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 1,459 records · Page 81Linked to original sources
PubMed2003

Improved mental rotation by near-sighted subjects.

Several studies have shown significant correlations among factors such as sex, age, and hormone levels with performance on mental rotation tasks. To perform spatial rotation also seems to be related to cognitive abilities such as musical skills. The present experiment investigated a possible relationship for enhanced spatial abilities, as observed in near-sighted subjects, with mental rotation performance. 39 near-sighted and 21 normal-sighted subjects were tested on a mental rotation task using two-dimensional representations of three-dimensional objects. Near-sighted subjects displayed fewer errors in possible rotations than normal-sighted subjects. There was no difference in errors between groups in identification of mirror images ("impossible rotations"). Results were interpreted in terms of a relation between enhanced reliance on nonvisual information by near-sighted subjects and improved spatial representation.

Adult↗
PubMed2004

Repetition priming in visual search: episodic retrieval, not feature priming.

Previous research has shown that when the targets of successive visual searches have features in common, response times are shorter. However, the nature of the representation underlying this priming and how priming is affected by the task remain uncertain. In four experiments, subjects searched for an odd-sized target and reported its orientation. The color of the items was irrelevant to the task. When target size was repeated from the previous trial, repetition of target color speeded the response. However, when target size was different from that in the previous trial, repetition of target color slowed responses, rather than speeding them. Our results suggest that these priming phenomena reflect the same automatic mechanism as the priming of pop-out reported by Maljkovic and Nakayama (1994). However, the crossover interaction between repetition of one feature and another rules out Maljkovic and Nakayama's (1994) theory of independent potentiation of distinct feature representations. Instead, we suggest that the priming pattern results from contact with an episodic memory representation of the previous trial.

Association Learning↗
PubMed2003

Redescription disembeds relations: evidence from relational transfer and use in problem solving.

How relational information becomes disembedded from its original context is an important issue for theories of cognition. Two experiments tested the hypothesis that a process called redescription disembeds relations, resulting in abstract and, therefore, more transferable and robust representations. In Experiment 1, participants solved simple problems involving an alternating sequence. Participants who discovered the alternating-sequence relation through redescription transferred the relation to a second type of problem more quickly and used it more consistently than did participants who had been directly instructed on the alternating-sequence strategy. Experiment 2 showed similar effects for participants who discovered the alternating-sequence relation through redescription, as compared with participants who had discovered the relation through information available in the display. The present results converge with previous experimental and correlational evidence that suggests that redescription creates abstract representations of relations.

Adolescent↗
PubMed2001

Priming complex words: evidence for supralexical representation of morphology.

Effects of morphologically related primes were examined in two masked prime experiments. Responses to both free root and derived suffixed word targets were facilitated when primes were derived suffixed words containing the target's root, and this facilitation effect showed a time course similar to that for the facilitation effect of repetition primes (though systematically smaller in magnitude). In a control experiment only the longest prime duration of Experiment 1 was used; responses to derived suffixed word targets were facilitated by both free root primes and derived suffixed word primes sharing the target's root (relative to unrelated and form-related control primes). The free root and derived suffixed word prime conditions did not differ significantly. In Experiment 2, only true derived word primes produced facilitation, whereas morphologically simple primes containing a pseudoroot did not influence performance relative to the unrelated prime condition. We argue that this supports a supralexical account of morphological representation.

Adult↗
PubMed1986

Identification and its vicissitudes as observed in adolescence: object loss and identification.

This paper presents clinical material from the analysis of a late adolescent whose father died when he was just 9 years old. It emphasizes the impact of his father's death upon subsequent identifications. It was through a reactivation of certain critical issues in the transference that the patient was enabled to mourn the loss of his father and to achieve a wider and more mature range of selective identifications. Relevant information is presented to demonstrate how his primordial development influenced his reaction to his father's death. His later identifications were affected by his mother, brother and especially his stepfather. The influence of the death upon the analysis and his selective identifications with the analyst are explored. A brief review of the relevant literature confirms Blum's observation that the splitting of the ego owing to the simultaneous denial and acceptance of the reality of parental loss may be accompanied by a splitting of both parent and self-representations into idealized and denigrated.

Adolescent↗
PubMed2004

Perirhinal and postrhinal contributions to remote memory for context.

