Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Ensemble 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 145 records · Page 8Linked to original sources
PubMed1993

Why spikes? Hebbian learning and retrieval of time-resolved excitation patterns.

Hebbian learning allows a network of spiking neurons to store and retrieve spatio-temporal patterns with a time resolution of 1 ms, despite the long postsynaptic and dendritic integration times. To show this, we introduce and analyze a model of spiking neurons, the spike response model, with a realistic distribution of axonal delays and with realistic postsynaptic potentials. Learning is performed by a local Hebbian rule which is based on the synchronism of presynaptic neurotransmitter release and some short-acting postsynaptic process. The time window of this synchronism determines the temporal resolution of pattern retrieval, which can be initiated by applying a short external stimulus pattern. Furthermore, a rate quantization is found in dependence upon the threshold value of the neurons, i.e., in a given time a pattern runs n times as often as learned, when n is a positive integer (n > or = 0). We show that all information about the spike pattern is lost if only mean firing rates (temporal average) or ensemble activities (spatial average) are considered. An average over several retrieval runs in order to generate a post-stimulus time histogram may also deteriorate the signal. The full information on a pattern is contained in the spike raster of a single run. Our results stress the importance, and advantage, of coding by spatio-temporal spike patterns instead of firing rates and average ensemble activity. The implications regarding modelling and experimental data analysis are discussed.

Animals↗
PubMed2003

Training in cortical control of neuroprosthetic devices improves signal extraction from small neuronal ensembles.

We have recently developed a closed-loop environment in which we can test the ability of primates to control the motion of a virtual device using ensembles of simultaneously recorded neurons /29/. Here we use a maximum likelihood method to assess the information about task performance contained in the neuronal ensemble. We trained two animals to control the motion of a computer cursor in three dimensions. Initially the animals controlled cursor motion using arm movements, but eventually they learned to drive the cursor directly from cortical activity. Using a population vector (PV) based upon the relation between cortical activity and arm motion, the animals were able to control the cursor directly from the brain in a closed-loop environment, but with difficulty. We added a supervised learning method that modified the parameters of the PV according to task performance (adaptive PV), and found that animals were able to exert much finer control over the cursor motion from brain signals. Here we describe a maximum likelihood method (ML) to assess the information about target contained in neuronal ensemble activity. Using this method, we compared the information about target contained in the ensemble during arm control, during brain control early in the adaptive PV, and during brain control after the adaptive PV had settled and the animal could drive the cursor reliably and with fine gradations. During the arm-control task, the ML was able to determine the target of the movement in as few as 10% of the trials, and as many as 75% of the trials, with an average of 65%. This average dropped when the animals used a population vector to control motion of the cursor. On average we could determine the target in around 35% of the trials. This low percentage was also reflected in poor control of the cursor, so that the animal was unable to reach the target in a large percentage of trials. Supervised adjustment of the population vector parameters produced new weighting coefficients and directional tuning parameters for many neurons. This produced a much better performance of the brain-controlled cursor motion. It was also reflected in the maximum likelihood measure of cell activity, producing the correct target based only on neuronal activity in over 80% of the trials on average. The changes in maximum likelihood estimates of target location based on ensemble firing show that an animal's ability to regulate the motion of a cortically controlled device is not crucially dependent on the experimenter's ability to estimate intention from neuronal activity.

Action Potentials↗
PubMed2006

Intelligent classification of electrocardiogram (ECG) signal using extended Kalman Filter (EKF) based neuro fuzzy system.

This study presents the development of a hybrid system consisting of an ensemble of Extended Kalman Filter (EKF) based Multi Layer Perceptron Network (MLPN) and a one-pass learning Fuzzy Inference System using Look-up Table Scheme for the recognition of electrocardiogram (ECG) signals. This system can distinguish various types of abnormal ECG signals such as Ventricular Premature Cycle (VPC), T wave inversion (TINV), ST segment depression (STDP), and Supraventricular Tachycardia (SVT) from normal sinus rhythm (NSR) ECG signal.

Algorithms↗
PubMed2026

Predicting enhancer-promoter interactions using a stacking-based ensemble strategy.

