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 541 records · Page 30Linked to original sources
PubMed2006

Infants form associations between memory representations of stimuli that are absent.

Traditional models of learning assume that an association can be formed only between cues that are physically present. Here, we report that when two objects that had never appeared together were simultaneously activated in memory, young human infants associated the representations of those objects. Neither object was physically present at the time the association was formed. The association remained latent for up to 2 weeks, when the infants used it to perform a deferred imitation task. These findings reveal that what infants merely see "brings to mind" what they saw before and combines it in new ways. In addition to challenging a fundamental tenet of classic learning models, these findings have major theoretical and practical implications for early cognitive development. Every day, in the same manner, young infants probably form numerous associations between activated memories of objects that are physically absent, creating a potential knowledge base of untold dimensions.

Association↗
PubMed2025

Refining sequence-to-activity models by increasing model resolution.

Decoding the cis-regulatory syntax that controls gene expression is essential for improving our understanding of cell differentiation and disease. To identify regulatory motifs and their regulatory syntax, deep learning based sequence-to-activity (S2A) models learn transcription factor binding motifs and their combinations from DNA sequence by modeling measured chromatin accessibility. Previously, we developed AI-TAC, a S2A model that predicts chromatin accessibility across various immune cell types in multi-task fashion, effectively decoding the regulatory syntax underlying immune cell differentiation. While ATAC-seq is commonly used to measure regional accessibility, it also provides high-resolution profiles, the distribution of Tn5 insertion sites, that offer additional insights into the precise location and strength of TF binding sites. Here we demonstrate that modeling ATAC-seq profiles alongside accessibility consistently improves predictions of differential chromatin accessibility across cell types. Moreover, we also find that multi-task learning across related immune cell types consistently outperforms single-task models. To understand what additional information bpAITAC learns from ATAC-seq profiles, we systematically compare sequence attributions from models trained with and without ATAC-seq profiles. We identify novel motifs with strong effect sizes that emerge only when profile data is included. Our findings suggest that modeling ATAC-seq at base-pair resolution enables the model to learn a more nuanced and sensitive representation of the cis-regulatory syntax driving immune cell-specific chromatin landscapes.

ATAC-seq↗
PubMed2000

Animat navigation using a cognitive graph.

This article describes a computational model of the hippocampus that makes it possible for a simulated rat to navigate in a continuous environment containing obstacles. This model views the hippocampus as a "cognitive graph", that is, a hetero-associative network that learns temporal sequences of visited places and stores a topological representation of the environment. Calling upon place cells, head direction cells, and "goal cells", it suggests a biologically plausible way of exploiting such a spatial representation for navigation that does not require complicated graph-search algorithms. Moreover, it permits "latent learning" during exploration, that is, the building of a spatial representation without the need of any reinforcement. When the rat occasionally discovers some rewarding place it may wish to rejoin subsequently, it simply records within its cognitive graph, through a series of goal and sub-goal cells, the direction in which to move from any given start place. Accordingly, the model implements a simple "place-recognition-triggered response" navigation strategy. Two implementations of place cell management are studied in parallel. The first one associates place cells with place fields that are given a priori and that are uniformly distributed in the environment. The second one dynamically recruits place cells as exploration proceeds and adjusts the density of such cells to the local complexity of the environment. Both implementations lead to identical results. The article ends with a few predictions about results to be expected in experiments involving simultaneous recordings of multiple cells in the rat hippocampus.

Animals↗
PubMed1997

Connectionism, phonology, reading, and regularity in developmental dyslexia.

Tests of the "phonological deficit" account of developmental dyslexia have produced apparently inconsistent results. We show how a connectionist approach to dyslexic reading development can resolve the paradox. A "dyslexic" model of reading was created by reducing the quality of the phonological representations available to the model during learning. The model behaved similarly to dyslexic children in that it had a selectively reduced ability to process nonwords, but showed normal effects of words' spelling-to-sound regularity. An experimental test of the model's predictions confirmed that dyslexic children perform similarly, in that they are impaired on irregular words to the same extent as nondyslexic children. It is concluded that developmentally dyslexic reading can indeed be understood in terms of impaired phonological representations and that the adoption of a modeling approach resolves an apparent paradox in the experimental literature.

Child↗
PubMed2026

CAKR: commutative algebra k-mer representations for genomics.