The perirhinal (PER) and postrhinal (POR) cortices, two components of the medial temporal lobe memory system, are reciprocally connected with the hippocampus both directly and via the entorhinal cortex. Damage to PER or POR before or shortly after training on a contextual fear conditioning task causes deficits in the subsequent expression of contextual fear, implicating these regions in the acquisition or expression of contextual memory. Here, we examined the contribution of PER and POR to the processing of remotely learned contextual information. Male Long-Evans rats were trained in an unsignaled contextual fear conditioning paradigm. After training, rats received bilateral neurotoxic lesions to PER or POR or sham control surgeries at three different training-to-lesion intervals: 1, 28, or 100 d after training. Two weeks after surgery, lesioned and control rats were returned to the training context to assess contextual fear as measured by freezing. Rats with PER or POR damage froze significantly less in the training context than control rats but were not different from each other. The severity of the deficit did not differ across training-to-lesion intervals for any group. This pattern of deficits differs from that of posttraining hippocampal lesions, for which longer training-to-lesion intervals produce significantly more fear-conditioned contextual freezing than shorter training-to-lesion intervals. In the absence of such a retrograde gradient in the present study, our interpretation is that PER and POR have an ongoing role in the storage or retrieval of representations for context. Alternatively, these regions may be involved in a more extended consolidation process that becomes apparent beyond 100 d after learning.

Animals↗
PubMed2000

Individual differences in spatial learning from computer-simulated environments.

A multivariate study examined relationships between the following factors: paper-and-pencil assessments of verbal and spatial ability, ability to form an accurate spatial representation of a large real-world environment, gender, computer attitudes and experience, proficiency with the navigational interface of a virtual environment (VE), and the ability to acquire and transfer spatial knowledge from a VE. Psychometrically assessed spatial ability and proficiency with the navigational interface were found to make substantial contributions to individual differences in the ability to acquire spatial information from a VE. Gender influenced many VE-related tasks, primarily through its relationship with interface proficiency and spatial ability. Measures of spatial knowledge of a VE maze were highly predictive of subsequent performance in a similar real-world maze, suggesting that VEs can be useful for training people about real-world spaces.

Adult↗
PubMed2004

Effector dependent sequence learning in the serial RT task.

At least five earlier studies could not find effector-dependent learning in the keying version of the serial reaction time (RT) task. Experiment 1 examined whether effector-dependent learning occurs when participants practice the serial RT task with three fingers of one hand for about 1,300 sequence repetitions instead of the more common 50-100 repetitions. The results confirm that, following extended practice, sequence learning produces an effector-dependent component. Specifically, an unpracticed hand executed a practiced sequence slower than a practiced hand. However, Experiment 2 showed that effector-dependent sequence learning develops only when fingers of one hand are used, suggesting that effector-dependent sequence learning involves adjustment to the mechanical interactions between the fingers of one hand. In addition, when sequences had been practiced with one hand, mirror versions of the practiced sequences in both experiments showed moderate transfer. But when practiced with two hands no transfer to a mirrored version of the sequence was observed. This suggests that only practice with one hand produces a representation that facilitates the execution of mirror sequences. Generally, the same results were found in more or less aware participants, congruent with the idea that the effector-dependent representation and the representation allowing transfer to mirror sequences are implicit.

Adult↗
PubMed1996

Redistribution of spatial representation in the hippocampus of aged rats performing a spatial memory task.

Young and old rats performed on a maze according to a forced-choice and then a spatial memory procedure either in the same or a different environment. Aged rats were slower to learn the spatial memory task when tested in the same, but not in a different, room. One interpretation of this pattern of results is that although old rats learn new rules as quickly as young rats, they show less flexibility with old rules and familiar spatial information. Impaired choice accuracy during asymptote performance suggests poor processing of trial-unique information by old rats. Spatial correlates of hippocampal CA1 and hilar cells varied with task demand: CA1 cells of aged rats showed more spatially selective place fields, whereas hilar cells showed more diffuse location coding during spatial memory, and not forced-choice, tests. Such representational reorganization may reflect a compensatory response to age-related neurobiological changes in hippocampus.

Aging↗
PubMed2001

Association of the distinct visual representations of faces and names: a PET activation study.

A PET study of seven normal individuals was carried out to investigate the neural populations involved in the retrieval of the visual representation of a face when presented with an associated name, and conversely. Face-name associations were studied by means of four experimental matching conditions, including the retrieval of previously learned (1) name-name (NN), (2) face-face (FF), (3) name-face (NF), and (4) face-name (FN) associations, as well as a resting scan with eyes closed. Before PET images acquisition, subjects were presented with 24 unknown face-name associations to encode in 12 male/female couples. During PET scanning, their task was to decide whether the presented pair was a previously learned association. The right fusiform gyrus was strongly activated in FF condition as compared to NN and Rest conditions. However, no specific activations were found for NN condition relative to FF condition. A network of three areas distributed in the left hemisphere, both active in (NF-FF) and (FN-NN) comparisons, was interpreted as the locus of the integration of visual faces and names representations. These three regions were localized in the inferior frontal gyrus (BA 45), the medial frontal gyrus (BA 6) and the supramarginal gyrus of the inferior parietal lobe (BA 40). An interactive model accounting for these results, with BA 40 seen as an amodal binding region, is proposed.