MOTIVATION: Enhancer-promoter interactions (EPIs) are essential for gene regulation and disease progression. Recent studies have shown that distal enhancers can regulate target genes through interactions with nearby promoters, providing important insights into transcriptional regulation mechanisms. Although high-throughput experimental techniques have enabled large-scale identification of EPIs, these methods are often costly and time-consuming. In addition, existing computational approaches still face challenges in effectively integrating heterogeneous feature representations from different cell lines. RESULTS: We propose a stacked ensemble framework for EPI prediction that integrates feature representations from diverse cell line datasets using multiple machine learning algorithms. The extracted complementary patterns are further combined by an XGBoost classifier to improve robustness against overfitting. Experiments on six independent datasets show that the proposed method achieves superior accuracy and generalization compared with existing EPI prediction models, with an average AUROC of 0.909 while maintaining computational efficiency. AVAILABILITY: The source code and its archived release are available at GitHub and Zenodo. The Zenodo archive provides a versioned snapshot of the repository: https://zenodo.org/records/19952998.

Promoter Regions, Genetic↗
PubMed2001

Learning a grating discrimination task broadens human spatial frequency tuning.

The effect of spatial frequency discrimination learning on spatial frequency detection tuning curves, obtained by a summation to threshold paradigm, has been investigated. Three human observers were exposed to a grating discrimination task for longer than two weeks, and their detection thresholds for compound Gabor gratings were measured before and after this time interval. Discrimination thresholds decreased continuously and substantially during the course of learning, while the spatial frequency detection tuning curves show significant broadening in the posttest. Calculating the discrimination resolution of an ensemble of sensory coding units shows that larger bandwidths lead to better spatial frequency discrimination performance if pattern discrimination rests on multidimensional comparison or one-dimensional scaling of the spatial frequency parameter. Further, it is shown that a multiple-mechanism nonlinear pooling model is capable of explaining the results if plasticity of coding unit bandwidth or adaptive weights of the coding unit responses at the stage of response integration is assumed. The alternative sources of plasticity and the consequences of the findings for psychophysical modeling are discussed.

Computer Simulation↗
PubMed2006

Are sparse-coding simple cell receptive field models physiologically plausible?

Olshausen and Field (1996) developed a simple cell receptive field model for natural scene processing in V1, based on unsupervised learning and non-orthogonal basis function optimization of an overcomplete representation of visual space. The model was originally tested with an ensemble of whitened natural scenes, simulating pre-cortical filtering in the retinal ganglia and lateral geniculate nucleus, and the basis functions qualitatively resembled the orientation-specific responses of V1 simple cells in the spatial domain. In this study, the quantitative tuning responses of the basis functions in the spectral domain are estimated using a Gaussian model, to determine their goodness-of-fit to the known bandwidths of simple cells in primate V1. Five simulation experiments which examined key features of the model are reported: changing the size of the basis functions; using a complete versus over-complete representation; changing the sparseness factor; using a variable learning rate; and mapping the basis functions with a whitening spatial function. The key finding of this study is that across all image themes, basis function sizes, number of basis functions, sparseness factors and learning rates, the spatial-frequency tuning did not closely resemble that of primate area 17 -- the model results more closely resembled the unclassified cat neurones of area 19 with a single exception, and not area 17 as predicted.

Action Potentials↗
PubMed2004

Olfactory learning.

The olfactory nervous systems of insects and mammals exhibit many similarities, suggesting that the mechanisms for olfactory learning may be shared. Neural correlates of olfactory memory are distributed among many neurons within the olfactory nervous system. Perceptual olfactory learning may be mediated by alterations in the odorant receptive fields of second and/or third order olfactory neurons, and by increases in the coherency of activity among ensembles of second order neurons. Operant olfactory conditioning is associated with an increase in the coherent population activity of these neurons. Olfactory classical conditioning increases the odor responsiveness and synaptic activity of second and perhaps third order neurons. Operant and classical conditioning both produce an increased responsiveness to conditioned odors in neurons of the basolateral amygdala. Molecular genetic studies of olfactory learning in Drosophila have revealed numerous molecules that function within the third order olfactory neurons for normal olfactory learning.

Action Potentials↗
PubMedDate not supplied

A theory of pragmatic information and its application to the quasi-species model of biological evolution.

