden state problem and provides a suitable representation for applying reinforcement learning in continuous space for action selection. A temporal-difference prediction scheme is used to learn sensorimotor mappings to perform goal-oriented navigation. Population vector coding is employed to interpret ensemble neural activity. The model is validated on a mobile Khepera miniature robot.

Cognition↗
PubMed1988

Learning the hidden structure of speech.

In the work described here, the backpropagation neural network learning procedure is applied to the analysis and recognition of speech. This procedure takes a set of input/output pattern pairs and attempts to learn their functional relationship; it develops the necessary representational features during the course of learning. A series of computer simulation studies was carried out to assess the ability of these networks to accurately label sounds, to learn to recognize sounds without labels, and to learn feature representations of continuous speech. These studies demonstrated that the networks can learn to label presegmented test tokens with accuracies of up to 95%. Networks trained on segmented sounds using a strategy that requires no external labels were able to recognize and delineate sounds in continuous speech. These networks developed rich internal representations that included units which corresponded to such traditional distinctions as vowels and consonants, as well as units that were sensitive to novel and nonstandard features. Networks trained on a large corpus of unsegmented, continuous speech without labels also developed interesting feature representations, which may be useful in both segmentation and label learning. The results of these studies, while preliminary, demonstrate that backpropagation learning can be used with complex, natural data to identify a feature structure that can serve as the basis for both analysis and nontrivial pattern recognition.

Computer Simulation↗
PubMed1998

Alignment and category learning.

Recent research shows that similarity comparisons involve an alignment process in which features are placed into correspondence. In 6 studies, the authors showed that alignment is involved in category learning as well. Within a category, aligned matches (feature matches occurring on the same dimension) facilitate learning more than nonaligned matches do (matches on different dimensions), although nonaligned matches still facilitate learning relative to nonmatches. Analogously, feature matches that cross category boundaries hurt learning more if they occur on the same versus a different dimension, and cross-category feature matches on different dimensions hurt learning relative to nonmatching features. Representational assumptions of category learning models must be modified to account for the differences between aligned and nonaligned feature matches.

Humans↗
PubMed1999

Hippocampus and context in classical conditioning.

Recent evidence suggests that contextual learning encompasses a variety of changes in learning and performance processes. Only some of these changes depend on the hippocampus. Specialized functions proposed for the hippocampus in contextual learning include the construction and consolidation of contextual memory representations, incidental contextual learning, and inhibitory contextual learning.

Animals↗
PubMed2001

Relational learning with and without awareness: transitive inference using nonverbal stimuli in humans.

Learning complex relationships among items and representing them flexibly have been shown to be highly similar in function and structure to conscious forms of learning. However, it is unclear whether conscious learning is essential for the exhibition of flexibility in learning. Successful performance on the transitive inference task requires representational flexibility. Participants learned four overlapping premise pairs (A > B, B > C, C > D, D > E) that could be encoded separately or as a sequential hierarchy (A > B > C > D > E). Some participants (informed) were told prior to training that the task required an inference made from premise pairs. Other participants (uninformed) were told simply that they were to learn a series of pairs by trial and error. Testing consisted of unreinforced trials that included the non-adjacent pair, B versus D, to assess capacity for transitive inference. Not surprisingly, those in the informed condition outperformed those in the uninformed condition. After completion of training and testing, uninformed participants were given a postexperimental questionnaire to assess awareness of the task structure. In contrast with expectations, successful performance on the transitive inference task for uninformed participants does not depend on or correlate with postexperimental awareness. The present results suggest that relational learning tasks do not necessarily require conscious processes.

Adult↗
PubMed2026

NanoSSL: attention mechanism-based self-supervised learning method for protein identification using nanopores.

MOTIVATION: Nanopores are cutting-edge interdisciplinary tools that can analyze biomolecules at the single-molecule level for many applications, e.g. DNA sequencing. Efforts are underway to extend nanopores to proteomics, including the development of machine learning algorithms for protein sequencing and identification. However, single-molecule data are intrinsically noisy and hard to process. Moreover, the development and performance of machine learning for nanopore is jeopardized by data scarcity. Self-supervised learning is an emerging method that may yield advantages in nanopore scenarios. RESULTS: We propose and experimentally validate Nanopore analysis using Self-Supervised Learning (NanoSSL), a generative self-supervised learning framework based on attention mechanisms for the identification of protein signals from nanopores. Leveraging a two-step approach consisting of self-supervised pre-training and supervised fine-tuning, NanoSSL learns useful feature representations from empirical data to facilitate downstream classification tasks. Inspired by the concept of fragmentation in conventional protein sequencing technologies, during pretraining each translocation event is split into multiple non-overlapping fragments of equal size, some of which are randomly masked and reconstructed using a masked autoencoder. Learning the feature representations of the reconstructed nanopore events facilitates molecular identification in fine-tuning. In this study, we retested a publicly available nanopore multiplexed protein sensing dataset for model iteration, and subsequently measured Alzheimer's disease biomarker Aβ1-42 using homemade solid-state nanopores. Empirical results indicated NanoSSL achieved an unprecedented performance across four metrics: accuracy, precision, recall, and F1 score, when classifying two mutated Aβ1-42, E22G and G37R. The self-supervised learning and attention mechanism were verified as the source of performance gains. AVAILABILITY AND IMPLEMENTATION: The main program is available at https://doi.org/10.5281/zenodo.17172822.

Nanopores↗
PubMed1999

Orbitofrontal cortex and representation of incentive value in associative learning.

Clinical evidence indicates that damage to ventromedial prefrontal cortex disrupts goal-directed actions that are guided by motivational and emotional factors. As a consequence, patients with such damage characteristically engage in maladaptive behaviors. Other research has shown that neurons in the corresponding orbital region of prefrontal cortex in laboratory animals encode information regarding the incentive properties of goals or expected events. The present study investigates the effect of neurotoxic orbitofrontal cortex (OFC) lesions in the rat on responses that are normally influenced by associations between a conditioned stimulus (CS) and the incentive value of reinforcement. Rats were first trained to associate a visual CS with delivery of food pellets to a food cup. As a consequence of learning, rats approached the food cup during the CS in anticipation of reinforcement. In a second training phase, injection of LiCl followed consumption of the food unconditioned stimulus (US) in the home cage, a procedure used to alter the incentive value of the US. Subsequently, rats were returned to the conditioning chamber, and their responding to the CS in the absence of the food US was tested. Lesions of OFC did not affect either the initial acquisition of a conditioned response to the light CS in the first training phase or taste aversion learning in the second training phase. In the test for devaluation, however, OFC rats exhibited no change in conditioned responding to the visual CS. This outcome contrasts with the behavior of control rats; after devaluation of the US a significant decrease occurred in approach to the food cup during presentation of the CS. The results reveal an inability of a cue to access representational information about the incentive value of associated reinforcement after OFC damage.

Animals↗