PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “active learning”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources
PubMed2025

Active learning of enhancers and silencers in the developing neural retina.

Deep learning is a promising strategy for modeling cis-regulatory elements. However, models trained on genomic sequences often fail to explain why the same transcription factor can activate or repress transcription in different contexts. To address this limitation, we developed an active learning approach to train models that distinguish between enhancers and silencers composed of binding sites for the photoreceptor transcription factor cone-rod homeobox (CRX). After training the model on nearly all bound CRX sites from the genome, we coupled synthetic biology with uncertainty sampling to generate additional rounds of informative training data. This allowed us to iteratively train models on data from multiple rounds of massively parallel reporter assays. The ability of the resulting models to discriminate between CRX sites with identical sequence but opposite functions establishes active learning as an effective strategy to train models of regulatory DNA. A record of this paper's transparent peer review process is included in the supplemental information.

Retina↗
PubMed2023

Active learning of enhancer and silencer regulatory grammar in photoreceptors.

Cis-regulatory elements (CREs) direct gene expression in health and disease, and models that can accurately predict their activities from DNA sequences are crucial for biomedicine. Deep learning represents one emerging strategy to model the regulatory grammar that relates CRE sequence to function. However, these models require training data on a scale that exceeds the number of CREs in the genome. We address this problem using active machine learning to iteratively train models on multiple rounds of synthetic DNA sequences assayed in live mammalian retinas. During each round of training the model actively selects sequence perturbations to assay, thereby efficiently generating informative training data. We iteratively trained a model that predicts the activities of sequences containing binding motifs for the photoreceptor transcription factor Cone-rod homeobox (CRX) using an order of magnitude less training data than current approaches. The model's internal confidence estimates of its predictions are reliable guides for designing sequences with high activity. The model correctly identified critical sequence differences between active and inactive sequences with nearly identical transcription factor binding sites, and revealed order and spacing preferences for combinations of motifs. Our results establish active learning as an effective method to train accurate deep learning models of cis-regulatory function after exhausting naturally occurring training examples in the genome.

Journal Article↗
PubMed2025

Boolean matrix logic programming for active learning of gene functions in genome-scale metabolic network models.

Reasoning about hypotheses and updating knowledge through empirical observations are central to scientific discovery. In this work, we applied logic-based machine learning methods to drive biological discovery by guiding experimentation. Genome-scale metabolic network models (GEMs) - comprehensive representations of metabolic genes and reactions - are widely used to evaluate genetic engineering of biological systems. However, GEMs often fail to accurately predict the behaviour of genetically engineered cells, primarily due to incomplete annotations of gene interactions. The task of learning the intricate genetic interactions within GEMs presents computational and empirical challenges. To efficiently predict using GEM, we describe a novel approach called Boolean Matrix Logic Programming (BMLP) by leveraging Boolean matrices to evaluate large logic programs. We developed a new system, [Formula: see text], which guides cost-effective experimentation and uses interpretable logic programs to encode a state-of-the-art GEM of a model bacterial organism. Notably, [Formula: see text] successfully learned the interaction between a gene pair with fewer training examples than random experimentation, overcoming the increase in experimental design space. [Formula: see text] enables rapid optimisation of metabolic models to reliably engineer biological systems for producing useful compounds. It offers a realistic approach to creating a self-driving lab for biological discovery, which would then facilitate microbial engineering for practical applications.

Active learning↗
PubMed1992

[Pharmacological procedures for treating enuresis and operant behavior therapy. A comparison].

In comparison with the highly significant effect of the apparative behavior therapy of enuresis, tricyclic antidepressants, desmopressin and anticholinergic drugs all fail to show statistically significant results. The author presents a model to illustrate and explain how the apparative behavior therapy of enuresis works. The therapy's main aim is for patients to learn active control of the bladder. From this theoretical point of view the aforementioned pharmacological treatments of enuresis cannot yield effects that last a long time after discontinuation of medication. By reducing the frequency of micturition at night only, they prevent active learning of bladder control.

Antidepressive Agents, Tricyclic↗
PubMedDate not supplied

Marijuana influenced changes in GSR activation peaking during paired-associate learning.

