Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “REINFORCEMENT 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 1,549 records · Page 86Linked to original sources
PubMed1962

The functional properties of a time out from an avoidance schedule.

Investigations are reported of some of the discriminative (cue) and reinforcing properties of stimuli (TO(av)) correlated with the discontinuation of a free-operant avoidance schedule. After several unsuccessful initial attempts to demonstrate the positively reinforcing effect of TO(av), a special technique was developed. The subjects were trained by means of a free-operant avoidance procedure to press a lever and postpone electric shock; by pressing a second lever, they could produce a cue correlated with the TO(av) period. The importance of a distinct cue correlated with TO(av) was demonstrated in experiments in which this signal was eliminated from the procedure. The reinforcing function of the TO(av) period was explored by showing its effectiveness under various schedules of reinforcement. Other observations were made concerning the course of extinction of both TO(av)-producing responses and avoidance behavior. In further experiments in which TO(av) was used as a reinforcer, the TO(av) behavior was also brought under separate stimulus control. The development of the various discriminations involved in the resulting performance were described.

Conditioning, Operant↗
PubMedDate not supplied

Enhanced passive avoidance learning and appetitive T-maze learning with post-trial rewarding hypothalamic stimulation.

Experiments were carried out to investigate the effects of post-trial reinforcing stimulation of the lateral hypothalamus on learning in rats. The reinforcing stimulation was always presented for a duration of 20--30 sec (0.2 sec on/0.8 sec off), and was administered either immediately, 30 sec delayed or 300 sec delayed after exposure to the learning situation. In experiment 1 post-trial stimulation led to improved passive avoidance learning of an alcove-avoidance task when presented 30 sec compared to immediately after the footshock. In Experiment 2 reversal learning of a one-way shuttle-box avoidance task was facilitated by 30 sec delayed, but not 300 sec delayed post-trial reinforcing stimulation. In Experiment 3 appetitive left-right discrimination was investigated using a T-maze task. Thirty sec delayed post-trial reinforcing stimulation presented contingent on errors facilitated learning of this task. Together, the 3 studies provide further support for the hypothesis that reinforcers directly influence labile memory processes (such as short-term memory) and thereby improve learning.

Animals↗
PubMed1999

Learning from mistakes.

We re-examine the commonly held view that learning and memory necessarily require potentiation of synapses. A simple neuronal model of self-organized learning with no positive reinforcement is presented. The strongest synapses are selected for propagation of activity. Active synaptic connections are temporarily "tagged" and subsequently depressed if the resulting output turns out to be unsuccessful. Thus, all learning occurs by mistakes. The model operates at a highly adaptive state with low activity. Previously stored patterns may be swiftly retrieved when the environment and the demands of the brain change. The combined process of: (i) activity selection by extremal "winner-take-all" dynamics; and (ii) the subsequent weeding out of synapses may be viewed as synaptic Darwinism. We argue that all the features of the model are biologically plausible and discuss our results in light of recent experiments by Fitzsimonds et al. on back-propagation of long-term depression, by Xu et al. on facilitation of long-term depression in the hippocampus by behavioural stress, and by Frey and Morris on synaptic tagging.

Brain↗
PubMed2004

[Medical studies at the University of Liège: renewing the art of teaching].

The Faculty of Medicine of the University of Liège has undergone since 2000, a striking change in the medical school, aiming at a better correlation between medical education and community needs and a shift from teaching to learning approach for education. Consequently, teaching matters related to the community medical practice interface have been reinforced. Furthermore, problem based learning has been implemented all along the cursus.

Curriculum↗
PubMed1997

Serial learning in the pigeon in a free correction procedure: role of reinforcing values.

Pigeons were shown four chromatic stimuli simultaneously (A, B, C, D) in daily 15-min sessions. A free correction procedure was used: production of the series A-->B-->C resulted in food delivery and ended a trial, while incorrect colour pecks had no consequences. The learning criterion (five consecutive correct series) was not achieved after 50 sessions, although the number of correct responses increased progressively. From around session 10, birds produced primarily transitions between items B and C or A and C and repeats on items B and C (BC 35%, AC 15%, BB 10%, CC 25%). Despite the low rate of the previously acquired A-->B transition (6%), its production was highly correlated with that of the A-->B-->C series. On the other hand, item C was predominantly pecked and a recency effect occurred for this item. These data extend previous findings indicating that acquisition of series in pigeons involves the formation of associative links between stimuli according to direct and indirect reinforcing values.

