SEARCH · PubMed Health

Results for “GENERALIZATION 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 685 records · Page 38Linked to original sources

PubMed1998

Parity still isn't a generalisation problem.

Clark & Thornton take issue with my claim that parity is not a generalisation problem, and that nothing can be inferred about back-propagation in particular, or learning in general, from failures of parity generalisation. They advance arguments to support their contention that generalisation is a relevant issue. In this continuing commentary, I examine generalisation more closely in order to refute these arguments. Different learning algorithms will have different patterns of failure: back-propagation has no special status in this respect. This is not to deny that a particular algorithm might fortuitously happen to produce the "intended" function in an (oxymoronic) parity-generalisation task.

Algorithms↗

PubMed2000

Artificial neural networks: fundamentals, computing, design, and application.

Artificial neural networks (ANNs) are relatively new computational tools that have found extensive utilization in solving many complex real-world problems. The attractiveness of ANNs comes from their remarkable information processing characteristics pertinent mainly to nonlinearity, high parallelism, fault and noise tolerance, and learning and generalization capabilities. This paper aims to familiarize the reader with ANN-based computing (neurocomputing) and to serve as a useful companion practical guide and toolkit for the ANNs modeler along the course of ANN project development. The history of the evolution of neurocomputing and its relation to the field of neurobiology is briefly discussed. ANNs are compared to both expert systems and statistical regression and their advantages and limitations are outlined. A bird's eye review of the various types of ANNs and the related learning rules is presented, with special emphasis on backpropagation (BP) ANNs theory and design. A generalized methodology for developing successful ANNs projects from conceptualization, to design, to implementation, is described. The most common problems that BPANNs developers face during training are summarized in conjunction with possible causes and remedies. Finally, as a practical application, BPANNs were used to model the microbial growth curves of S. flexneri. The developed model was reasonably accurate in simulating both training and test time-dependent growth curves as affected by temperature and pH.

Humans↗

PubMed1991

Asymmetric Boltzmann machines.

We study asymmetric stochastic networks from two points of view: combinatorial optimization and learning algorithms based on relative entropy minimization. We show that there are non trivial classes of asymmetric networks which admit a Lyapunov function L under deterministic parallel evolution and prove that the stochastic augmentation of such networks amounts to a stochastic search for global minima of L. The problem of minimizing L for a totally antisymmetric parallel network is shown to be associated to an NP-complete decision problem. The study of entropic learning for general asymmetric networks, performed in the non equilibrium, time dependent formalism, leads to a Hebbian rule based on time averages over the past history of the system. The general algorithm for asymmetric networks is tested on a feed-forward architecture.

Algorithms↗

PubMed1993

Stimulus configuration, spatial learning, and hippocampal function.

Schmajuk and DiCarlo (Psychol. Rev., 99 (1992) 268-305) introduced a neural network, which utilizes a biologically plausible backpropagation procedure, to describe configural paradigms in classical conditioning. The model correctly describes many experimental results under the assumption that aspiration lesions of the hippocampus eliminate (a) the competition between simple and configural stimuli to gain association with the unconditioned stimulus and (b) the adjustment of initially random configural stimuli. The present study extends the network to describe place learning. Under the assumption that ibotenic acid lesions of the hippocampus only impair the adjustment of initially random configural stimuli, the model correctly shows that ibotenic acid lesions might spare a configural discrimination but impair place learning. In general, the results are taken to support a hippocampal role in the modulation of stimulus configuration.

Animals↗

PubMed1998

Psychotherapy with the chronic pain patient using coping skills development: outcome study.

This study provided outcome data measuring variables of a psychoeducational group approach to self-management of 88 chronic pain patients and 25 controls. Most of the injuries were back injuries and work related. At the completion of the 16 group sessions, patients reported decreased depression, pain severity, interference, and affective distress and increased life control and general activity. At 1-year follow-up, there was continued increased return to work, lowered workers' compensation, fewer health care visits, and less prescribed pain medication than demographically similar controls. Results suggest that a group psychoeducational program involving learning of general coping skills primarily and pain coping skills secondarily produces an effective approach for the management of chronic pain.

