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 1,369 records · Page 76Linked to original sources
PubMed2000

Proactive interference of a sequence of tones in a two-tone pitch comparison task.

Subjects compared pitches of a standard tone and a comparison tone separated by 1,300-3,000 msec and responded according to whether the comparison tone sounded higher or lower in pitch than the standard tone. Three interfering tones at 300-msec intervals were presented before each pair of tones. Their pitch range varied, being either below or above the pitch of the standard tone; in some of the trials, their pitches were identical to the pitch of the standard tone (no interference). The highest error rate in performance was found when the interfering tones and the comparison tone deviated in the same direction in pitch from the standard tone. In turn, their deviations in the opposite directions resulted in the lowest error rate. This effect was not found to be dependent on whether the interfering tones were randomly ordered or monotonically ordered, together with the standard tone, into melodically ascending/descending sequences. An intermediate error rate in performance was found when the interfering tones and the standard tone were identical. The results support earlier hypotheses, presented in the context of retroactive interference, by demonstrating proactive interference of a tone sequence at the level of representations of individual tones.

Acoustic Stimulation↗
PubMed2005

Effects of sensory learning on intracortical synaptic transmission in the barrel cortex of mice.

Pairing tactile stimulation of a row of whiskers with a tail shock results in an expansion of the functional representation of the stimulated whiskers within the primary somatosensory cortex of mice. Using the same paradigm, the present study examined field potentials evoked in ex vivo slices of the barrel cortex. The amplitude of responses, evoked by single and repetitive stimuli in layer IV-layer II/III pathway contained within the barrel column corresponding to the whisker stimulated during training, was unchanged. In contrast, in a transcolumnar pathway from the "trained" barrel to layer II/III of the neighboring, "untrained" column, the amplitude of responses was reduced and responses to trains of stimuli applied at 40 Hz, but not at lower frequencies, depressed faster. These data are suggestive of a selective weakening of excitatory transmission and/or enhancement of inhibitory transmission in transcolumnar pathways, which accompany associative learning-induced cortical plasticity.

Afferent Pathways↗
PubMed2001

Comparison of three databases with a decision tree approach in the medical field of acute appendicitis.

Decision trees have been successfully used for years in many medical decision making applications. Transparent representation of acquired knowledge and fast algorithms made decision trees one of the most often used symbolic machine learning approaches. This paper concentrates on the problem of separating acute appendicitis, which is a special problem of acute abdominal pain from other diseases that cause acute abdominal pain by use of an decision tree approach. Early and accurate diagnosing of acute appendicitis is still a difficult and challenging problem in everyday clinical routine. An important factor in the error rate is poor discrimination between acute appendicitis and other diseases that cause acute abdominal pain. This error rate is still high, despite considerable improvements in history-taking and clinical examination, computer-aided decision-support and special investigation, such as ultrasound. We investigated three different large databases with cases of acute abdominal pain to complete this task as successful as possible. The results show that the size of the database does not necessary directly influence the success of the decision tree built on it. Surprisingly we got the best results from the decision trees built on the smallest and the biggest database, where the database with medium size (relative to the other two) was not so successful. Despite that we were able to produce decision tree classifiers that were capable of producing correct decisions on test data sets with accuracy up to 84%, sensitivity to acute appendicitis up to 90%, and specificity up to 80% on the same test set.

Abdomen, Acute↗
PubMed1985

Analysis of the unique cue in configural discriminations.

Four experiments used an autoshaping procedure with pigeons to investigate the basis of configural discriminations. The elements of both a negative patterning (A+, B+, AB-) and a conditional discrimination (AC+, BD+, AD-, BC-) were paired, in a second-order procedure with two new key lights, X and Y. Responding was then tested to X and Y presented in compound with each other and with A and B. The pattern of responding to compounds containing X and Y was like the pattern of responding to compounds containing their associates, A and B. This suggests that A and B can be replaced by their associates without disrupting responding to their compounds. Because X and Y are physically different from A and B, this in turn suggests that any unique cue controlling responding to their compounds does not depend on the physical presence of the component stimuli. Instead the unique stimulus appears to arise from the joint activation of memory representations.