'Standard' information theory says nothing about the semantic content of information. Nevertheless, applications such as evolutionary theory demand consideration of precisely this aspect of information, a need that has motivated a largely unsuccessful search for a suitable measure of an 'amount of meaning'. This paper represents an attempt to move beyond this impasse, based on the observation that the meaning of a message can only be understood relative to its receiver. Positing that the semantic value of information is its usefulness in making an informed decision, we define pragmatic information as the information gain in the probability distributions of the receiver's actions, both before and after receipt of a message in some pre-defined ensemble. We then prove rigorously that our definition is the only one that satisfies obvious desiderata, such as the additivity of information from logically independent messages. This definition, when applied to the information 'learned' by the time evolution of a process, defies the intuitions of the few previous researchers thinking along these lines by being monotonic in the uncertainty that remains after receipt of the message, but non-monotonic in the Shannon entropy of the input ensemble. It also follows that the pragmatic information of the genetic 'messages' in an evolving population is a global Lyapunov function for Eigen's quasi-species model of biological evolution. A concluding section argues that a theory such as ours must explicitly acknowledge purposeful action, or 'agency', in such diverse fields as evolutionary theory and finance.

Biological Evolution↗
PubMed1998

Development of localized oriented receptive fields by learning a translation-invariant code for natural images.

Neurons in the mammalian primary visual cortex are known to possess spatially localized, oriented receptive fields. It has previously been suggested that these distinctive properties may reflect an efficient image encoding strategy based on maximizing the sparseness of the distribution of output neuronal activities or alternately, extracting the independent components of natural image ensembles. Here, we show that a strategy for transformation-invariant coding of images based on a first-order Taylor series expansion of an image also causes localized, oriented receptive fields to be learned from natural image inputs. These receptive fields, which approximate localized first-order differential operators at various orientations, allow a pair of cooperating neural networks, one estimating object identity ('what') and the other estimating object transformations ('where'), to simultaneously recognize an object and estimate its pose by jointly maximizing the a posteriori probability of generating the observed visual data. We provide experimental results demonstrating the ability of such networks to factor retinal stimuli into object-centred features and object-invariant transformation estimates.

Animals↗
PubMed2026

GraphyloVar: predicting the impact of non-coding variants using a multi-species sequence model.

MOTIVATION: Understanding the functional impact of genetic variants is a key problem for precision medicine. Tools like CADD, PhyloP, and PhastCons are useful, but they often look at each position in the genome in isolation. This means they can miss important information from the evolutionary history that connects different species. In this paper, we extend our previous model, Graphylo, to predict the effects of variants. Our new model, GraphyloVar, is built to directly utilize the phylogenetic tree that relates the species. RESULTS: GraphyloVar is a deep learning model that considers both DNA sequence and evolutionary patterns from many species. It uses two main components: Graph Convolutional Networks (GCNs) to process the phylogenetic tree, and Transformer encoders to extract features from the DNA sequences. Pre-trained to predict population-level allele frequencies on the TOPMed whole-genome sequencing cohort, GraphyloVar achieves an AUROC of 0.6246 zero-shot on &#x223c;149M held-out variants, and an ensemble with CADD reaches 0.6442 (+0.020, P<10-15). Fine-tuned GraphyloVar achieves the highest AUROC across all 13 MPRA benchmark datasets. By integrating deep learning with explicit phylogenetic input, GraphyloVar offers a powerful and complementary approach to variant effect prediction that utilizes the full evolutionary history from many species to better identify and prioritize important non-coding variants. AVAILABILITY AND IMPLEMENTATION: Code and datasets are available at https://github.com/DongjoonLim/GraphyloVar under DOI: 10.5281/zenodo.20616818.

Phylogeny↗
PubMed1991

Are memory traces localized or distributed?

Evidence supports the view that "memory traces" are formed in the hippocampus and in the cerebellum in classical conditioning of discrete behavioral responses (e.g. eyeblink conditioning). In the hippocampus, learning results in long-lasting increases in excitability of pyramidal neurons that appear to be localized to these neurons (i.e. changes in membrane properties and receptor function). However, these learning-altered pyramidal neurons are distributed widely throughout CA3 and CA1. Although it plays a key role in certain aspects of classical conditioning, the hippocampus is not necessary for learning and memory of the basic conditioned responses. The cerebellum and its associated brain stem circuitry, on the other hand, does appear to be essential (necessary and sufficient) for learning and memory of the conditioned response. Evidence to date is most consistent with a localized trace in the interpositus nucleus and multiple localized traces in cerebellar cortex, each involving relatively large ensembles of neurons. Perhaps "procedural" memory traces are relatively localized and "declarative" traces more widely distributed.