Despite the availability of various sequence analysis models, comparative genomic analysis remains a challenge in genomics, genetics, and phylogenetics. Commutative algebra, a fundamental tool in algebraic geometry and number theory, has rarely been used in data and biological sciences. In this study, we introduce commutative algebra k-mer representations as a nonlinear algebraic framework for analyzing genomic sequences. This representation bridges commutative algebra, algebraic topology, combinatorics, and machine learning to establish a mathematical framework for comparative genomic analysis. We evaluate its effectiveness on three tasks including genetic variant classification, phylogenetic tree reconstruction, and viral classification, typically requiring alignment-based, alignment-free, and machine-learning approaches, respectively. In this work, we show that commutative algebra k-mer representations outperform five state-of-the-art sequence analysis methods across twelve primary datasets, with two additional supplementary fragment-placement benchmarks, especially in viral classification, and maintain relatively stable predictive accuracy as dataset size increases, underscoring scalability and robustness.

Genomics↗
PubMed2004

Increase in reaction time for solving problems during learning-set formation.

Six rhesus monkeys were tested for a change in reaction time for problem-solving during a learning-set task, in which they showed progressive improvement in the rate of learning successive problems of visual discrimination. To evaluate the processing time for cognitive processes in problem-solving, the differences in release latency and movement time between the visual discrimination task and the visuomotor control task were defined. In their first experience, the monkeys required several hundreds of trials for solving the problem, and the Deltarelease latency was constant throughout the learning. With increasing experience, they solved problems within fewer trials than with the first problem. At this stage, the Deltarelease latency was high at the beginning and then decreased. The rise in the Deltarelease latency within the learning acquisition period increased depending on the amount of experience with problems they had solved, whereas the Deltamovement time within that period was not significantly affected by the experience with problems. The present findings suggest that the number of problem-solving experiences could promote profound cognitive processing, which may be related to a conceptual representation that actualizes the flexibility of learning, namely, the learning set.

Animals↗
PubMed1982

Synthesis of nonlinear control surfaces by a layered associative search network.

An approach to solving nonlinear control problems is illustrated by means of a layered associative network composed of adaptive elements capable of reinforcement learning. The first layer adaptively develops a representation in terms of which the second layer can solve the problem linearly. The adaptive elements comprising the network employ a novel type of learning rule whose properties, we argue, are essential to the adaptive behavior of the layered network. The behavior of the network is illustrated by means of a spatial learning problem that requires the formation of nonlinear associations. We argue that this approach to nonlinearity can be extended to a large class of nonlinear control problems.

Animals↗
PubMed1999

Knowledge extraction: a comparison between symbolic and connectionist methods.

The use of a linguistic representation for expressing knowledge acquired by learning systems is an important issue as regards to user understanding. Under this assumption, and to make sure that these systems will be welcome and used, several techniques have been developed by the artificial intelligence community, under both the symbolic and the connectionist approaches. This work discusses and investigates three knowledge extraction techniques based on these approaches. The first two techniques, the C4.5 and CN2 symbolic learning algorithms, extract knowledge directly from the data set. The last technique, the TREPAN algorithm extracts knowledge from a previously trained neural network. The CN2 algorithm induces if...then rules from a given data set. The C4.5 algorithm extracts decision trees, although it can also extract ordered rules, from the data set. Decision trees are also the knowledge representation used by the TREPAN algorithm.

Algorithms↗
PubMed2001

The rules of formation of the olfactory representations found in the orbitofrontal cortex olfactory areas in primates.

Approximately 35% of neurons in the orbitofrontal cortex taste and olfactory areas with olfactory responses provide a representation of odour that depends on the taste with which the odour has been associated previously. This representation is produced by a slowly acting learning mechanism that learns associations between odour and taste. Other neurons in the orbitofrontal cortex respond to both the odour and to the mouth feel of fat. The representation of odour thus moves for at least some neurons in the orbitofrontal cortex beyond the domain of physico-chemical properties of the odours to a domain where the ingestion-related significance of the odour determines the representation provided. Olfactory neurons in the primate orbitofrontal cortex decrease their responses to a food eaten to satiety, but remain responsive to other foods, thus contributing to a mechanism for olfactory sensory-specific satiety. It has been shown in neuroimaging studies that the human orbitofrontal cortex provides a representation of the pleasantness of odour, in that the activation produced by the odour of a food eaten to satiety decreases relative to another food-related odour not eaten in the meal. In the same general area there is a representation of the pleasantness of the smell, taste and texture of a whole food, in that activation in this area decreases to a food eaten to satiety, but not to a food that has not been eaten in the meal.

Animals↗
PubMed1999

Learning the parts of objects by non-negative matrix factorization.

Is perception of the whole based on perception of its parts? There is psychological and physiological evidence for parts-based representations in the brain, and certain computational theories of object recognition rely on such representations. But little is known about how brains or computers might learn the parts of objects. Here we demonstrate an algorithm for non-negative matrix factorization that is able to learn parts of faces and semantic features of text. This is in contrast to other methods, such as principal components analysis and vector quantization, that learn holistic, not parts-based, representations. Non-negative matrix factorization is distinguished from the other methods by its use of non-negativity constraints. These constraints lead to a parts-based representation because they allow only additive, not subtractive, combinations. When non-negative matrix factorization is implemented as a neural network, parts-based representations emerge by virtue of two properties: the firing rates of neurons are never negative and synaptic strengths do not change sign.

