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 271 records · Page 15Linked to original sources
PubMed2005

Organizing learning processes on risks by using the bow-tie representation.

The Aramis method proposes a complete and efficient way to manage risk analysis by using the bow-tie representation. This paper shows how the bow-tie representation can also be appropriate for experience learning. It describes how a pharmaceutical production plant uses bow-ties for incident and accident analysis. Two levels of bow-ties are constructed: standard bow-ties concern generic risks of the plant whereas local bow-ties represent accident scenarios specific to each workplace. When incidents or accidents are analyzed, knowledge that is gained is added to existing local bow-ties. Regularly, local bow-ties that have been updated are compared to standard bow-ties in order to revise them. Knowledge on safety at the global and at local levels is hence as accurate as possible and memorized in a real time framework. As it relies on the communication between safety experts and local operators, this use of the bow-ties contributes therefore to organizational learning for safety.

Accidents, Occupational↗
PubMed2006

The structure and function of explanations.

Generating and evaluating explanations is spontaneous, ubiquitous and fundamental to our sense of understanding. Recent evidence suggests that in the course of an individual's reasoning, engaging in explanation can have profound effects on the probability assigned to causal claims, on how properties are generalized and on learning. These effects follow from two properties of the structure of explanations: explanations accommodate novel information in the context of prior beliefs, and do so in a way that fosters generalization. The study of explanation thus promises to shed light on core cognitive issues, such as learning, induction and conceptual representation. Moreover, the influence of explanation on learning and inference presents a challenge to theories that neglect the roles of prior knowledge and explanation-based reasoning.

Adult↗
PubMed2005

Unconscious symmetrical inferences: A role of consciousness in event integration.

Explicit and implicit learning have been attributed to different learning processes that create different types of knowledge structures. Consistent with that claim, our study provides evidence that people integrate stimulus events differently when consciously aware versus unaware of the relationship between the events. In a first, acquisition phase participants sorted words into two categories (A and B), which were fully predicted by task-irrelevant primes-the labels of two other, semantically unrelated categories (C and D). In a second, test phase participants performed a lexical decision task, in which all word stimuli stemmed from the previous prime categories (C and D) and the (now nonpredictive) primes were the labels of the previous target categories (A and B). Reliable priming effects in the second phase demonstrated that bidirectional associations between the respective categories had been formed in the acquisition phase (A<-->C and B<-->D), but these effects were found only in participants that were unaware of the relationship between the categories! We suggest that unconscious, implicit learning of event relationships results in the rather unsophisticated integration (i.e., bidirectional association) of the underlying event representations, whereas explicit learning takes the meaning of the order of the events into account, and thus creates unidirectional associations.

Adult↗
PubMed1992

A note on schema and exemplar approaches to motor skill representation in memory.

Given the need for a memory representation of well-learned motor skills, a common assumption in motor behavior is that this knowledge is stored in a central, abstracted form. Active production of motor skills has not been used in experimental designs that have provided empirical support for this view of representation, however. Much of the faith in centralized, abstracted forms of memory representation for motor skills is due to the popularity of Schmidt's schema theory, which has adapted the prototype abstraction model from category learning research to the representation of motor skills. Since schema theory was proposed, however, an alternative view that seriously questions the preeminence of the prototype abstraction model for the central representation of knowledge has arisen in the category learning literature. This particular view, termed the specific exemplar model, has led a number of researchers in cognition to develop mixed models that involve both prototypic abstraction and specific exemplar elements. This note, then, identifies what can be perceived as a gap in the empirical knowledge base in motor behavior and discusses the possibility of using the debate about representation for category learning as a stimulus for initiating a similar investigation into the representation of motor skills. A hypothetical specific exemplar model for the memory representation of motor skills is outlined, and possible empirical comparisons between this model and the schema abstraction model are suggested.

Journal Article↗
PubMed2006

Haven't we met before? The effect of facial familiarity on repetition priming.