Animals↗
PubMed1999

Using neural networks and genetic algorithms to enhance performance in an electronic nose.

Sensitivity, repeatability, and discernment are three major issues in any classification problem. In this study, an electronic nose with an array of 32 sensors was used to classify a range of odorous substances. The collective time response of the sensor array was first partitioned into four time segments, using four smooth time-windowing functions. The dimension of the data associated with each time segment was then reduced by applying the Karhunen-Loéve (truncated) expansion (KLE). An ensemble of the reduced data patterns was then used to train a neural network (NN) using the Levenberg-Marquardt (LM) learning method. A genetic algorithm (GA)-based evolutionary computation method was used to devise the appropriate NN training parameters, as well as the effective database partitions/features. Finally, it was shown that a GA-supervised NN system (GANN) outperforms the NN-only classifier, for the classes of the odorants investigated in this study (fragrances, hog farm air, and soft beverages).

Algorithms↗
PubMed2002

On the amazing olivocerebellar system.

Over the last four decades elegant sets of a single-cell studies, originating from various research groups, have contributed significantly to our understanding of olivary and cerebellar physiology. Nevertheless questions relating to the dynamic properties of olivocerebellar network, as a system, remain unsolved. We may be reaching the limits of what can be learned using the single-cell recordings. Further research on this subject may require study of the spatiotemporal activity profiles of ensemble neuronal activity. This paper summarizes results obtained using voltage-sensitive dye imaging in inferior olive slices, and the use of mathematical modeling to address such activity profiles.

Animals↗
PubMed2026

Machine Learning and Metabolomics to Characterize Warburg-Like Metabolic Subtypes in Human Retinal Endothelial Cells Exposed to Risk Factors Associated With Proliferative Diabetic Retinopathy.

PURPOSE: High glucose (HG), hypoxia (Hyp), and their combination are major risk factors for proliferative diabetic retinopathy (PDR). Although these conditions induce features of the Warburg-like metabolic reprogramming in human retinal endothelial cells (HRECs), it remains unclear whether they produce distinct metabolic and angiogenic subtypes. This study aimed to characterize the Warburg-like-associated metabolic heterogeneity induced by these PDR-related risk factors and evaluate the ability of supervised machine-learning models to distinguish these subtypes. METHODS: HRECs were cultured under normoglycemic, HG, Hyp (2% O2), and combined HG-Hyp conditions. Untargeted LC-MS/MS metabolomics quantified metabolites spanning carbohydrates, amino acids, nucleotides, and lipids. Principal component analysis (PCA) assessed overall metabolic variation, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis identified metabolic pathways associated with angiogenesis. In vitro angiogenesis assays measured endothelial tube formation and branching. Nine supervised classifiers (decision tree, logistic regression, na&#xef;ve Bayes, random forest, K-Nearest Neighbors, neural network, gradient boosting, AdaBoost, and Support Vector Machine) were trained on the highest-ranked metabolites selected by the Information Gain Ratio feature-ranking approach. Model performance was evaluated using 10-fold cross-validation, leave-one-out cross-validation (LOOCV), permutation testing, and a classifier stability analysis under biologically meaningful distributional shift using an independent chemically induced hypoxia model (CoCl2). RESULTS: PCA revealed partial separation of metabolic profiles across conditions, indicating different Warburg-like metabolic subtypes. The combined HG-Hyp condition exhibited enhanced angiogenic potential relative to either HG or Hyp alone. KEGG pathway enrichment analysis identified fatty acid biosynthesis and elongation among the most significantly enriched pathways in HRECs under combined HG-Hyp conditions, alongside amino sugar and nucleotide sugar metabolism, glycerophospholipid metabolism, the pentose phosphate pathway, and glycolysis/gluconeogenesis. Supervised machine-learning classifiers distinguished these metabolic subtypes, with AdaBoost and gradient Boosting showing the most balanced, reproducible performance across 10-fold cross-validation, LOOCV, and permutation testing, and remaining the most reliable classifiers under domain-shift testing (area under the curve = 0.88, P = 0.0061). CONCLUSIONS: In this exploratory analysis, HG, Hyp, and their combination drive metabolically and functionally distinct subtypes of Warburg-like metabolic reprogramming in HRECs, with HG-Hyp in combination producing a highly angiogenic phenotype. Boosting-based ensemble classifiers provide a promising framework for detecting these subtypes even under domain-shift conditions, warranting validation in larger independent datasets. TRANSLATIONAL RELEVANCE: Integrating metabolomics with machine-learning classification offers a strategy to identify Warburg-like metabolic subtypes in retinal endothelial cells, providing insights into angiogenic mechanisms and guiding the development of targeted diagnostics or therapeutics for PDR.