Algorithms↗
PubMed1997

Hippocampal synaptic plasticity: role in spatial learning or the automatic recording of attended experience?

Allocentric spatial learning can sometimes occur in one trial. The incorporation of information into a spatial representation may, therefore, obey a one-trial correlational learning rule rather than a multi-trial error-correcting rule. It has been suggested that physiological implementation of such a rule could be mediated by N-methyl-D-aspartate (NMDA) receptor-dependent long-term potentiation (LTP) in the hippocampus, as its induction obeys a correlational type of synaptic learning rule. Support for this idea came originally from the finding that intracerebral infusion of the NMDA antagonist AP5 impairs spatial learning, but studies summarized in the first part of this paper have called it into question. First, rats previously given experience of spatial learning in a watermaze can learn a new spatial reference memory task at a normal rate despite an appreciable NMDA receptor blockade. Second, the classical phenomenon of 'blocking' occurs in spatial learning. The latter finding implies that spatial learning can also be sensitive to an animal's expectations about reward and so depend on more than the detection of simple spatial correlations. In this paper a new hypothesis is proposed about the function of hippocampal LTP. This hypothesis retains the idea that LTP subserves rapid one-trial memory, but abandons the notion that it serves any specific role in the geometric aspects of spatial learning. It is suggested that LTP participates in the automatic recording of attended experience': a subsystem of episodic memory in which events are temporarily remembered in association with the contexts in which they occur. An automatic correlational form of synaptic plasticity is ideally suited to the online registration of context event associations. In support, it is reported that the ability of rats to remember the most recent place they have visited in a familiar environment is exquisitely sensitive to AP5 in a delay-dependent manner. Moreover, new studies of the lasting persistence of NMDA-dependent LTP, known to require protein synthesis, point to intracellular mechanisms that enable transient synaptic changes to be stabilized if they occur in close temporal proximity to important events. This new property of hippocampal LTP is a desirable characteristic of an event memory system.

2-Amino-5-phosphonovalerate↗
PubMed1998

Distance learning, problem based learning and dynamic knowledge networks.

This paper is an attempt to develop a distance learning model grounded upon a strict integration of problem based learning (PBL), dynamic knowledge networks (DKN) and web tools, such as hypermedia documents, synchronous and asynchronous communication facilities, etc. The main objective is to develop a theory of distance learning based upon the idea that learning is a highly dynamic cognitive process aimed at connecting different concepts in a network of mutually supporting concepts. Moreover, this process is supposed to be the result of a social interaction that has to be facilitated by the web. The model was tested by creating a virtual classroom of medical and nursing students and activating a learning session on the concept of knowledge representation in health sciences.

Computer Communication Networks↗
PubMed2003

Watch how to do it! New advances in learning by observation.

Recent data demonstrate that the cerebellum contributes to the internal representation of action. This representation is used not only to generate motor actions, but also to understand and learn the actions and skills of others by imitation. The cerebellar networks appear to be indispensable for acquiring complex behaviors and procedures. The cerebellar role in the acquisition of procedural competencies is particularly evident in spatial information processing. The cerebellum allows acquiring by observation competencies in exploration behaviors as efficient as the competencies acquired by actually performing the same task. The specificity of the cerebellar role in the acquisition phases of learning by observation is demonstrated by the complete absence of spatial learning when the observational training is performed in presence of a cerebellar lesion. This datum is further corroborated by the evidence that, once acquired, spatial procedures can be efficiently performed even in the presence of cerebellar damage, in agreement with the neuroimaging findings of low cerebellar activation after prolonged practice. The finding that the cerebellum is involved in procedural acquisition and in observational learning allowed us to dissect a complex behavior into single behavioral units forming a complete procedural sequence, demonstrating that such behavioral units do exist and can be independently acquired.

Animals↗
PubMed1998

Is the hippocampus a Kalman filter?

Based on a large body of neurophysiological, neuroanatomical, and behavioral data, it has been suggested that the hippocampal formation serves as a spatial learning and localization system. This spatial representation is metric in nature and arises as a result of associations between sensory inputs and dead-reckoning information generated by the animal. However, despite the fact that these two information streams provide uncertain information (e.g., recognition errors, dead-reckoning drifts, etc.), the hippocampal computational models suggested to date have not explicitly addressed information fusion from erroneous sources. In this paper we develop a computational model of hippocampal spatial learning and relate its functioning to a probabilistic tool used for uncertain sensory fusion in robots: the Kalman filter. This parallel allows us to derive statistically optimal update expressions for the localization performed by our computational model.