Within the word recognition literature, word-frequency and hence familiarity has been shown to affect the degree of repetition priming. The current paper reports two experiments which examine whether familiarity also affects the degree of repetition priming for faces. The results of Experiment 1 confirmed that familiarity did moderate the degree of priming in a face recognition task. Low familiarity faces were primed to a significantly greater degree than high familiarity faces in terms of accuracy, speed, and efficiency of processing. Experiment 2 replicated these results but additionally, demonstrated that familiarity moderates priming for name recognition as well as face recognition. These results can be accommodated within both a structural account of repetition priming (Burton, Bruce & Johnston, 1990) and an Episodic Memory account of repetition priming (see Roediger, 1990), and are discussed in terms of a common mechanism for priming, learning and the representation of familiarity.

Adolescent↗
PubMed2004

Impaired learning-dependent cortical plasticity in Huntington's disease transgenic mice.

Huntington's disease (HD) is a genetically transmitted neurodegenerative disorder. The neuropathology in HD is a selective neuronal cell death in several brain regions including cortex. Although changes in synaptic plasticity were shown within the hippocampus and striatum of HD transgenic mice, there are no studies considering neocortical synaptic plasticity abnormalities in HD. We examined the impact of the HD transgene upon learning-dependent plasticity of cortical representational maps. The effect of associative learning, in which stimulation of a row of vibrissae was paired with appetitive stimulus, upon functional representations of vibrissae in the barrel cortex, was investigated with 2-deoxyglucose brain mapping in presymptomatic R6/1 HD mice. In wild-type mice, cortical representation of the row of vibrissae involved in the training was expanded, while in HD mice the representation of this row was not expanded. The results suggest that presymptomatic R6/1 HD transgenic mice show deficits in plasticity of primary somatosensory cortex.

Animals↗
PubMed2005

Learning and transfer of bimanual multifrequency patterns: effector-independent and effector-specific levels of movement representation.

Current behavioural theories consider that during motor learning, an effector-independent memory representation of the acquired skill is built up. Using a transfer paradigm, we addressed the nature of the memory representation for a 2:1 multifrequency co-ordination task, requiring, for example, the left arm to cycle twice as fast as the right. After learning this 2:1 pattern, transfer to its converse pattern (i.e., the right arm cycles twice as fast as the left) revealed powerful evidence for negative transfer. The converse task arrangement revealed similar effects. These observations suggest a reconsideration of current viewpoints on movement representations, which emphasize effector independence. Based on the present findings, we propose a new model of motor memory, consisting of an abstract, effector-independent and an effector-specific layer. The abstract code is hypothesized to represent general spatiotemporal movement features, whereas the specific representation refers to effector-related movement commands. This concept is consistent with recent neuroscientific evidence in animal and human species, and invites a reconsideration of current behavioural theories of motor learning and memory.

Adult↗
PubMed1992

ALCOVE: an exemplar-based connectionist model of category learning.

ALCOVE (attention learning covering map) is a connectionist model of category learning that incorporates an exemplar-based representation (Medin & Schaffer, 1978; Nosofsky, 1986) with error-driven learning (Gluck & Bower, 1988; Rumelhart, Hinton, & Williams, 1986). Alcove selectively attends to relevant stimulus dimensions, is sensitive to correlated dimensions, can account for a form of base-rate neglect, does not suffer catastrophic forgetting, and can exhibit 3-stage (U-shaped) learning of high-frequency exceptions to rules, whereas such effects are not easily accounted for by models using other combinations of representation and learning method.

Attention↗
PubMed2002

Representation of time in time-place learning.

Ordinal, interval, and circadian mechanisms of solving a time-place task were tested. Rats searched for food twice in the morning and once in the afternoon (Group AB-C, n = 5) or once in the morning and twice in the afternoon (Group A-BC, n = 5) in a box with four food troughs. The location of the food depended on the time of day in a 12:12-h light:dark cycle. Acquisition was documented by food-site inspections at the correct locations prior to food availability. On nonrewarded probes, the time of the middle search (B) was shifted late (for Group AB-C) or early (for Group A-BC). The rats visited Location B at chance, contrary to an ordinal mechanism. When the posttesting meal and light-dark transitions were omitted, the rats visited correct locations with impaired performance but at above-chance levels on nonrewarded probes. The results are consistent with interval and circadian representations of time.

Animals↗
PubMed2004

Neural systems for individual song recognition in adult birds.