Animals↗
PubMed1993

[Haptic mental rotation in patients with congenital blindness, acquired blindness and normal vision persons].

Nineteen congenitally blind subjects (Ss), eleven Ss who went blind later in their life, and 42 control Ss with normal vision were tested in a tactile version of a "mental rotation" task. A specially constructed tactile display was used which enabled a rotation of the alphanumeric symbols in multiples of 60 degrees without any form distortion. Ss had to compare two successively presented symbols regarding whether the second was just a rotated version or also a mirror image of the first. Decision latencies and errors were recorded. Neither of dependent variables reveal any differences between the three groups of subjects. Moreover, the well known linear increase of decision latency as a function of angular disparity could also be observed with the tactile stimuli in all three groups. Both results are consistent with the assumption that mental rotation effects are neither linked to the visual modality nor to a manipulation of visual-analog representations. If the task is solved by so called analog processes then these seem to operate with non-visual spatial representations.

Adult↗
PubMed1996

The serial organisation of behaviour by non-human primates; an evaluation of experimental paradigms.

On close examination of research programs which focus, either implicitly or explicitly, on the problem of the organisation of serial order in non-human primates, it is possible to detect some limitations in the paradigms conventionally used. Serial learning studies, which focus on the acquisition of arbitrary lists of unconnected elements point towards a distinction between the representation of ordered series formed by monkeys and pigeons. However, the use of unconnected items prevents an assay of the degree to which primates might be able to impose a structure over the list to be reported. The study of transitive reasoning has been implemented by means of a paradigm where the order of a series is conveyed be presenting a common item in pairs of binary discriminations. Animals tested with this paradigm develop control strategies using more information than that provided by reward contingencies alone. A restriction of this paradigm is that, in its binary form, it does not allow a differentiation between the performance of monkeys and pigeons and even simple models account for a transitive bias in the task. On the basis of these observations it is proposed that novel paradigms which go beyond the binary context, feature multiple connected items, and accord a high degree of spontaneity to the subjects might allow better than traditional ones to uncover qualitative different ways in which different organisms serially organise their behaviour. Some recent research programs based on such rationale, and implemented as search tasks, are then outlined and compared with other approaches to the study of search behaviour. Preliminary results obtained from these studies indicate that the spontaneous serial organisation of multiple connected items might uncover new dimensions of promising comparative relevance.

Animals↗
PubMed1987

The central representation of an aversive event maintains the opioid and nonopioid forms of analgesia.

Exposure to an aversive event, such as shock, can elicit either an opioid or nonopioid analgesia in rats. We suggest that the central representation of an aversive event in working memory activates both forms of analgesia. We formalize this basic hypothesis by coupling it with a current model of animal learning and memory, SOP (Wagner, 1981). SOP is designed to capture the standard operating procedures that govern memory systems. Our application of SOP suggests that manipulations which disrupt the maintenance of information in working memory should alter the magnitude and time course of analgesia. Three experiments are reported that support our proposal. Experiment 1 showed that analgesia decays more rapidly if the representation of the aversive event is displaced from working memory by presenting a postshock distractor. Experiment 2 demonstrated that the postshock distractor alters the magnitude and time course of both the opioid and nonopioid forms of analgesia. Experiment 3 demonstrated that pharmacologically disrupting working memory, by administering a high dose of pentobarbital, prevents mild shock from inducing a strong change in pain reactivity. Implications of the results are discussed.

Animals↗
PubMed2026

mmContext: an open framework for multimodal contrastive learning of omics and text data.

