PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Reinforcement 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 937 records · Page 52Linked to original sources
PubMed1996

A fuzzy adaptive learning control network with on-line structure and parameter learning.

This paper addresses a general connectionist model, called Fuzzy Adaptive Learning Control Network (FALCON), for the realization of a fuzzy logic control system. An on-line supervised structure/parameter learning algorithm is proposed for constructing the FALCON dynamically. It combines the backpropagation learning scheme for parameter learning and the fuzzy ART algorithm for structure learning. The supervised learning algorithm has some important features. First of all, it partitions the input state space and output control space using irregular fuzzy hyperboxes according to the distribution of training data. In many existing fuzzy or neural fuzzy control systems, the input and output spaces are always partitioned into "grids". As the number of input/output variables increase, the number of partitioned grids will grow combinatorially. To avoid the problem of combinatorial growing of partitioned grids in some complex systems, the proposed learning algorithm partitions the input/output spaces in a flexible way based on the distribution of training data. Second, the proposed learning algorithm can create and train the FALCON in a highly autonomous way. In its initial form, there is no membership function, fuzzy partition, and fuzzy logic rule. They are created and begin to grow as the first training pattern arrives. The users thus need not give it any a priori knowledge or even any initial information on these. In some real-time applications, exact training data may be expensive or even impossible to obtain. To solve this problem, a Reinforcement Fuzzy Adaptive Learning Control Network (RFALCON) is further proposed. The proposed RFALCON is constructed by integrating two FALCONs, one FALCON as a critic network, and the other as an action network. By combining temporal difference techniques, stochastic exploration, and a proposed on-line supervised structure/parameter learning algorithm, a reinforcement structure/parameter learning algorithm is proposed, which can construct a RFALCON dynamically through a reward/penalty signal. The ball and beam balancing system is presented to illustrate the performance and applicability of the proposed models and learning algorithms.

Algorithms↗
PubMed2001

Functional imaging studies on dopamine and motor control.

In this review, the insights that PET and MR activation studies have given us concerning the role of dopamine in motor control are reviewed. Regional cerebral blood flow (rCBF) changes when normal subjects and Parkinson's disease patients perform simple finger movements, motor sequence learning, problem solving, and tasks financially rewarded according to success are compared. Additionally, dopamine release during rewarded and unrewarded actions, as reflected by relative levels of striatal binding of the reversible dopamine antagonist 11C-raclopride, are contrasted. It is argued that during unrewarded familiar actions tonic dopamine release in the basal ganglia acts to focus and filter cortical output so optimising the running of motor programmes. During motor learning, novel, and financially rewarded tasks additional dopamine is phasically released, acting to both alert the subject and to reinforce motor learning.

Animals↗
PubMed2002

Can a postgraduate course for general practitioners deliver perceived benefit for learners, patients and the NHS?: a qualitative study.

The focus of continuing professional development in general practice is shifting towards professional development and away from following personal areas of interest. Previous work has suggested that much CPD has not had an obvious impact in the three areas of professional development: the needs of individual doctors, patients and the needs of the NHS. We report on the results of a programme of study where developments in all three were perceived as being achieved. This outcome was realized by basing learning around real problems course members encountered in their daily work, using these real situations to identify theory, then reinforcing this learning through practical application.

Clinical Competence↗
PubMed1980

[Electrical activity of the dorsal hippocampus during DRL learning in the rabbit (author's transl)].

The change of electrical activity of the dorsal hippocampus during differential reinforcement of low rate (DRL) learning was examined in six rabbits. Bipolar electrodes were implanted stereotaxically into the dorsal hippocampus (3 mm posterior to bregma, 4-4.5 mm lateral and 7-7.5 mm ventral from bregma). After completion of continuous reinforcement learning, the animals learned DRL on a 10 sec. schedule. Rhythmic theta waves (6-8 Hz, 0.6-1.5 mV) were predominant in the early sessions of the DRL schedule. With further sessions, regular theta waves decreased and desynchronized waves or slightly irregular theta waves increased especially during the 10 second waiting period. Contrary to the report of Bennett et al., synchronized waves were always seen from 1-2 sec. before lever-press until completing of lever-press. Moreover, these preceding synchronized waves became more pronounced and of higher frequencies (8-10 Hz.) as learning proceeded. These high frequency regular waves (alpha-bursts) were seen in 5.6-35.8% of reinforced lever-press in the 10th session, and 79.5-93.8% in the 60th session. Alpha bursts were sometimes seen after 8-9 sec. waiting in the later sessions, but never seen within five secs.. During the extinction of DRL learning, the occurrence rate of alpha-bursts on adequate but non-rewarded lever-pressing decreased. These results cannot be explained satisfactorily in terms of an association between synchronized waves and voluntary movement. Frequency increase was seen when rabbits expected reinforcement after waiting for about 10 sec. or more. These results indicate that frequency increase of hippocampal synchronized waves is seen in association with higher central processing, most likely with memory.