Adult↗
PubMed1996

Emergence of simple-cell receptive field properties by learning a sparse code for natural images.

The receptive fields of simple cells in mammalian primary visual cortex can be characterized as being spatially localized, oriented and bandpass (selective to structure at different spatial scales), comparable to the basis functions of wavelet transforms. One approach to understanding such response properties of visual neurons has been to consider their relationship to the statistical structure of natural images in terms of efficient coding. Along these lines, a number of studies have attempted to train unsupervised learning algorithms on natural images in the hope of developing receptive fields with similar properties, but none has succeeded in producing a full set that spans the image space and contains all three of the above properties. Here we investigate the proposal that a coding strategy that maximizes sparseness is sufficient to account for these properties. We show that a learning algorithm that attempts to find sparse linear codes for natural scenes will develop a complete family of localized, oriented, bandpass receptive fields, similar to those found in the primary visual cortex. The resulting sparse image code provides a more efficient representation for later stages of processing because it possesses a higher degree of statistical independence among its outputs.

Algorithms↗
PubMed2004

Cortical activity reductions during repetition priming can result from rapid response learning.

Recent observation of objects speeds up their subsequent identification and classification. This common form of learning, known as repetition priming, can operate in the absence of explicit memory for earlier experiences, and functional neuroimaging has shown that object classification improved in this way is accompanied by 'neural priming' (reduced neural activity) in prefrontal, fusiform and other cortical regions. These observations have led to suggestions that cortical representations of items undergo 'tuning', whereby neurons encoding irrelevant information respond less as a given object is observed repeatedly, thereby facilitating future availability of pertinent object knowledge. Here we provide experimental support for an alternative hypothesis, in which reduced cortical activity occurs because subjects rapidly learn their previous responses. After a primed object classification (such as 'bigger than a shoebox'), cue reversal ('smaller than a shoebox') greatly slowed performance and completely eliminated neural priming in fusiform cortex, which suggests that these cortical item representations were no more available for primed objects than they were for new objects. In contrast, prefrontal cortex activity tracked behavioural priming and predicted the degree to which cue reversal would slow down object classification--highlighting the role of the prefrontal cortex in executive control.

Adolescent↗
PubMed2003

An examination of the value of the generation effect for learning new material.

The generation effect involves an improvement in memory when learners must complete or modify materials. Several researchers have suggested that this effect involves enhanced access to learners' existing memory representations; therefore, the effect should be less effective with meaningless, low meaningful, or unfamiliar material. In the present study, the authors conducted 4 experiments in which legal nonwords were used, and they found no generation effect. In another 2 experiments in which familiar clichés were contrasted with new sentences and with unfamiliar sentences from textbooks, the results showed a greatly reduced generation effect for the new, unfamiliar material. Those findings suggest that memory strategies that depend on the generation effect will have limited effectiveness when they are applied to new or unfamiliar material.

Adolescent↗
PubMedDate not supplied

Influence of the shape of the experimental room on spatial learning in rats.

Rats and other mammals can orient themselves in their habitat and locate various goals using different learning strategies. When a cartographic or local strategy is used, the animals construct and store in their long-term memories a complex representation of the extramaze space. However, exactly what constitutes the behaviorally key components of the extramaze space is still little known. The purpose of this study was to investigate whether the shape of the environment that surrounds the animal is an informative factor when a cartographic strategy is used. Two experiments were performed in which the rats were trained to find food in the goal arm of a four-arm, plus-shaped maze. Experiment 1 shows that lesions to the hippocampus impair the acquisition of this task, which suggests that in fact, a cartographic strategy is necessary to solve it. In Experiment 2, in which only one group of neurologically intact rats was used, we directly manipulated the shape of the space around the maze. The results showed that once the animals have learned the task, the hiding of all the extramaze single landmarks while keeping intact the general shape of the experimental room does not affect performance. Only when the shape of the surrounding space was modified did the performance deteriorate significantly. These results are discussed in relation to the different components of the extramaze environment and the contribution of each of them to the spatial learning.

Animals↗
PubMed2001

Retrospective revaluation in humans: learning or memory?