Adaptation, Psychological↗

PubMed1987

Variables influencing stimulus overselectivity and "tunnel vision" in developmentally delayed children.

Three variables (diagnosis, location of cues, and MA of learners) influencing stimulus control and stimulus overselectivity in autistic children were assessed. Eight autistic and 8 intellectually average children, matched on MA, were trained on two discrimination tasks; one task contained two "within-stimulus" (i.e., physically connected) cues; the other contained the same two cues presented "extra-stimulus" (i.e., physically separate). Generalization gradients were used following training to measure the degree of stimulus control acquired by each cue. Results showed: autistic subjects tended to respond overselectively only in the extra-stimulus condition; MA was positively correlated with breadth of learning; and when autistic children were overselective to one cue, some stimulus control was also acquired by the second cue. The notion of tunnel vision was discussed, as it may represent a "keystone" deficit interfering with stimulus control and learning by autistic children.

Adolescent↗

PubMed1995

Learning a new visuomotor transformation: error correction and generalization.

The use of an aiming tool requires learning a new transformation between visual and proprioceptive information and motor command. We have examined this question by quantifying the kinematics of the movement during the transitory phase of adaptation to a rotational bias (60 degrees counterclockwise, then clockwise) added to a standard mouse-cursor device in the plane of the screen. Control-aiming movements were almost linear with a bell-shaped velocity profile. The bias induced an equivalent initial directional error which was usually corrected within 20 trials. The learning trajectories were combinations of spirals and fast or slow straight movements. The posture of the hand was slightly (less than 10 degrees) modified by the bias. These features suggest three corrective processes: on-line continuous correction based on evaluation of the relative cursor-to-target position, discrete correction based on assessment of the discrepancy angle between the cursor-to-target direction and the effective cursor direction, and memorization of trial-to-trial correction. These results are interpreted in the light of neurophysiological data and neural net modeling, which suggest that the visuomotor transformation performed by cortical areas for reaching is effected by projecting the visual information on a reference frame that rotates with the arm. The initial directional error reappeared when the direction of the target was changed and increased with degree of change. The limited generalization suggests that bias correction is stored in relation to the coding of the target direction and that movement towards a new direction is computed as a projection of the previously learned bias on the new visual direction.

Feedback↗

PubMed1968

Some effects of discriminative training with equated frequency of reinforcement.

Pigeons were exposed to a multiple schedule which provided equally frequent reinforcement in the presence of two stimuli but which produced markedly different rates of key-pecking. Generalization gradients were displaced away from the stimulus associated with the lower rate of key-pecking. Another group of pigeons had similar training, except that a low rate of key-pecking was established in a stimulus with a much higher frequency of food reinforcement. In this case, the generalization gradients were not affected by the training on the schedule producing a low response rate.

Animals↗

PubMed1998

The learning curve in laparoscopic inguinal hernia repair for the community general surgeon.

OBJECTIVE: To determine the learning curve (number of operations required) to stabilize operating times and complication rates for a general surgeon doing laparoscopic inguinal hernia repair in a community practice. DESIGN: A prospective analysis. SETTING: A 256-bed secondary-care community hospital. PATIENTS: Ninety-eight consecutive patients booked for elective laparoscopic hernia repair on an outpatient basis. INTERVENTIONS: Using the transbdominal preperitoneal approach, 100 operations were carried out to repair 138 groins and a total of 164 separate hernial defects. OUTCOME MEASURES: The number of operations required to decrease operative times and complication rates to a steady level. RESULTS: There were no deaths. There were 5 conversions and 10 admissions, all occurring between the 1st and 46th operations. Two reoperations for reasons other than recurrence were required between the 45th and 55th operations. There were 24 other complications. Complications and surgical times began to level off after 50 operations. The 1 readmission was after the 42nd operation. There were 4 recurrences (2.9% recurrence rate), 2 in each group of 50 operations. Both groups of 2 recurrence occurred within the first 10 operations involving the use of a new stapler. Twenty-two other patients had open hernia repairs because laparoscopy was unsuitable for them. CONCLUSION: The learning curve for laparoscopic inguinal hernia repair in the hands of a general surgeon in community practice who is experienced in open herniorraphy and laparoscopic cholecystectomy is at least 50 operations.