Alpha Rhythm↗
PubMed1980

Independence of sensitivity to relative reinforcement rate and discriminability in signal detection.

Six pigeons were trained to detect differences between two white stimuli, S1 and S2, differing in duration and arranged probabilistically on the center key of a three-key chamber. Detection performance was measured at two levels of discriminability. At one level, S1 was five seconds and S2 was thirty seconds. At the other level, S1 was twenty seconds and S2 was thirty seconds. The procedure was a standard signal-detection yes-no design in which stimulus-presentation probability was varied from .1 to .9 at both discriminability levels. On completion of the center-key stimulus, a peck on the center key darkened the center-key light and turned on the two red side keys. A left-key response was "correct" on S1 trials, and a right-key response was "correct" on S2 trials. Correct responses produced food reinforcement on a variable-ratio 1.3 schedule. Incorrect responses produced three second blackout. Discriminability was higher for the five-second versus thirty-second conditions than for the twenty-second versus thirty-second conditions, but there were no differences in sensitivity of behavior to reinforcement variation for the two stimulus pairs. Response bias was a function of the relative reinforcement rate for correct choice responses.

Animals↗
PubMed1982

Stimulus discriminability in free-operant and discrete-trial detection procedures.

Six pigeons were trained to discriminate different light intensities in four experimental procedures. Experiment 1 compared stimulus discriminability in a yes-no signal-detection task with discriminability measures obtained from two free-operant procedures. Discriminability estimates were significantly lower in the detection procedure. Experiment 2 showed this lowered discriminability to be a function of the delay between stimulus presentation and the availability of the choice-response keys in the standard detection task. In addition, reinforcement sensitivity was lowest when correct choice responses were intermittently, rather than continuously, reinforced.

Animals↗
PubMed1980

Immediate and long-term memory for reinforcement context: the development of learned expectancies in early infancy.

Infants were pretrained to move crib mobiles containing 6 or 10 identical objects by means of footkicking and were then exposed to a reinforcer containing only 2 objects. Relative to infants with no prior history of reinformcement with either of the "larger mobiles, infants shifted to the "smaller" mobile had higher kick rates, and, in addition, their visual attention decreased and negative vocalizations increased. These effects were not transient but persisted 24 hours later in infants experiencing the larger (10- to 2-object) shift. The results suggested that infants develop reward-expectation habits which continue to influence behavior for periods of at least 24 hours.

Arousal↗
PubMed1990

Drugs to decrease alcohol drinking.

A wide variety of drugs have been tested in experimental animals and several have been found that reduce voluntary alcohol drinking. The available evidence suggested that the same drugs also reduce alcohol drinking in alcoholics. Various factors limited or prevented the clinical use of the these drugs. Our working hypothesis has been that alcohol drinking is a learned response, reinforced primarily from alcohol in the brain, and that an alcoholic is a person in which this response and the related craving have become so strong that they dominate the behaviour and interfere with normal functioning. Learned responses are extinguished if they are made repeatedly while the reinforcement is blocked, and opiate antagonists appear to block the reinforcement from alcohol. A series of experiments support the hypothesis that drinking alcohol while an antagonist is present extinguishes the alcohol-drinking response in rats. The antagonists are non-addictive and at least naloxone appears to be safe. Clinical trials are now needed, but the present results suggest that this extinction procedure might be a useful adjunct to the treatment of alcoholism.

Alcohol Drinking↗
PubMed2005

The Edinburgh Electronic Veterinary Curriculum: an online program-wide learning and support environment for veterinary education.