The songbird auditory system is an excellent model for neuroethological studies of the mechanisms that govern the perception and cognition of natural stimuli (i.e., song), and the translation of corresponding representations into natural behaviors. One common songbird behavior is the learned recognition of individual conspecific songs. This chapter summarizes the research effort to identify the brain regions and mechanisms mediating individual song recognition in European starlings, a species of songbird. The results of laboratory behavioral studies are reviewed, which show that when adult starlings learn to recognize other individual's songs, they do so by memorizing large sets of song elements, called motifs. Recent data from single neurons in the caudal medial portion of the mesopallium are then reviewed, showing that song recognition learning leads to explicit representation of acoustic features that correspond closely to specific motifs, but only to motifs in the songs that birds have learned to recognize. This suggests that the strength and tuning of high-level auditory object representations, of the sort that presumably underlie many forms of vocal communication, are shaped by each animal's unique experience.

Animal Communication↗
PubMed2000

Spatial cognition and neuro-mimetic navigation: a model of hippocampal place cell activity.

A computational model of hippocampal activity during spatial cognition and navigation tasks is presented. The spatial representation in our model of the rat hippocampus is built on-line during exploration via two processing streams. An allothetic vision-based representation is built by unsupervised Hebbian learning extracting spatio-temporal properties of the environment from visual input. An idiothetic representation is learned based on internal movement-related information provided by path integration. On the level of the hippocampus, allothetic and idiothetic representations are integrated to yield a stable representation of the environment by a population of localized overlapping CA3-CA1 place fields. The hippocampal spatial representation is used as a basis for goal-oriented spatial behavior. We focus on the neural pathway connecting the hippocampus to the nucleus accumbens. Place cells drive a population of locomotor action neurons in the nucleus accumbens. Reward-based learning is applied to map place cell activity into action cell activity. The ensemble action cell activity provides navigational maps to support spatial behavior. We present experimental results obtained with a mobile Khepera robot.

Animals↗
PubMed2006

Making working memory work: a computational model of learning in the prefrontal cortex and basal ganglia.

The prefrontal cortex has long been thought to subserve both working memory (the holding of information online for processing) and executive functions (deciding how to manipulate working memory and perform processing). Although many computational models of working memory have been developed, the mechanistic basis of executive function remains elusive, often amounting to a homunculus. This article presents an attempt to deconstruct this homunculus through powerful learning mechanisms that allow a computational model of the prefrontal cortex to control both itself and other brain areas in a strategic, task-appropriate manner. These learning mechanisms are based on subcortical structures in the midbrain, basal ganglia, and amygdala, which together form an actor-critic architecture. The critic system learns which prefrontal representations are task relevant and trains the actor, which in turn provides a dynamic gating mechanism for controlling working memory updating. Computationally, the learning mechanism is designed to simultaneously solve the temporal and structural credit assignment problems. The model's performance compares favorably with standard backpropagation-based temporal learning mechanisms on the challenging 1-2-AX working memory task and other benchmark working memory tasks.

Algorithms↗
PubMed2006

Associative activation of stimulus representations restores lost salience: implications for perceptual learning.

In 3 experiments, rats received preexposure to presentations of a compound flavor BX. The effective salience of B was then tested by assessing its ability to interfere with the aversion controlled by another flavor or the tendency to drink a saline solution after the induction of a salt need. It was found that the effective salience of B was maintained when during preexposure, presentations of BX alternated with presentations of X alone. This was true both when BX was presented as a simultaneous compound (Experiment 1) and as a serial compound (X-->B; Experiments 2 and 3); salience was not maintained when the serial compound took the form B-->X (Experiments 2 and 3a). It was argued that the salience of B declines during preexposure but is restored when presentations of X are able to activate the representation of B by way of the associative X-B link.

Animals↗
PubMed2003

Development of object concepts in infancy: Evidence for early learning in an eye-tracking paradigm.

