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 1,495 records · Page 83Linked to original sources
PubMed1994

The evaluative response: primitive but necessary.

We do not believe the claim that "evaluative conditioning is a qualitatively distinct form of classical conditioning" [Davey (1994). Behaviour Research and Therapy, 32, 291-299]. We view all classical conditioning as a learning process which leads organisms to assign a positive or negative value to previously neutral stimuli and to respond to them accordingly. The evaluative response is a necessary component of this process and hence it is central to all classical conditioning, not a separate type. Davey concentrates on a signal-based information processing view of learning, i.e. the formation of linear associations between CS and UCS of which human subjects are aware and which they can verbalize. We propose a more primitive and more general model in which stimulus evaluation (like/dislike) occurs with a minimal degree of processing, and enters into a representation of stimulus (CS and UCS) and response (CR and UCR) characteristics which is reintegrative rather than associative.

Animals↗
PubMed2000

Reversible inactivation of the hippocampal mossy fiber synapses in mice impairs spatial learning, but neither consolidation nor memory retrieval, in the Morris navigation task.

The role played by hippocampal mossy fibers in the learning and memory processes implemented in the Morris swimming navigation task has been studied in C57BL/6 mice by selective and reversible inactivation of mossy fiber synaptic fields by diethyldithiocarbamate. The functional integrity of the mossy fibers proved essential for the storage of the spatial representation on the modifiable synapses of the recurrent collaterals of the CA3 pyramidal cells, whereas it is not necessary for the consolidation and recall of spatial memories. The results suggest that mossy fibers are preferentially involved in new learning. They are consistent with the hypothesis that the hippocampal CA3 region might act as an autoassociation memory.

Animals↗
PubMed2005

Cortical responses in rats predict perceptual sensitivities to complex sounds.

The common assumption that perceptual sensitivities are related to neural representations of sensory stimuli has seldom been directly demonstrated. The authors analyzed the similarity of spike trains evoked by complex sounds in the rat auditory cortex and related cortical responses to performance in an auditory task. Rats initially learned to identify 2 highly different periodic, frequency-modulated sounds and then were tested with increasingly similar sounds. Rats correctly classified most novel sounds; their accuracy was negatively correlated with acoustic similarity. Rats discriminated novel sounds with slower modulation more accurately than sounds with faster modulation. This asymmetry was consistent with similarities in cortical representations of the sounds, demonstrating that perceptual sensitivities to complex sounds can be predicted from the cortical responses they evoke.

Acoustics↗
PubMed2001

Improving text comprehension strategies in upper primary school children: a design experiment.

BACKGROUND: With respect to the acquisition of competence in reading, new standards for primary education stress more than before the importance of learning and teaching cognitive and metacognitive strategies that facilitate text comprehension. Therefore, there is a need to design a research-based instructional approach to strategic reading comprehension. AIMS: The design experiment aimed at developing, implementing and evaluating a research-based, but also practically applicable learning environment for enhancing skilled strategy use in upper primary school children when reading a text. Four text comprehension strategies (activating prior knowledge, clarifying difficult words, making a schematic representation of the text, and formulating the main idea) and a metacognitive strategy (regulating one's own reading process) were trained through a variety of highly interactive instructional techniques, namely modelling, whole class discussion, and small group work in the format of reciprocal teaching. SAMPLE: Participants in the study were four experimental 5th grade classes (79 children) and eight comparable control classes (149 pupils). METHOD: The effects of the learning environment were measured using a pretest-post-test-retention design. Multilevel hierarchical linear regression models were used to analyse the quantitative data of a Reading Strategy Test, a standardised Reading Comprehension Test, a Reading Attitude Scale, a Transfer Test and an interview about strategy use during reading. RESULTS: The data of the Reading Strategy Test, the Transfer Test and the interviews about strategy use showed that the experimental group out-performed the control group in terms of the strategy adoption and application during text reading. Whilst the experimental group also scored higher on the Reading Comprehension Test than the control group, the difference was not significant. CONCLUSIONS: This design experiment shows that it is possible to foster pupils' use and transfer of strategic reading comprehension skills in regular classrooms by immersing them in a powerful learning environment. But this intervention does not automatically result in improvement of performance on a standardised reading comprehension test.

Belgium↗
PubMed2001

Configural learning without configural training.