Adult↗

PubMed1996

Transition in an era of reform.

Transition services for students with learning disabilities have not always been a priority in public schools. Although school personnel have long acknowledged the transition needs of students considered to have more severe disabilities, such as mental retardation, physical problems, and emotional disturbance, they do not always consider that students with learning disabilities also need transition programming. The result has been a neglect of these students' transition needs, which has likely contributed to some of the problems experienced by young adults with learning disabilities. This article focuses on the effect of educational reform on transition services for students with learning disabilities. General education reform--fueled by some of the critical reports of the 1980s--inclusion, and some of the more recent efforts to reform public education (namely, Goals 2000) are all discussed as they relate to the transition needs of students with learning disabilities.

Education, Special↗

PubMed2006

The development and evaluation of the use of a virtual learning environment (Blackboard 5) to support the learning of pre-qualifying nursing students undertaking a human anatomy and physiology module.

Students commence nurse education with varying levels of understanding of human anatomy and physiology due to a wide range of previous exposure to the topic. All students, however, are required to attain a broad knowledge of this topic prior to qualification. This paper describes the use of a Virtual Learning Environment (VLE), Blackboard 5, and the associated development of appropriate resources aimed at supporting nursing students undertaking a human anatomy and physiology module at Higher Education Level 1. The VLE was used as part of a blended learning approach. The results suggested that the majority of students utilised the VLE throughout the academic year. Opportunities for independent and self-directed learning were available in that students chose when and where to learn. Students generally commented favourably on ease of use and type of resources available. Frequency of use of the VLE, however, did not correlate strongly with the final examination mark achieved. Overall the VLE and the associated available resources appeared useful in supporting student learning and has been adopted for use in subsequent years.

Anatomy↗

PubMed1997

Gene targeting and the biology of learning and memory.

The general goal of genetic studies of learning and memory is to develop and test theories that explain the animal's behavior in neuroanatomical, neurophysiological, cellular, and molecular terms. In this review we describe the role that gene targeting and other transgenic techniques have had in the study of mammalian learning and memory. We focus especially on the hippocampus, a brain structure that is thought to be central to the processing and temporary storage of complex information. We also discuss the main issues that confront this young field, as well as our vision for its future.

Animals↗

PubMed1987

Rolipram as a discriminative stimuli: transfer to phosphodiesterase inhibitors.

Rats were trained to discriminate (+/-)-rolipram (0.32 mg/kg, i.p.) from saline in a two-lever food-reinforced drug discrimination procedure (FR 10). (+/-)-Rolipram served as a drug discriminative stimulus, and its discrimination was readily established with a mean of 42 training sessions to achieve criterion performance (at least 80% correct response in the consecutive generalization tests with both saline and (+/-)-rolipram). Thereafter, this stimulus was stably maintained. The (-)-isomer of rolipram was generalized at about one-third of the training dose of the (+/-)-compound, but the (+)-isomer was generalized only at 10 times the dose of the (+/-)-compound. This finding suggests that the (-)-compound is more extensively involved as a stimulus than either (+/-)-rolipram or the (+)-isomer. Ro20-1724, caffeine and theophylline, which are phosphodiesterase inhibitors, were generalized to (+/-)-rolipram. This result strongly suggests that phosphodiesterase inhibition may be an important factor involved in (+/-)-rolipram discrimination.

Animals↗

PubMed1998

Increased extracellular dopamine in the nucleus accumbens of the rat during associative learning of neutral stimuli.