SUMMARY: Multimodal approaches are increasingly leveraged for integrating omics data with textual biological knowledge. Yet there is still no accessible, standardized framework that enables systematic comparison of omics representations with different text encoders within a unified workflow. We present mmContext, a lightweight and extensible multimodal embedding framework built on top of the open-source Sentence Transformers library. The software allows researchers to train or apply models that jointly embed omics and text data using any numeric representation stored in an AnnData.obsm layer and any text encoder available in Hugging Face. mmContext supports integration of diverse biological text sources and provides pipelines for training, evaluation, and data preparation. We train and evaluate models for a RNA-Seq and text integration task, and demonstrate their utility through zero-shot classification of cell types and diseases across four independent datasets. By releasing all models, datasets, and tutorials openly, mmContext enables reproducible and accessible multimodal learning for omics-text integration. AVAILABILITY AND IMPLEMENTATION: Pretrained checkpoints and full source code for our custom MMContextEncoder are available on Hugging Face huggingface.co/jo-mengr. The Python package github.com/mengerj/mmcontext provides the model implementation and training and evaluation scripts for custom training. The releases for the publication can be accessed via zenodo: adata_hf_datasets: doi.org/10.5281/zenodo.19185217 and mmContext: doi.org/10.5281/zenodo.19185493.

Computational Biology↗
PubMed1993

The effect of age on children's learning of problems that require a configural association solution.

Elemental associations permit subjects to solve problems when the significance of the relevant stimulus elements are consistent from trial to trial but do not allow subjects to solve problems that require them to construct and use representations of stimulus conjunctions or configurations to guide their behavior. Recent research with brain-damaged and developing animals has led several theorists to propose that elemental and configural associations depend on different neural systems. Some evidence suggests that changes in children's learning that occur when they are about 4.5 years old may be partially due to developmental differences in access to the elemental and configural association systems. Two experiments are reported that support this hypothesis. Children at least 4.5 years old were able to solve the transverse patterning problem and a conditional discrimination, two problems that require configural association solutions. Younger children did not solve these problems but were able to solve problems constructed from the same stimulus materials that permitted an elemental association solution. These results suggest that children may not gain access to the configural association system until they are about 4.5 years old and support the hypothesis that the configural association system depends on different, later developing neural structures than does the elemental association system.

Association Learning↗
PubMed1998

Speech sound perception and learning: biologic bases.

Historically, auditory research has focused predominantly on how relatively simple acoustic signals are represented in the neuronal responses of the auditory periphery. However, in order to understand the neurophysiology underlying speech perception, the ultimate objective is to discover how speech sounds are represented in the central auditory system and to relate that representation to the perception of speech as a meaningful acoustic signal. This paper reviews three areas pertaining to the central auditory representation of speech: (1) the differences in neural representation of speech sounds at different levels of the auditory system, (2) the relation between the representation of sound in the auditory pathway and the perception/misperception of speech, and (3) the plasticity of speech-sound neural representation and speech perception.

Auditory Pathways↗
PubMed2006

Episodic memory--from brain to mind.

Neuronal mechanisms of episodic memory, the conscious recollection of autobiographical events, are largely unknown because electrophysiological studies in humans are conducted only in exceptional circumstances. Unit recording studies in animals are thus crucial for understanding the neurophysiological substrate that enables people to remember their individual past. Two features of episodic memory--autonoetic consciousness, the self-aware ability to "travel through time", and one-trial learning, the acquisition of information in one occurrence of the event--raise important questions about the validity of animal models and the ability of unit recording studies to capture essential aspects of memory for episodes. We argue that autonoetic experience is a feature of human consciousness rather than an obligatory aspect of memory for episodes, and that episodic memory is reconstructive and thus its key features can be modeled in animal behavioral tasks that do not involve either autonoetic consciousness or one-trial learning. We propose that the most powerful strategy for investigating neurophysiological mechanisms of episodic memory entails recording unit activity in brain areas homologous to those required for episodic memory in humans (e.g., hippocampus and prefrontal cortex) as animals perform tasks with explicitly defined episodic-like aspects. Within this framework, empirical data suggest that the basic structure of episodic memory is a temporally extended representation that distinguishes the beginning from the end of an event. Future research is needed to fully understand how neural encodings of context, sequences of items/events, and goals are integrated within mnemonic representations of autobiographical events.

Action Potentials↗
PubMed1999

Blocking in the spatial domain with arrays of discrete landmarks.