Rats learned discriminations in which 2 familiar compound visual stimuli AB and CD were nonreinforced, whereas less-familiar compounds (EF, GH, IJ...) were reinforced (constant-negative paradigm). Such discriminations can in principle be acquired through elemental (nonconfigural) learning. Three different types of compound--object/object, object/position, and shape/fill-pattern--were used, and the number of familiar compounds extended to 3 or 4. In all cases, when rats were tested for the first time with previously unseen recombinations of the familiar elements--for example, AC, BD--they preferred these to the familiar compounds, implying that they had encoded configural information during the original training. Moreover, preference remained constant despite extended exposure to configural training in the test phase. The findings extend the evidence that configural representations are formed independently of explicit configural training.

Animals↗
PubMed2005

A joint compression-discrimination neural transformation applied to target detection.

Many image recognition algorithms based on data-learning perform dimensionality reduction before the actual learning and classification because the high dimensionality of raw imagery would require enormous training sets to achieve satisfactory performance. A potential problem with this approach is that most dimensionality reduction techniques, such as principal component analysis (PCA), seek to maximize the representation of data variation into a small number of PCA components, without considering interclass discriminability. This paper presents a neural-network-based transformation that simultaneously seeks to provide dimensionality reduction and a high degree of discriminability by combining together the learning mechanism of a neural-network-based PCA and a backpropagation learning algorithm. The joint discrimination-compression algorithm is applied to infrared imagery to detect military vehicles.

Algorithms↗
PubMed1984

Ontogenesis of learning: IV. Dissociation of memory and perceptual-altering processes mediating taste neophobia in the rat.

The rat's neophobic response to tastes can be conceptualized as depending on two processes: memory processes that store a representation of an experienced taste so that it can be recognized as familiar on subsequent encounters, and perceptual-altering processes that respond differentially to novel versus familiar tastes. We investigated the ontogeny of these processes by studying the preweanling rat's behavioral reaction to a 10% sucrose solution. Experiments I and II suggest that the memory processes mature earlier (by 7-8 days of age) than the perceptual-altering processes (about 11 days of age). Experiment III suggests that pups do not acquire an aversion to sucrose paired with illness until they are 12-days-old, implying a close correspondence between the maturation of processes mediating neophobia and taste aversion learning. These findings are consistent with our previous work (Vogt & Rudy, 1984), and our view that ontogenetic dissociations in taste-guided behavior reflect caudal-to-rostral maturation in the ascending gustatory system.

Aging↗
PubMed1984

Timing, learning, and forgetting.

A computer-simulation model (AL) was developed for molecular and molar operant conditioning data. AL assumes a simple forgetting rule and all-or-nothing associations between memory representations of responses and reinforcers. AL describes acquisition and steady-state behavior in both concurrent interresponse time (IRT) schedules and in ordinary concurrent schedules. It describes functions relating the relative frequency of an IRT to IRT duration and to reinforcement frequency. AL accurately predicts for ordinary concurrent variable-interval schedules that the log ratio of responses to the two alternatives is a linear function of the log ratio of the reinforcers they deliver, and that the slope of this straight line approximates the degree of undermatching that characterizes the behavior of real animals. In addition, AL successfully describes the fact that in real data of Nevin, Heyman, and others, the probability of a changeover from one response to the other is roughly independent of the number of consecutive responses to the same alternative preceding the changeover. A previous version of AL assimilated data on temporal psychophysics, such as that animals bisect two temporal intervals at the geometric mean, and that the Weber fraction in DRL schedules is approximately constant. AL therefore describes transient and steady-state molecular and molar performances as well as data on temporal psychophysics.

Animals↗
PubMed1998

Influences of prior knowledge on selective weighting of category members.

Three experiments addressed how prior theories affect categorization, comparing the influence of theory-congruent versus theory-incongruent category members. Subjects observed descriptions of persons, some congruent with prior knowledge and some incongruent, then made transfer judgments. In Experiment 1, subjects were given a relatively long time to study each description, whereas in Experiment 2 study time was manipulated between subjects. In Experiment 3, learning was self-paced by each subject. It was found that, with enough study time, prior knowledge had 2 distinct influences. First, prior knowledge provided an initial representation, subsequently revised in light of new observations. Second, incongruent observations had more impact than congruent observations on categorization. In comparison, when study time was more limited, revision proceeded in a Bayesian manner, in that congruent and incongruent observations had equal impacts.