Humans↗
PubMed2005

Recognizing partially occluded, expression variant faces from single training image per person with SOM and soft kappa-NN ensemble.

Most classical template-based frontal face recognition techniques assume that multiple images per person are available for training, while in many real-world applications only one training image per person is available and the test images may be partially occluded or may vary in expressions. This paper addresses those problems by extending a previous local probabilistic approach presented by Martinez, using the self-organizing map (SOM) instead of a mixture of Gaussians to learn the subspace that represented each individual. Based on the localization of the training images, two strategies of learning the SOM topological space are proposed, namely to train a single SOM map for all the samples and to train a separate SOM map for each class, respectively. A soft kappa nearest neighbor (soft kappa-NN) ensemble method, which can effectively exploit the outputs of the SOM topological space, is also proposed to identify the unlabeled subjects. Experiments show that the proposed method exhibits high robust performance against the partial occlusions and variant expressions.

Algorithms↗
PubMed1996

Long-term potentiation and learning.

Long-term potentiation (LTP), a relatively long-lived increase in synaptic strength, remains the mot popular model for the cellular process that may underlie information storage within neural systems. The strongest arguments for a role of LTP in memory are theoretical and involve Hebb's Postulate, Marr's theory of hippocampal function, and neural network theory. Considering LTP research as a whole, few studies have addressed the essential question: Is LTP a process involved in learning and memory? The present manuscript reviews research that attempts to link LTP with learning and memory, focusing on studies utilizing electrophysiological, pharmacological, and molecular biological methodologies. Most evidence firmly supports a role for LTP in learning memory. However, an unequivocal experimental demonstration of a contribution of LTP to memory is hampered by our lack of knowledge of the biological basis of memory and of the ways in which memories are represented in ensembles of neurons, the existence of a variety of cellular forms of LTP, and the likely resistance of distributed memory stores to degradation by treatments that incompletely disrupt LTP.

Animals↗
PubMed2006

The representation of information about faces in the temporal and frontal lobes.

Neurophysiological evidence is described showing that some neurons in the macaque inferior temporal visual cortex have responses that are invariant with respect to the position, size and view of faces and objects, and that these neurons show rapid processing and rapid learning. Which face or object is present is encoded using a distributed representation in which each neuron conveys independent information in its firing rate, with little information evident in the relative time of firing of different neurons. This ensemble encoding has the advantages of maximising the information in the representation useful for discrimination between stimuli using a simple weighted sum of the neuronal firing by the receiving neurons, generalisation and graceful degradation. These invariant representations are ideally suited to provide the inputs to brain regions such as the orbitofrontal cortex and amygdala that learn the reinforcement associations of an individual's face, for then the learning, and the appropriate social and emotional responses, generalise to other views of the same face. A theory is described of how such invariant representations may be produced in a hierarchically organised set of visual cortical areas with convergent connectivity. The theory proposes that neurons in these visual areas use a modified Hebb synaptic modification rule with a short-term memory trace to capture whatever can be captured at each stage that is invariant about objects as the objects change in retinal view, position, size and rotation. Another population of neurons in the cortex in the superior temporal sulcus encodes other aspects of faces such as face expression, eye gaze, face view and whether the head is moving. These neurons thus provide important additional inputs to parts of the brain such as the orbitofrontal cortex and amygdala that are involved in social communication and emotional behaviour. Outputs of these systems reach the amygdala, in which face-selective neurons are found, and also the orbitofrontal cortex, in which some neurons are tuned to face identity and others to face expression. In humans, activation of the orbitofrontal cortex is found when a change of face expression acts as a social signal that behaviour should change; and damage to the orbitofrontal cortex can impair face and voice expression identification, and also the reversal of emotional behaviour that normally occurs when reinforcers are reversed.

Animals↗