Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “sparse 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 199 records · Page 11Linked to original sources
PubMed2005

Degenerate coding in neural systems.

When the dimensionality of a neural circuit is substantially larger than the dimensionality of the variable it encodes, many different degenerate network states can produce the same output. In this review I will discuss three different neural systems that are linked by this theme. The pyloric network of the lobster, the song control system of the zebra finch, and the odor encoding system of the locust, while different in design, all contain degeneracies between their internal parameters and the outputs they encode. Indeed, although the dynamics of song generation and odor identification are quite different, computationally, odor recognition can be thought of as running the song generation circuitry backwards. In both of these systems, degeneracy plays a vital role in mapping a sparse neural representation devoid of correlations onto external stimuli (odors or song structure) that are strongly correlated. I argue that degeneracy between input and output states is an inherent feature of many neural systems, which can be exploited as a fault-tolerant method of reliably learning, generating, and discriminating closely related patterns.

Animals↗
PubMed2006

Learning weighted metrics to minimize nearest-neighbor classification error.

In order to optimize the accuracy of the Nearest-Neighbor classification rule, a weighted distance is proposed, along with algorithms to automatically learn the corresponding weights. These weights may be specific for each class and feature, for each individual prototype, or for both. The learning algorithms are derived by (approximately) minimizing the Leaving-One-Out classification error of the given training set. The proposed approach is assessed through a series of experiments with UCI/STATLOG corpora, as well as with a more specific task of text classification which entails very sparse data representation and huge dimensionality. In all these experiments, the proposed approach shows a uniformly good behavior, with results comparable to or better than state-of-the-art results published with the same data so far.

Algorithms↗
PubMed2026

RCoxNet: A Deep Learning Framework Integrating Random Walk with Restart, Mutation, and Clinical Data for Cancer Survival Prediction.

Accurate survival prediction in cancer remains challenging due to the sparsity of somatic mutation profiles and the failure of existing models to capture higher-order gene-gene dependencies. Network diffusion methods such as Random Walk with Restart (RWR) can propagate mutation signals across protein-protein interaction (PPI) networks to address sparsity, yet their integration within a deep learning Cox survival framework has not been comprehensively benchmarked across multiple cancer cohorts. We present RCoxNet, a deep learning framework that maps somatic mutation profiles onto a ConsensusPathDB-derived PPI network via RWR, selects prognostic genes by log-rank filtering, and processes network-informed mutation scores through three fully connected hidden layers feeding into a Cox proportional hazards output. RCoxNet was evaluated on The Cancer Genome Atlas (TCGA) cohorts for four cancer types (breast invasive carcinoma [BRCA], lung adenocarcinoma [LUNG], glioblastoma multiforme [GBM], and ovarian serous cystadenocarcinoma [OV]) using 20 independent random splits. The model achieved mean C-index values of 0.807 ± 0.044 (BRCA), 0.750 ± 0.039 (LUNG), 0.704 ± 0.041 (GBM), and 0.668 ± 0.036 (OV), consistently outperforming DeepSurv, Cox-nnet, SurvivalNet, Cox Elastic-Net (Cox-EN), and DeepHit, with statistically significant gains over Cox-EN, Cox-nnet, SurvivalNet, and DeepHit across the majority of cohorts. RCoxNet demonstrates that embedding sparse mutation profiles into a PPI network context substantially improves cancer survival prediction and yields biologically interpretable prognostic features relevant to precision oncology.

cancer survival prediction↗
PubMed2005

Learning deterministic finite automata with a smart state labeling evolutionary algorithm.

Learning a Deterministic Finite Automaton (DFA) from a training set of labeled strings is a hard task that has been much studied within the machine learning community. It is equivalent to learning a regular language by example and has applications in language modeling. In this paper, we describe a novel evolutionary method for learning DFA that evolves only the transition matrix and uses a simple deterministic procedure to optimally assign state labels. We compare its performance with the Evidence Driven State Merging (EDSM) algorithm, one of the most powerful known DFA learning algorithms. We present results on random DFA induction problems of varying target size and training set density. We also studythe effects of noisy training data on the evolutionary approach and on EDSM. On noise-free data, we find that our evolutionary method outperforms EDSM on small sparse data sets. In the case of noisy training data, we find that our evolutionary method consistently outperforms EDSM, as well as other significant methods submitted to two recent competitions.