Activation Peaking (AP) refers to a patterned physiological response occurring during learning. Marijuana has been found to interfere with both paired-associate learning and phasic GSR activity. Therefore, a study was performed to assess the effects of marijuana intoxication on paired-associate learning and concomitant GSR AP. Two marijuana usage categories were employed--light and heavy usage Ss. Within each category four groups were run in a design to test state-dependent effects. Each S was seen twice with a seven-day inter-session interval. The groups were P-P, P-M, M-M and P-M with P equals placebo and M equals 14 mg delta-9 THC. At each sessions S learned a nine-word paired-associate list to a criterion of one correct recitation, and then received 100 percent overlearning. No usage or group differences were found in level of basal conductance, except lights showed habituation over sessions and heavies did not. Magnitude of phasic GSR activation, aligned for AP, was significantly reduced for both heavy usage and marijuana intoxicated Ss. Also, only on placebo days was an AP effect evident. The results were discussed in terms of marijuana's effects on learning and physiology with emphasis on possible mechanisms of action.

Adolescent↗
PubMed1992

Active avoidance learning in old rats chronically treated with levocarnitine acetyl.

The aging laboratory animal is recognized as a suitable experimental model for the investigation on drugs potentially able to retard the age-dependent decline in cognitive functions. There is robust evidence that levocarnitine acetyl (ALCAR), the acetyl derivative of carnitine, when administered chronically, prevents some age-related deficits of the central nervous system, mainly at the hippocampal level. On the basis of this evidence and because learning of active avoidance was demonstrated to become impaired with age, we decided to investigate the effect of ALCAR in rats. For statistical evaluation of results, the Cluster Analysis technique was chosen. This procedure pointed out the great heterogeneity of the old population and allowed the classification of the animals into homogeneous groups according to their response pattern. The effect of ALCAR was evident in the higher number of treated old animals yielding escape responses, indicating that ALCAR can preserve, at least partially, learning and memory from the natural decay occurring with age.

Acetylcarnitine↗
PubMed1978

Proactive and retroactive effects of hippocampal stimulation on active avoidance learning, hippocampal EEG and brain acetylcholinesterase activity in cats.

The subject of this investigation were the effects of electrical stimulation of the hippocampus on the acquisition of active avoidance response (AAR) in a shuttle-box. The stimulation (200 microamperes, 50/s, negative rectangular pulses of 1.0 ms duration) was applied once for 10 s before or after each training session. It was found that the application of hippocampal stimulation before each session facilitated the acquisition of AAR; discontinuation of the stimulation after training did not cause a decrease of AAR performance. Application of the stimulation after each session inhibited learning in four out of six cats. However, the level of AAR performance increased rapidly in these cats after inversion of the trials-stimulation sequence. It was also found that the intensity of the somatic and vegetative symptoms evoked by stimulation (stupor, salivation, twitching of facial muscles, pupil dilatation, crying) increased gradually in successive experimental sessions, suggesting the development of the 'kindling effect in cats stimulated before each session. In cats stimulated after each session the intensity of these symptoms was greatly diminished as compared to sessions where the stimulation was not preceded by the avoidance training, or they did appear at all. However, normal sensitivity to stimulation returned after several applications of hippocampal stimulation before each experimental session. Electroencephalographic studies showed that hippocampal stimulation with the use of the same parameters as those used during training evoked hippocampal afterdischarges lasting 5-60 s. No changes of aceltycholinesterase activity in different brain regions were found in consequence of such stimulation.

Acetylcholinesterase↗
PubMed1979

[Rat brain nuclease activity during learning with emotionally different reinforcement].

Acid and alkaline activity of nucleases of the rats trained with emotional positive or negative reinforcement was estimated in the neocortex, hippocampus, midbrain, and in caudal portions of the brain-stem, using native and denaturated DNA as a substrate. The results showed the total increase in nuclease activity during learning. Nevertheless the dynamics of enzyme activation was different depending on the emotional state of rats during learning. The most active enzyme was found in the caudal portion of the brain-stem.

Animals↗
PubMed1975

Activity and learning in neonatally hormone treated rats.

The degree of activity and the performance in active avoidance learning were determined in adulthood, from neonatal hormone treated and intact rats. The treated groups were androgenized females and castrated males and females. Normal males and females composed the intact groups. Activity was measured in two tests situations, i.e., open field and shuttle box. Similar results were obtained in both tests, and a positive correlation coefficient was found between them. Intact females showed more activity than males; castration and testosterone injection in females decreased their activity. On the contrary, castration in males increased it. The effects of the treatments on the performance in the active avoidance task were similar to those described for activity, with the only exception that castration in males did not produce a significant effect in this test. Nevertheless, no correlation was found between activity and avoidance learning scores. The validity of the classification of different types of behavior as sexual or non-sexual is discussed.

Animals↗