Adult↗
PubMed2003

Learning to control a brain-machine interface for reaching and grasping by primates.

Reaching and grasping in primates depend on the coordination of neural activity in large frontoparietal ensembles. Here we demonstrate that primates can learn to reach and grasp virtual objects by controlling a robot arm through a closed-loop brain-machine interface (BMIc) that uses multiple mathematical models to extract several motor parameters (i.e., hand position, velocity, gripping force, and the EMGs of multiple arm muscles) from the electrical activity of frontoparietal neuronal ensembles. As single neurons typically contribute to the encoding of several motor parameters, we observed that high BMIc accuracy required recording from large neuronal ensembles. Continuous BMIc operation by monkeys led to significant improvements in both model predictions and behavioral performance. Using visual feedback, monkeys succeeded in producing robot reach-and-grasp movements even when their arms did not move. Learning to operate the BMIc was paralleled by functional reorganization in multiple cortical areas, suggesting that the dynamic properties of the BMIc were incorporated into motor and sensory cortical representations.

Animals↗
PubMed2004

Neural mechanisms of vocal sequence generation in the songbird.

Little is known about the biophysical and neuronal circuit mechanisms underlying the generation and learning of behavioral sequences. Songbirds provide a marvelous animal model in which to study these phenomena. By use of a motorized microdrive to record the activity of single neurons in the singing bird, we are beginning to understand the circuits that generate complex vocal sequences. We describe recent experiments elucidating the role of premotor song-control nucleus HVC in the production of song. We find that HVC neurons projecting to premotor nucleus RA each generate a single burst of spikes at a particular time in the song and may form a sparse representation of temporal order. We incorporate this observation into a working hypothesis for the generation of vocal sequences in the songbird, and examine some implications for song learning.

Animals↗
PubMed2005

Modeling the influence of task on attention.

We propose a computational model for the task-specific guidance of visual attention in real-world scenes. Our model emphasizes four aspects that are important in biological vision: determining task-relevance of an entity, biasing attention for the low-level visual features of desired targets, recognizing these targets using the same low-level features, and incrementally building a visual map of task-relevance at every scene location. Given a task definition in the form of keywords, the model first determines and stores the task-relevant entities in working memory, using prior knowledge stored in long-term memory. It attempts to detect the most relevant entity by biasing its visual attention system with the entity's learned low-level features. It attends to the most salient location in the scene, and attempts to recognize the attended object through hierarchical matching against object representations stored in long-term memory. It updates its working memory with the task-relevance of the recognized entity and updates a topographic task-relevance map with the location and relevance of the recognized entity. The model is tested on three types of tasks: single-target detection in 343 natural and synthetic images, where biasing for the target accelerates target detection over twofold on average; sequential multiple-target detection in 28 natural images, where biasing, recognition, working memory and long term memory contribute to rapidly finding all targets; and learning a map of likely locations of cars from a video clip filmed while driving on a highway. The model's performance on search for single features and feature conjunctions is consistent with existing psychophysical data. These results of our biologically-motivated architecture suggest that the model may provide a reasonable approximation to many brain processes involved in complex task-driven visual behaviors.

Attention↗
PubMed1997

The flexible use of multiple cue relationships in spatial navigation: a comparison of water maze performance following hippocampal, medial septal, prefrontal cortex, or posterior parietal cortex lesions.

Rats prepared with lesions of the prefrontal cortex, posterior parietal cortex, hippocampus, or medial septal area were tested for acquisition of a number of variations of the open-field water maze using a version of place learning assessment described by Eichenbaum, Stewart, and Morris (1991). Specifically, the individual role of the aforementioned cortical and subcortical structures in tasks with differing representational demands on navigation were assessed. The results suggest that the sham-operated control, posterior parietal cortex-lesioned rats, and medial septal area-lesioned rats were able to navigate effectively under changing task conditions. Conversely, the navigational performances of the prefrontal cortex- and hippocampal formation-lesioned rats were impaired when task demands changed. The results are discussed in terms of the flexible use of multiple distal cues to guide navigation and the resulting loss of this flexibility after lesions to either the prefrontal cortex or the hippocampus.

Animals↗
PubMed1990

Experimental and theoretical results for some models of random-dot-pattern discrimination.