Animals↗
PubMed1982

Sensory reinforcement in the mentally handicapped and autistic child: a review.

Sensory reinforcement was first studied by learning theorists working with animals in the 1950s. Attempts to examine the phenomenon with children followed in the 1960s, and the studies demonstrated that sensory stimuli could act like any other reinforcers with normal young children. Similar work with the autistic and mentally handicapped child arose in relation to both the study of receptor development and more treatment-oriented research. It now seems that even profoundly handicapped children can learn to operate simple levers when reinforced by sensory stimuli, and some handicapped children have learned quite complex skills through sensory reinforcement. There also appears to be a close relationship between stereotyped behavior and sensory reinforcement. The clinical implications of the studies reviewed are discussed.

Adolescent↗
PubMed1997

Contribution of CNS nicotine metabolites to the neuropharmacological effects of nicotine and tobacco smoking.

Nicotine, the principal alkaloid in tobacco products, is generally accepted to be the active pharmacological agent responsible for CNS effects resulting from tobacco use. Arguments are presented in this commentary which take issue with this popular dogma, by providing evidence that nicotine metabolites may also be responsible for the CNS effects commonly attributed to nicotine. CNS effects attributed to nicotine include reinforcing effects, mood elevation, arousal, locomotor stimulant effects, and learning and memory enhancement. The reinforcing and locomotor stimulant effects of nicotine have been suggested to be the result of activation of CNS dopaminergic systems, and nicotine-induced modulation of dopaminergic neurotransmission has been studied in detail. Nicotine acts at a family of nicotinic receptor subtypes composed of multiple subunits; however, the exact composition of the subunits in native nicotinic receptors and the functional significance of the receptor subtype diversity are currently unknown. This nicotinic subtype diversity increases the complexity of the potential mechanisms of action of nicotine and its metabolites. Although peripheral metabolism of nicotine has been studied extensively, metabolism in the CNS has not been investigated to any great extent. Recently, studies from our laboratory have demonstrated that several nicotine metabolites are present in the CNS after acute nicotine administration. Moreover, nicotine metabolites are pharmacologically active in neurochemical and behavioral assays. Thus, CNS effects resulting from nicotine exposure may not be due solely to nicotine, but may result, at least in part, from the actions of nicotine metabolites.

Animals↗
PubMed1979

A behavioral version of catecholamine involvement in new learning rewarded by brain stimulation.

Numerous studies in the rat indicate that catecholamines (CA) mediate rewarding properties of self-administered electrical stimulation to the brain. One such property is the learning of new response-reinforcement relationships. In the present experiment, amphetamine which potentiates CA at the synapse produces stereotypical responding but does not interfere with the learning of new response-reinforcement relationships. Apomorphine, which mimics dopamine (DA) at DA receptors, also produces stereotypy and interferes with learning. The results suggest that DA released by stimulation mediates the stereotyped responding seen in intracranial self-stimulation (ICS) but norepinephrine mediates reward of newly learned responses.

Amphetamine↗
PubMed1968

The discrimination of stimulus duration by pigeons.

Pigeons were trained to discriminate the duration of a stimulus. One response in a psychophysical choice situation was reinforced, given the immediately prior presentation of a stimulus duration in one class of durations called short durations, and the other response was reinforced given the immediately prior presentation of a stimulus duration in a second class called long durations. Durations of equal logarithmic difference from the cutoff, whether in the short or long class, yielded equal accuracy. Accuracy was a function not only of the properties of the stimuli to be discriminated, but also of the experimental contingencies used. Accuracy was greater in variable-ratio than in fixed-ratio schedules of reinforcement of the discriminative responses, and was lower at the beginning than later in individual fixed ratios. Proportion of short or long responses (response bias) was affected by sequential dependencies among long and short durations and was effectively controlled through the use of asymmetric reinforcement and fixed-ratio contingencies.

Animals↗