Algorithms↗
PubMed1996

Representation of olfactory information in the primate orbitofrontal cortex.

1. To analyze the information represented about individual odor stimuli in the responses of single olfactory neurons in the primate orbitofrontal area, neuronal responses were measured to a set of seven to nine odorants in macaques performing an olfactory discrimination task. The population of neurons analyzed had responses that were significantly differential to the odorants. 2. Information theoretic analyses were applied to the responses of the neurons, and information measures were calculated from the firing rate of the responses and from the principal components of the responses. The information reflected by the firing rate of the response accounted for the majority of the information present (86%) when compared with the information derived from the first three principal components of the spike train. This indicated that temporal encoding had a very minor role in the encoding of olfactory information by single orbitofrontal olfactory cells. 3. The average information about which odorant was presented, averaged across the 38 neurons, was 0.09 bits, a figure that is low when compared with the information values previously published for the responses of temporal lobe face-selective neurons. 4. Application of information theoretic analyses to the responses of these neurons showed how much information about which stimulus was delivered was present in the responses of individual neurons. It was found that for the majority of the neurons significant amounts of information were reflected about one or two of the odorants presented. 5. For each neuron, the information reflected in the responses of that neuron about the reinforcement value and the information about the identity of the odorants were calculated. It is shown that many neurons carry information about which of the odorants was presented; in addition, some neurons reflect information only about the taste association of the stimuli and not about odorant identity. 6. Measurements of the sparseness of the representation indicated that a broadly distributed representation of the identity of odorants was present in this population of neurons.

Animals↗
PubMed2002

Funding the new biologics--what can we learn from infliximab? The CCOHTA report: a gastroenterologist's viewpoint.

The treatment of severe Crohn's disease is difficult, and approximately 20% of patients do not respond to conventional therapy, including corticosteroids and immunosuppressives. Infliximab is one of the only treatments of proven efficacy in this group. Awareness of its benefits and risks is incomplete, because the drug has only recently been introduced and published research data are relatively sparse. Economic analyses help to evaluate the value of interventions that are both effective and expensive, but their validity is compromised by input data that involve questionable assumptions. They should not, therefore, be the only basis for funding decisions. Patients with severe Crohn's disease are frequently unable to be gainfully employed and thus incur significant indirect costs. In a recent study by the Canadian Coordinating Office for Health Technology Assessment (CCOHTA), infliximab was deemed to not meet commonly accepted standards of cost effectiveness. This economic analysis did not incorporate indirect costs, and thus was inherently flawed and likely underestimated infliximab's value. The CCOHTA report also used population data from a period of time during which treatment of Crohn's disease was undergoing major transition. The study population was based in Minnesota and thus might not be applicable to Canada. Although they are routinely used in cost effectiveness models, quality-adjusted life-years gained are difficult to translate into practice, and the health care resources required to induce remission (or some other clinically meaningful result) might be a preferable measure. The CCOHTA report also did not include concomitant therapy with immunosuppressives, despite growing evidence for its benefit. Instead, it considered 'usual care', but, given the lack of effective treatment for many patients with Crohn's disease, such an option constitutes 'no care' and continued suffering. Economic analyses should not be the only basis on which decisions regarding the funding of infliximab, or other new agents, are made.

Antibodies, Monoclonal↗
PubMed1998

Neuronal regulation versus synaptic unlearning in memory maintenance mechanisms.

Hebbian learning, the paradigm of memory formation, needs further mechanisms to guarantee creation and maintenance of a viable memory system. One such proposed mechanism is Hebbian unlearning, a process hypothesized to occur during sleep. It can remove spurious states and eliminate global correlations in the memory system. However, the problem of spurious states is unimportant in the biologically interesting case of memories that are sparsely coded on excitatory neurons. Moreover, if some memories are anomalously strong and have to be weakened to guarantee proper functioning of the network, we show that it is advantageous to do that by neuronal regulation (NR) rather than synaptic unlearning. Neuronal regulation can account for dynamical maintenance of memory systems that undergo continuous synaptic turnover. This neuronal-based mechanism, regulating all excitatory synapses according to neuronal average activity, has recently gained strong experimental support. NR achieves synaptic maintenance over short time scales by preserving the average neuronal input field. On longer time scales it acts to maintain memories by letting the stronger synapses grow to their upper bounds. In ageing, these bounds are increased to allow stronger values of remaining synapses to overcome the loss of synapses that have perished.