A characteristic feature of associative conditioning is that learning a predictive relationship between two events can block later learning about an added event. It is not yet well established whether blocking occurs in the spatial domain or the circumstances in which it does. We now report, using rats trained to search for hidden food near landmarks in an open field arena, that blocking can occur in spatial learning. The animals noticed the added landmark at the start of the blocking phase and explored it, but either failed to incorporate it into their spatial map or developed a representation in which only some landmarks actually control behavior. Additionally, performance at asymptote was controlled by the shape of the landmark array rather than the individual landmarks comprising it, indicating that blocking in the spatial domain may represent a failure to alter the encoded geometry of a learned array.

Animals↗
PubMed1992

The representation and integration in memory of spatial and nonspatial information.

A series of experiments investigated whether people could integrate nonspatial information about an object with their knowledge of the object's location in space. In Experiments 1 and 3, subjects learned the locations of cities on a fictitious road map; in Experiments 2, 4, and 5, subjects were already familiar with the locations of buildings on a campus. The subjects then learned facts about the cities on the maps or the buildings on the campus. The question of interest was whether or not these nonspatial facts would be integrated in memory with the spatial knowledge. After learning the facts, subjects were given a location-judgment test in which they had to decide whether an object was in one region of the space or another. Knowledge integration was assessed by comparing levels of performance in two conditions: (a) when a city or a building name was primed by a fact about a neighboring city or building, and (b) when a city or a building name was primed by a fact about a distant city or building. Results showed that responses in Condition a were faster or more accurate, or both faster and more accurate, than responses in Condition b. These results indicate that the spatial and nonspatial information were encoded in a common memory representation.

Adult↗
PubMed2004

Asymmetry of generalization decrement in causal learning.

Two experiments required volunteers to learn which of various "planes" caused high levels of pollution. Novel test items were then rated as causes of pollution. Items created by adding novel features were rated at the same level as that of the original training items but items created by removing features received reduced ratings. This asymmetry of generalization decrement was not predicted by a well-known configural model of stimulus representation (Pearce, 1987, 1994) but was predicted by a recently proposed model of stimulus representation, the replaced-elements model (Brandon, Vogel, & Wagner, 2000).

Adult↗
PubMed1997

Effects of orbital frontal and anterior cingulate lesions on object and spatial memory in rhesus monkeys.

Object memory processes, evaluated in rhesus monkeys by delayed nonmatching-to-sample with trial-unique stimuli and object reversal learning, were more severely impaired by orbital frontal than by anterior cingulate lesions. Spatial memory processes, assessed by spatial delayed response and spatial reversal learning, showed a weak trend in the opposite direction, though on these tasks neither lesion produced a serious loss. Comparison of the present results with those of earlier studies on the effects of various limbic system lesions suggests that object memory processes, including object recognition and object-reward association, are served by a circuit consisting mainly of the rhinal cortex, orbitofrontal cortex, and the magnocellular division of the medial dorsal thalamic nucleus. Although both the rhinal and orbitofrontal components of this circuit appear to participate in both functions, evidence from the present and earlier studies suggests that the orbitofrontal component is the more important one for associative memory, i.e. the formation across trials of associations between particular objects or classes of objects and reward, whereas the rhinal component is the more critical one for recognition memory, i.e. the storage and retrieval within trials of the representations of particular objects.

Animals↗
PubMed1996

A nonlinear Hebbian network that learns to detect disparity in random-dot stereograms.

An intrinsic limitation of linear, Hebbian networks is that they are capable of learning only from the linear pairwise correlations within an input stream. To explore what higher forms of structure could be learned with a nonlinear Hebbian network, we constructed a model network containing a simple form of nonlinearity and we applied it to the problem of learning to detect the disparities present in random-dot stereograms. The network consists of three layers, with nonlinear sigmoidal activation functions in the second-layer units. The nonlinearities allow the second layer to transform the pixel-based representation in the input layer into a new representation based on coupled pairs of left-right inputs. The third layer of the network then clusters patterns occurring on the second-layer outputs according to their disparity via a standard competitive learning rule. Analysis of the network dynamics shows that the second-layer units' nonlinearities interact with the Hebbian learning rule to expand the region over which pairs of left-right inputs are stable. The learning rule is neurobiologically inspired and plausible, and the model may shed light on how the nervous system learns to use coincidence detection in general.

Artificial Intelligence↗