Concepts of objects as enduring and complete across space and time have been documented in infants within several months after birth, but little is known about how such concepts arise during development. Current theories that stress innate knowledge may neglect the potential contributions of experience to guide acquisition of object concepts. To examine whether learning plays an important role in early development of object representations, we used an eye-tracking paradigm with 4- and 6-month-old infants who were provided with an initial period of experience viewing an unoccluded trajectory, or no experience with this particular stimulus. After exposure to the unoccluded trajectory for only 2 min, there was a reliable increase in 4-month-old infants' anticipatory eye movement when the infants subsequently viewed occluded-trajectory displays, relative to 4-month-old infants who did not receive this experience. This effect of training in 4-month-old infants was found to generalize to another category of trajectory orientation. Older infants received no additional benefit from training, most likely because they enter the task capable of forming robust object representations under these conditions. This finding provides compelling evidence that very brief training facilitated formation of object representations, and suggests more generally that infants learn such representations from real-world experience viewing objects undergoing occlusion and disocclusion.

Age Factors↗
PubMed2002

Adults' representations of division facts: a consequence of learning history?

There has been a recent increase in the study of adults' performance on simple division problems. Researchers up to now have focused on the relationship between multiplication and division and have found that multiplication often has a mediating role in the solution of division problems (Campbell, 1997, 1999; LeFevre & Morris, 1999; Mauro, LeFevre, & Morris, 2002). In this study, division was exclusively examined to determine the strategies that are used to solve these problems and to identify factors relating to particular strategy use. Thirty-two participants were asked to solve two sets of 64 simple division problems (from 4 divided by 2 to 81 divided by 9) and error, latency, and strategy report data were collected. Fewer errors were made on easy problems, which were also solved more quickly than difficult problems. Participants used retrieval, multiplication, and other strategies to solve the problems and tended to use retrieval more on easy than difficult problems and used multiplication more on difficult problems than easy problems. Unexpected age differences in strategy use were also found. Older participants tended to rely more heavily on retrieval than younger participants. These results suggest that older participants may have stronger representations for simple division problems than younger participants.

Adolescent↗
PubMed1995

Internal representations of the motor apparatus: implications from generalization in visuomotor learning.

Recent computational studies have proposed that the motor system acquires internal models of kinematic transformations, dynamic transformations, or both by learning. Computationally, internal models can be characterized by 2 extreme representations: structured and tabular (C. G. Atkeson, 1989). Tabular models do not need prior knowledge about the structure of the motor apparatus, but they lack the capability to generalize learned movements. Structured models, on the other hand, can generalize learned movements, but they require an analytical description of the motor apparatus. In investigating humans' capacity to generalize kinematic transformations, we examined which type of representation humans' motor system might use. Results suggest that internal representations are nonstructured and nontabular. Findings may be due to a neural network model with a medium number of neurons and synapses.

Adult↗
PubMed2002

Multiple parallel memory systems in the brain of the rat.

A theory of multiple parallel memory systems in the brain of the rat is described. Each system consists of a series of interconnected neural structures. The "central structures" of the three systems described are the hippocampus, the matrix compartment of the dorsal striatum (caudate-putamen), and the amygdala. Information, coded as neural signals, flows independently through each system. All systems have access to the same information from situations in which learning occurs, but each system is specialized to represent a different kind of relationship among the elements (stimulus events, responses, reinforcers) of the information that flows through it. The speed and accuracy with which a system forms a coherent representation of a learning situation depend on the correspondence between the specialization of the system and the relationship among the elements of the situation. The coherence of these stored representations determines the degree of control exerted by each system on behavior in the situation. Although they process information independently the systems interact in at least two ways: by simultaneous parallel influence on behavioral output and by directly influencing each other. These interactions can be cooperative (leading to similar behaviors) or competitive (leading to different behaviors). Experimental findings consistent with these ideas, mostly from experiments with rats, are reviewed.

Amygdala↗
PubMed2003

Slow feature analysis: a theoretical analysis of optimal free responses.

Temporal slowness is a learning principle that allows learning of invariant representations by extracting slowly varying features from quickly varying input signals. Slow feature analysis (SFA) is an efficient algorithm based on this principle and has been applied to the learning of translation, scale, and other invariances in a simple model of the visual system. Here, a theoretical analysis of the optimization problem solved by SFA is presented, which provides a deeper understanding of the simulation results obtained in previous studies.

Algorithms↗