Adult↗
PubMed2006

Singing-related activity of identified HVC neurons in the zebra finch.

High vocal center (HVC) is part of the premotor pathway necessary for song production and is also a primary source of input to the anterior forebrain pathway (AFP), a basal ganglia-related circuit essential for vocal learning. We have examined the activity of identified HVC neurons of zebra finches during singing. Antidromic activation was used to identify three classes of HVC cells: neurons projecting to the premotor nucleus RA, neurons projecting to area X in the AFP, and putative HVC interneurons. HVC interneurons are active throughout the song and display tonic patterns of activity. Projection neurons exhibit highly phasic stereotyped firing patterns. X-projecting (HVC((X))) neurons burst zero to four times per motif, whereas RA-projecting neurons burst extremely sparsely--at most once per motif. The bursts of HVC projection neurons are tightly locked to the song and typically have a jitter of <1 ms. Population activity of interneurons, but not projection neurons, was significantly correlated with syllable patterns. Consistent with the idea that HVC codes for the temporal order in the song rather than for sound, the vocal dynamics and neural dynamics in HVC occur on different and uncorrelated time scales. We test whether HVC((X)) neurons are auditory sensitive during singing. We recorded the activity of these neurons in juvenile birds during singing and found that firing patterns of these neurons are not altered by distorted auditory feedback, which is known to disrupt learning or to cause degradation of song already learned.

Acoustic Stimulation↗
PubMed2006

An fMRI investigation of syllable sequence production.

Fluent speech comprises sequences that are composed from a finite alphabet of learned words, syllables, and phonemes. The sequencing of discrete motor behaviors has received much attention in the motor control literature, but relatively little has been focused directly on speech production. In this paper, we investigate the cortical and subcortical regions involved in organizing and enacting sequences of simple speech sounds. Sparse event-triggered functional magnetic resonance imaging (fMRI) was used to measure responses to preparation and overt production of non-lexical three-syllable utterances, parameterized by two factors: syllable complexity and sequence complexity. The comparison of overt production trials to preparation only trials revealed a network related to the initiation of a speech plan, control of the articulators, and to hearing one's own voice. This network included the primary motor and somatosensory cortices, auditory cortical areas, supplementary motor area (SMA), the precentral gyrus of the insula, and portions of the thalamus, basal ganglia, and cerebellum. Additional stimulus complexity led to increased engagement of the basic speech network and recruitment of additional areas known to be involved in sequencing non-speech motor acts. In particular, the left hemisphere inferior frontal sulcus and posterior parietal cortex, and bilateral regions at the junction of the anterior insula and frontal operculum, the SMA and pre-SMA, the basal ganglia, anterior thalamus, and the cerebellum showed increased activity for more complex stimuli. We hypothesize mechanistic roles for the extended speech production network in the organization and execution of sequences of speech sounds.

Adult↗
PubMed1999

Neuronal population activity and functional imaging.

Human functional brain imaging detects blood flow changes that are thought to reflect the activity of neuronal populations and, thus, the responses of neurons that carry behaviourally relevant information. Since this relationship is poorly understood, we explored the link between the activity of single neurons and their neuronal population. The functional imaging results were in good agreement with levels of population activation predicted from the known effects of sensory stimulation, learning and attention on single cortical neurons. However, the nature of the relationship between population activation and single neuron firing was very surprising. Population activation was strongly influenced by those neurons firing at low rates and so was very sensitive to the baseline or 'spontaneous' firing rate. When neural representations were sparse and neurons were tuned to several stimulus dimensions, population activation was hardly influenced by the few neurons whose firing was most strongly modulated by the task or stimulus. Measures of population activation could miss changes in information processing given simultaneous changes in neurons' baseline firing, response modulation or tuning width. Factors that can modulate baseline firing, such as attention, may have a particularly large influence on population activation. The results have implications for the interpretation of functional imaging signals and for cross-calibration between different methods for measuring neuronal activity.

Brain↗
PubMed2002

Bayesian probabilistic network modeling of remifentanil and propofol interaction on wakeup time after closed-loop controlled anesthesia.