Brain microdialysis was used to study changes in dopamine in the nucleus accumbens and the dorsal striatum during associative learning between two neutral stimuli, flashing light and tone, presented on a paired schedule during stage 1 of a sensory preconditioning paradigm. The tone was subsequently paired with mild footshock using standard aversive conditioning procedures and the formation of a conditioned association between the flashing light and the tone in stage 1 was assessed by measuring the ability of the flashing light to elicit the same conditioned response as the tone when presented at test. The first experiment used behavioural monitoring only, to establish stimulus parameters for subsequent microdialysis experiments. Animals receiving paired presentation of the light and tone in stage 1 showed a conditioned suppression of licking to the light as well as to the tone, indicating that associative learning between the flashing light and the tone had occurred during stage 1, whilst in a separate group of animals given the same stimuli over the same time period but on an explicitly non-paired schedule, the conditioned emotional response was seen to the tone, but not to the light, showing that no association had been formed between the two stimuli during stage 1. In dialysis experiments using the same procedure, we measured a two-fold rise in dopamine in the nucleus accumbens during paired presentation of flashing light and tone, but not during non-paired presentation of the two stimuli. On subsequent test presentation of the two stimuli, we saw increases in accumbal dopamine on presentation of the tone in both groups, reflecting the formation of an association with the footshock in both. However the flashing light elicited an increase in dopamine only in the group which had received paired presentation at stage 1. Thus accumbal dopamine release at test is correlated to the ability of the stimulus to evoke a conditioned response measured behaviourally. Hypotheses of the behavioural function of the mesolimbic dopamine system centre on its role in mediating the effects of biological reinforcers, both rewarding and aversive, conditioned and unconditioned. The present results, showing increases in extracellular dopamine in the nucleus accumbens when an association is formed between two stimuli of which neither is a biological reinforcer nor, prior to formation of the association, affects dopamine levels, suggest a role for accumbal dopamine in the modulation of associative learning in general, not only that involving reinforcement.

3,4-Dihydroxyphenylacetic Acid↗

PubMed1987

Extinction reveals stimulus control: latent learning of feature-negative discriminations in pigeons.

Various discriminations based on the presence versus absence of a single feature are supposedly learned much better when the feature appears on reinforced rather than nonreinforced trials. However, failures to show discriminative acquisition with the feature on negative trials could reflect a deficiency in control of performance rather than a lack of learning. Five experiments supported this alternative possibility. Pigeons that had yielded little or no evidence of learning (with distinguishing features like a small white square on the response key or a tone located some distance away) revealed clear differences between keypecking to the formerly positive and negative stimuli when all food was removed from the situation. Besides extinction, several other procedures for decreasing the positive predictiveness of the most informative stimulus element also unmasked feature-negative learning, whereas general and specific contextual changes did not. Incompletely mastered feature-positive discriminations improved during extinction, too. The findings of better discrimination performance in extinction were related to analogous effects in previous generalization and discrimination research employing other tasks and arrangements. A sign-tracking analysis could not completely account for the present results.

Animals↗

PubMedDate not supplied

Mathematics instruction for secondary students with learning disabilities.

Secondary students with learning disabilities generally make inadequate progress in mathematics. Their achievement is often limited by a variety of factors, including prior low achievement, low expectations for success, and inadequate instruction. This article will discuss techniques that have been demonstrated to be effective with secondary students who have learning disabilities in mathematics.

Achievement↗

PubMed1990

How to build a connectionist idiot (savant).

This paper describes the development of a connectionist model of an idiot savant date calculator using a multi-layer back-propagation network. The model mimics the behaviour of idiot savant date calculators in learning to perform the date calculation task without the need either to develop arithmetical skills or to encode explicit knowledge about regularities in the structure of dates. However, the performance of the model highlights limitations on the use of back-propagation as a general-purpose learning algorithm in complex problem domains. Even with a relatively simple problem like date calculation back-propagation can only succeed when forced to take a highly structured and modular approach to learning.

Algorithms↗

PubMed1975

Effects of stimulus similarity in discrimination training upon wavelength generalization in pigeons.

Thirty pigeons were given variable interval training to peck a 555-nm. light and then were tested for wavelength generalization. The subjects were later assigned to 1 of 3 groups, matched for both relative generalization slope and response rate. One group then received successive discrimination training between the 555-nm. stimulus (S+) and a vertical white line on a 555-nm. background (S minus); another group experienced the same S+ but a vertical white line on a black background as S minus. A third group received a comparable amount of single stimulus training with the 555-nm. value. On a second wavelength generalization test, the first group yielded greater sharpening of generalization than the second group, whereas the third group showed no change from Test 1. These results indicate that the sharpening of generalization gradients by discrimination training is directly related to the similarity of the discrimination training stimuli.

Animals↗