The Edinburgh Electronic Veterinary Curriculum (EEVeC) is a purpose-built virtual learning support environment for the veterinary medicine program at the University of Edinburgh. It is Web based and adapted from a system developed for the human medical curriculum. It is built around a set of databases and learning objects and incorporates features such as course materials, personalized timetables, staff and student contact pages, a notice board, and discussion forums. The EEVeC also contains global or generic resources such as information on quality enhancement and research options. Many of these features contribute to the aim of building a learning community, but a challenge has been to introduce specific features that enhance student learning. One of these is a searchable lecture database in which learning activities such as quizzes and computer-aided learning exercises (CALs) can be embedded to supplement a synopsis of the lecture and address the key needs of integration and reinforcement of learning. Statistics of use indicate extensive student activity during evenings and weekends, with a pattern of increased usage over the years as more features become available and staff and students progressively engage with the system. An essential feature of EEVeC is its flexibility and the way in which it is evolving to meet the changing needs of the teaching program.

Curriculum↗
PubMed1963

A note on chaining and temporal discrimination.

Four pigeons were exposed to a two-key DRL procedure. At the start of a trial, key A was illuminated. A response to the lighted key turned it off and simultaneously illuminated key B. Reinforcement was available for responses on key B which followed the initial key A response by more than 2 sec. In the course of exposure to these conditions, all birds acquired superstitious response chains on key A. The distribution of the number of responses on key A preceding a key B response and the distribution of intervals elapsing from the initial key A response to the key B response were of the same form. The suggestion is made that the superstitious responding on key A served to mediate the required delay interval. However, when intervals between successive key A responses were recorded for one subject, they were found to be regularly spaced in time. Thus, the problem remains of how this behavior is itself timed.

Animals↗
PubMed2026

A systematic review of human avoidance learning: Cognition, computation, and methods.

Avoidance behaviour is fundamental for survival but can become maladaptive in clinical conditions. A large body of literature has accumulated on the dynamics of human avoidance learning. However, current theories and overviews do not provide an exhaustive account of this evidence. In this systematic review, we identify N = 116 studies on human avoidance learning. We analyse these studies with the goal of distilling robust empirical phenomena as a basis for theory-building, and examine their diagnostic value in differentiating between competing theories. We find that the evidence is difficult to reconcile with foundational two-factor and classical safety-signal accounts, and most strongly supports expectancy- and inference-based views, in which avoidance responses are selected with respect to represented consequences. At the same time, no current framework provides a complete account of the evidence: several findings point to an additional role for operant valuation, Pavlovian influences, and contextual or latent-state control over the expression of avoidance. Methodologically, we observe that the problem setting in the most common experimental paradigms is radically simpler than real-world avoidance and therefore unlikely to expose the limits of inferential or reflective mechanisms. Consequently, we argue that paradigms with greater computational demands and more realistic action affordances are required to identify the mechanisms underlying avoidance learning. Collectively, these insights provide a foundation for theoretical refinement, computational modelling, and methodological innovation, with implications for advancing interventions targeting maladaptive avoidance.

Humans↗
PubMed1999

Reinforcement enhances hippocampal acetylcholine release in rats: an in vivo microdialysis study.

Rats were trained to press a lever under a 'Multiple FI-60s and Extinction' schedule with food reinforcements. After learning the task, an in vivo microdialysis probe was inserted into the dentate gyrus and CA3 regions of the hippocampus, and the sequential changes in the dialysate acetylcholine (ACh) concentration were analyzed. In the session, two fixed-interval of reinforcement (FI) components (for 20 min) and two extinction (EXT) components (for 30 min) were alternated to examine the correlation between behavioral and neurochemical outcomes. The dialysate ACh concentration increased during the FI component and returned to the baseline during the EXT component of the schedule. Next, in order to dissociate the effect of discrimination from the effect of rewarding on the neurochemical changes in the hippocampus, we used a final TEST period (for 20 min) during which the actual schedule was extinction but the discriminative stimulus was on, i.e. the manifest condition of the test period was reinforcement. In the TEST period, the animals pressed the lever with almost the same frequency as during the FI component; however, the dialysate ACh concentration did not increase above the baseline concentration. These results suggest that ACh release in the rat hippocampus is associated with reinforcement but not with discrimination in operant conditioning.

Acetylcholine↗