OBJECTIVE: Until now, the knowledge of combining anesthetics to obtain an adequate level of anesthesia and to economize wakeup time has been empirical and difficult to represent in quantitative models. Since there is no reason to expect that the effect of non-opioid and opioid anesthetics can be modeled in a simple linear manner, the use of a new computational approach with Bayesian belief network software is demonstrated. METHODS: A data set from a pharmacodynamic study was used where remifentanil was randomly given in three fixed target concentrations (2, 4, and 8 ng/ml) to 62 subjects. Target concentrations of propofol were controlled according to the closed-loop system feedback of the auditory evoked potential index to render modeling unbiased by the level of anesthesia. Time to open eyes was measured to represent wakeup time after surgery. The NETICA version 1.37 software was used on a personal computer for network building, validation, and prediction. RESULTS: After the learning phase, the network was used to generate a series of random cases whose probability distribution matches that of the compiled network. The sampling algorithms used are precise, so that the frequencies of the simulated cases will exactly approach the probabilities of the network and that of the data learned. The graphical display of the predicted wakeup time shows less variability but a more complex interaction pattern than with the unadjusted original data. CONCLUSIONS: Model building and evaluation with Bayesian networks does not depend on underlying linear relationships. Bayesian relationships represent true features of the represented data sample. Data may be sparse, uncertain, stochastic, or imprecise. Multiple platform software that is easy to use is increasingly available. Bayesian networks promise to be versatile tools for building valid, nonlinear, predictive instruments to further gain insight into the complex interaction of anesthetics.

Anesthesia Recovery Period↗
PubMed1999

The use of ethanol to monitor fluid absorption during transurethral resection of the prostate.

Ethanol monitoring is a method for assessing fluid absorption during transurethral resection of the prostate. The method is based on the principle that absorption can be measured by expired-breath tests provided that an irrigating fluid containing ethanol 1% is used. A nomogram for transformation of the ethanol data into meaningful indices of fluid absorption, such as the volume of irrigant absorbed and the degree of hyponatraemia, has been approved for routine use within the European Union. To make a correct interpretation, however, one must learn how to distinguish the pattern of breath ethanol changes typical of absorption of fluid directly into the blood from the pattern associated with the rare intraoperative events in which fluid is deposited in a pool in the retro- or intraperitoneal space. The ethanol method is well documented, but it is still sparsely used outside Scandinavia, due to its lack of a patent and differing opinions concerning the need for measuring fluid absorption.

Absorption↗
PubMed1997

Differentiation and cell polarity during renal cyst formation in the Han:SPRD (cy/+) rat, a model for ADPKD.

Despite the recent positional cloning of genes responsible for autosomal dominant polycystic kidney disease (ADPKD), the exact pathogenetic mechanisms underlying this disorder are still unclear. To learn more about cyst formation, we investigated cell differentiation and cell polarity in the Han:SPRD (cy/+) rat between 21 days and 60 wk of age. At early stages of cyst development, alkaline phosphatase, aquaporin-1, NaSi-1 cotransporter, and Na(+)-K(+)-adenosinetriphosphatase (Na(+)-K(+)-ATPase) were expressed normally. Clusterin mRNA was only sparsely expressed at the onset of cystic degeneration and increased thereafter, being highest in noncystic nephron segments. In cyst wall cells, clusterin on the one hand and alkaline phosphatase, aquaporin-1, NaSi-1-cotransporter, and Na(+)-K(+)-ATPase on the other were expressed in a mutually exclusive fashion. No change in cell polarity could be observed at any stage. Our data therefore argue against a change in cell polarity and against an early arrest in normal tubular development during cyst formation in the Han:SPRD (cy/+) rat model of ADPKD but favor the hypothesis that tubular epithelia develop in an orderly fashion and degenerate thereafter.

Aging↗
PubMed1999

[Comparison of stromal remodeling and keratocyte response after corneal incision and photorefractive keratectomy].