Animals↗
PubMed2006

ESPERR: learning strong and weak signals in genomic sequence alignments to identify functional elements.

Genomic sequence signals - such as base composition, presence of particular motifs, or evolutionary constraint - have been used effectively to identify functional elements. However, approaches based only on specific signals known to correlate with function can be quite limiting. When training data are available, application of computational learning algorithms to multispecies alignments has the potential to capture broader and more informative sequence and evolutionary patterns that better characterize a class of elements. However, effective exploitation of patterns in multispecies alignments is impeded by the vast number of possible alignment columns and by a limited understanding of which particular strings of columns may characterize a given class. We have developed a computational method, called ESPERR (evolutionary and sequence pattern extraction through reduced representations), which uses training examples to learn encodings of multispecies alignments into reduced forms tailored for the prediction of chosen classes of functional elements. ESPERR produces a greatly improved Regulatory Potential score, which can discriminate regulatory regions from neutral sites with excellent accuracy ( approximately 94%). This score captures strong signals (GC content and conservation), as well as subtler signals (with small contributions from many different alignment patterns) that characterize the regulatory elements in our training set. ESPERR is also effective for predicting other classes of functional elements, as we show for DNaseI hypersensitive sites and highly conserved regions with developmental enhancer activity. Our software, training data, and genome-wide predictions are available from our Web site (http://www.bx.psu.edu/projects/esperr).

Algorithms↗
PubMed2005

Neuromuscular control of the point to point and oscillatory movements of a sagittal arm with the actor-critic reinforcement learning method.

In this study, we have used a single link system with a pair of muscles that are excited with alpha and gamma signals to achieve both point to point and oscillatory movements with variable amplitude and frequency.The system is highly nonlinear in all its physical and physiological attributes. The major physiological characteristics of this system are simultaneous activation of a pair of nonlinear muscle-like-actuators for control purposes, existence of nonlinear spindle-like sensors and Golgi tendon organ-like sensor, actions of gravity and external loading. Transmission delays are included in the afferent and efferent neural paths to account for a more accurate representation of the reflex loops.A reinforcement learning method with an actor-critic (AC) architecture instead of middle and low level of central nervous system (CNS), is used to track a desired trajectory. The actor in this structure is a two layer feedforward neural network and the critic is a model of the cerebellum. The critic is trained by state-action-reward-state-action (SARSA) method. The critic will train the actor by supervisory learning based on the prior experiences. Simulation studies of oscillatory movements based on the proposed algorithm demonstrate excellent tracking capability and after 280 epochs the RMS error for position and velocity profiles were 0.02, 0.04 rad and rad/s, respectively.

Arm↗
PubMed2004

Atypical brainstem representation of onset and formant structure of speech sounds in children with language-based learning problems.

This study investigated how the human auditory brainstem represents constituent elements of speech sounds differently in children with language-based learning problems (LP, n = 9) compared to normal children (NL, n = 11), especially under stress of rapid stimulation. Children were chosen for this study based on performance on measures of reading and spelling and measures of syllable discrimination. In response to the onset of the speech sound /da/, wave V-V(n) of the auditory brainstem response (ABR) had a significantly shallower slope in LP children, suggesting longer duration and/or smaller amplitude. The amplitude of the frequency following response (FFR) was diminished in LP subjects over the 229-686 Hz range, which corresponds to the first formant of the/da/ stimulus, while activity at 114 Hz, representing the fundamental frequency of /da/, was no different between groups. Normal indicators of auditory peripheral integrity suggest a central, neural origin of these differences. These data suggest that poor representation of crucial components of speech sounds could contribute to difficulties with higher-level language processes.

Child↗
PubMed1981

The effect of barriers on spatial representations.

48 first graders and 48 fifth graders learned the arrangement of 6 objects in a novel environment containing 2 large barriers. 3 nonoverlapping paths, each connecting all locations, were devised. Children were assigned to 4 experimental conditions representing the factorial combination of number of trials of walking (2 vs. 3) and path experience (same vs. different). Those in the same path-experience group walked the same path over trials, while those in the different path-experience group walked a different path on each trial. Following these walking experiences, the children used markers to reconstruct the environmental configuration from memory. Based on the interpoint distances calculated from these placements, the presence of barriers led to overestimations of distance only for those children who received less experience, which was distributed across a variety of paths (different path experience/2-trials group). These overestimations occurred primarily for barrier-present interpoint distances viewed but not directly walked. These results are discussed in terms of the opportunities for coordinating and integrating spatial perspectives provided by these environmental experiences.

Child↗