The phenomenon of retrospective revaluation has posed considerable problems for many associative learning theories as it involves a change in the associative strength of a cue on trials on which that cue is absent. The present series of experiments pursues this idea of changes in associative strength between evoked representations of cues, in an effort to establish, de novo, an excitatory connection between two cues simultaneously activated in memory. Given the finding of Dwyer, Mackintosh, and Boakes (1998) that simultaneous activation of absent cues in the memory of rats resulted in learning comparable to that seen in retrospective revaluation, we expected that if retrospective revaluation was found in humans, then excitatory learning due to simultaneous activation would also be seen. This was not the case. The implications of our results are discussed in terms of Dickinson and Burke's (1996) modified SOP model and a version of McLaren's (1993) APECS network. We conclude that many of the effects attributed to learning in retrospective revaluation studies are better thought of as due to changes in the retrievability of items in memory.

Adult↗
PubMed2004

Changes in brain activation during the acquisition of a new bimanual coodination task.

Motor skill acquisition is associated with the development of automaticity and induces neuroplastic changes in the brain. Using functional magnetic resonance imaging (fMRI), the present study traced learning-related activation changes during the acquisition of a new complex bimanual skill, requiring a difficult spatio-temporal relationship between the limbs, i.e., cyclical flexion-extension movements of both hands with a phase offset of 90 degrees. Subjects were scanned during initial learning and after the coordination pattern was established. Kinematics of the movements were accurately registered and showed that the new skill was acquired well. Learning-related decreases in activation were found in right dorsolateral prefrontal cortex (DLPFC), right premotor, bilateral superior parietal cortex, and left cerebellar lobule VI. Conversely, learning-related increases in activation were observed in bilateral primary motor cortex, bilateral superior temporal gyrus, bilateral cingulate motor cortex (CMC), left premotor cortex, cerebellar dentate nuclei/lobule III/IV/Crus I, putamen/globus pallidus and thalamus. Accordingly, bimanual skill learning was associated with a shift in activation among cortico-subcortical regions, providing further evidence for the existence of differential cortico-subcortical circuits preferentially involved during the early and advanced stages of learning. The observed activation changes account for the transition from highly attention-demanding task performance, involving processing of sensory information and corrective action planning, to automatic performance based on memory representations and forward control.

Adult↗
PubMed1999

Syllable onsets: clusters and adjuncts in acquisition.

The Sonority Sequencing Principle is a presumed universal that governs the permissible sequences of consonants within syllables. In two single-subject experiments, we evaluated this principle as applied to the acquisition of onset clusters and adjuncts by children exhibiting functional phonological delays (age in years;months: 3;2 to 7;8). Experiment 1 tested the hypothesis that children abide by the Sonority Sequencing Principle in development, such that the occurrence and use of marked true clusters implies unmarked clusters, but not vice versa. This claim was validated, in part, by the gradient generalization learning patterns of children who were taught marked clusters. Others who were taught unmarked clusters exhibited limited learning characteristic of within-class generalization, with apparent gaps in sonority sequencing. Experiment 2 examined the role of adjunct sequences /sp, st, sk/, whose markedness status is questionable given their violation of the Sonority Sequencing Principle. Results indicated that children learned adjuncts consistent with patterns of within-class generalization, thereby supporting the view that these sequences are unmarked in structure. Experimental findings are integrated in discussion of the representation of onset clusters and their course of emergence in phonological acquisition relative to the Sonority Sequencing Principle.

Articulation Disorders↗
PubMed1997

Case-based learning of plans and goal states in medical diagnosis.

We introduce a case-based system, BOLERO, that learns both plans and goal states. The major aim is that of improving the performance of a rule-based diagnosis system by adapting its behavior using the most recent information available about a patient. On the one hand BOLERO gets knowledge from cases in the form of diagnostic plans that are represented as sequences of decision steps. The advantages of this representation include: (1) retrieval and adaptation of parts of plans (steps) appropriate to the current problem state; (2) generation of new plans not previously available in memory; and (3) learning from experience, both from successful or failed plans. On the other hand, since goal states are sets of final diagnosis likelihoods they are not known beforehand, i.e. goal states are not defined and the system has to learn to recognize them. For this reason BOLERO has a case-based method that uses solutions of past cases to recognize a diagnostic state as a goal state of a new planning problem. BOLERO and a rule-based system are integrated into a meta-level architecture in which we emphasize the collaboration of both systems in solving problems. The rule-based system executes the plans generated by BOLERO. As a consequence of the execution of plans, the rule-based system furnishes BOLERO with new information with which BOLERO can generate a new plan to adapt the reasoning process of the rule-based system into correspondence with the recent available data. All the methods have been designed to be useful for medical diagnosis and have been tested in the domain of diagnosing pneumonia.

Artificial Intelligence↗