PURPOSE: We investigated the keratocyte response and stromal remodeling after corneal incision and photorefractive keratectomy, respectively to learn the difference between the two surgeries histophysiologically and immunohistochemically. METHODS: We performed corneal incision or photorefractive keratectomy in rabbits or rats, and then we chronologically observed the histological changes and the changes in localization of extracellular matrix proteins. RESULTS: In both types of surgery, the keratocyte population in the damaged stroma became sparse, and the cells were undergoing apoptosis immediately after the procedures. After that, activated keratocytes migrated into the acellular zone, and the cells formed multiple layers at the resurfaced subepithelial regions. Deposition of amorphous substances was observed between migrated keratocytes, and stromal remodeling began. Three months after the surgery, corneal structure had recovered to near normal condition in the corneal incisions. In photorefractive keratectomy, however, strong immunoreactivity of extracellular matrix proteins was observed in the subepithelial regions. CONCLUSIONS: These results suggested that stromal wound healing processes were similar in both corneal incision and photorefractive keratectomy. Corneal incision may induce transient keratocyte response during the stromal remodeling, but photorefractive keratectomy may induce sustained keratocyte response.

Animals↗
PubMed2005

Temporal codes and sparse representations: a key to understanding rapid processing in the visual system.

Where neural information processing is concerned, there is no debate about the fact that spikes are the basic currency for transmitting information between neurons. How the brain actually uses them to encode information remains more controversial. It is commonly assumed that neuronal firing rate is the key variable, but the speed with which images can be analysed by the visual system poses a major challenge for rate-based approaches. We will thus expose here the possibility that the brain makes use of the spatio-temporal structure of spike patterns to encode information. We then consider how such rapid selective neural responses can be generated rapidly through spike-timing-dependent plasticity (STDP) and how these selectivities can be used for visual representation and recognition. Finally, we show how temporal codes and sparse representations may very well arise one from another and explain some of the remarkable features of processing in the visual system.

Action Potentials↗
PubMed2006

Parameter estimation for stiff equations of biosystems using radial basis function networks.

BACKGROUND: The modeling of dynamic systems requires estimating kinetic parameters from experimentally measured time-courses. Conventional global optimization methods used for parameter estimation, e.g. genetic algorithms (GA), consume enormous computational time because they require iterative numerical integrations for differential equations. When the target model is stiff, the computational time for reaching a solution increases further. RESULTS: In an attempt to solve this problem, we explored a learning technique that uses radial basis function networks (RBFN) to achieve a parameter estimation for biochemical models. RBFN reduce the number of numerical integrations by replacing derivatives with slopes derived from the distribution of searching points. To introduce a slight search bias, we implemented additional data selection using a GA that searches data-sparse areas at low computational cost. In addition, we adopted logarithmic transformation that smoothes the fitness surface to obtain a solution simply. We conducted numerical experiments to validate our methods and compared the results with those obtained by GA. We found that the calculation time decreased by more than 50% and the convergence rate increased from 60% to 90%. CONCLUSION: In this work, our RBFN technique was effective for parameter optimization of stiff biochemical models.

Algorithms↗
PubMedDate not supplied

Role of the lecturer practitioner: 1.

The lecturer practitioner role is a practice-led educational development which has been sparsely studied. This two-part article aims to show that the role is central to dynamic development within the nursing profession and that, despite many potential problems, it can be both rewarding and valuable. Part 1 describes the development of the lecturer practitioner role and examines various aspects of the role, including clinical learning and assessment.

Education, Nursing, Continuing↗
PubMed1982

Identifying the health risks from very low-dose sparsely ionizing radiation.

The health risks from low-dose sparsely ionizing (low-LET) radiation have been the subject of continued debate. At present, quantitative estimates of risk are extremely uncertain due to the controversy surrounding both the dosimetry for A-bomb survivor data and the choice of mathematical models for extrapolating risk from high to low doses. Nevertheless, much can be learned about the nature of the health risks by reviewing the epidemiologic literature. We present a summary of diseases which have been associated with low-LET radiation (less than 1000 rad) in at least two independent studies, according to the mean cumulative organ dose at which the disease was observed. At organ doses of less than or equal to 50 rad, the only diseases that have been reported consistently are thyroid cancer, salivary gland tumors, and leukemia. The first two diseases were observed in association with x-ray epilation of the scalp for tinea capitis, a therapy which is no longer employed. On the other hand, leukemia has been observed repeatedly to occur at cumulative doses of greater than or equal to 30 rad low-LET radiation.

Dose-Response Relationship, Drug↗