The popular, but low-power, two-stimulus signal-detection paradigms are extended to generate response-criterion-dependent psychometric functions. Maximumlikelihood estimation within a sufficiently powerful experimental design is described, and theoretical conditions for the resulting psychometric functions to be monotonically increasing are presented. These results are applied to an experiment on the detection of spatial-density differences in briefly presented random-dot patterns. The results confirm, extend, and complement current notions put forth by H.B. Barlow and his co-workers on mechanisms of internal spatial-density representation and the efficiency of intermediate levels of visual processing.

Attention↗
PubMed1991

The perception and metaperception of architectural designs communicated by video-computer imaging.

Non-architects answered multiple-choice questions about depth, size, shape, texture, and conformation in exterior building scenes presented by manual rendering or video-computer techniques including a single viewpoint, multiple viewpoints, and animation. They also rated (a) confidence in their answers, (b) overall judgments of clarity, and (c) overall impressions of realism. The viewing time needed for judgments was also measured. Presentation technique did not affect the viewing time, but it did affect perceptual accuracy and ratings of confidence, clarity, and realism. Animation yielded higher ratings in all categories. It also produced the highest accuracy for depth perception, but the lowest accuracy for size and texture perception. The low accuracy for size and texture was exacerbated by high confidence in the wrong answers. Multiple viewpoints caused neither false confidence nor reduced accuracy in any category, and yielded the highest accuracy in two categories: size and texture. The combined effects on perception and metaperception (knowledge about one's own perceptual processes) suggest that animation creates the best overall impression and is the best for conveying depth information. However, multiple viewpoints are better than animation for conveying size and texture information. These results are discussed with respect to theories of perceptual representation.

Adult↗
PubMedDate not supplied

Categorization in early infancy and the continuity of development.

Arguments and evidence are presented for the conclusion that the young infant's perceptually based categorical representations for natural kinds--animals in this case--are the basis for their mature conceptual counterparts. In addition, it is argued that conceptual development is continuous in nature and without the need for special developmental processes. A consideration of the development of the syllabic, segmental, and featural categories of phonology shows a more complex pattern of change--one marked by both continuities and discontinuities in the representations themselves and the processes that produce them.

Attention↗
PubMed2005

The influence of writing practice on letter recognition in preschool children: a comparison between handwriting and typing.

A large body of data supports the view that movement plays a crucial role in letter representation and suggests that handwriting contributes to the visual recognition of letters. If so, changing the motor conditions while children are learning to write by using a method based on typing instead of handwriting should affect their subsequent letter recognition performances. In order to test this hypothesis, we trained two groups of 38 children (aged 3-5 years) to copy letters of the alphabet either by hand or by typing them. After three weeks of learning, we ran two recognition tests, one week apart, to compare the letter recognition performances of the two groups. The results showed that in the older children, the handwriting training gave rise to a better letter recognition than the typing training.

Child, Preschool↗
PubMed2004

Fluoxetine facilitates use-dependent excitability of human primary motor cortex.

OBJECTIVES: In poststroke patients, fluoxetine, a selective serotonin-reuptake inhibitor, as an adjunct to physical therapy provided a better functional recovery from motor deficits. The aim of this study was to investigate the effect of a single dose of 20 mg fluoxetine on motor learning and associated cortical changes in healthy right-handed subjects in order to get deeper insight into its facilitating influence on human motor cortex. METHODS: Subjects performed a motor task consisting of a simultaneous co-contraction of the abductor pollicis brevis (APB) and the deltoid muscle with and without fluoxetine in a placebo-controlled double-blinded crossover study design. Immediately before and after motor learning motor output maps of the APB muscle were assessed in order to get insight into plastic changes of the muscle representation. RESULTS: We found a significantly improved motor performance under both conditions without having substantial differences between placebo and fluoxetine. After the completion of the motor task there was a medial shift of the APB muscle motor output map. Only after the administration of fluoxetine the sum of MEP amplitudes (SOA) increased and the motor output map enlarged. CONCLUSIONS: These findings provide evidence for a use-dependent facilitating effect of fluoxetine on cortical excitability but not on motor performance. SIGNIFICANCE: Our findings are not in line with previous experiments in poststroke patients. However, long-term treatment with fluoxetine may additionally improve motor function by upregulating serotonergic receptors. Further studies investigating the influence of long-term treatment on cortical excitability and psychophysics may therefore provide deeper insight into a possible therapeutical efficiency of fluoxetine in poststroke patients.

Adult↗