The least mean squares network with information coding: A model of cue learning.

Information coding, a method for coding variables in the least mean squares (LMSs) network, is proposed as an associative model of human cue learning. According to this model, participants in cue learning experiments assume the linear hypothesis and implement this hypothesis by updating their estimate of cue–target correlation on each trial. Because absence of a stimulus may give information about correlation, this model codes an absent stimulus as a nonzero value under some conditions. Thus, Information Coding differs from R. A. Rescorla and A. R. Wagner's (1972) model, a version of the LMSs network in which absence is always coded as 0. In 5 experiments, participants estimated the likelihood that a hypothetical patient had a disease given the patient's symptoms. The results support Information Coding, whereas the Rescorla-Wagner model, G. B. Chapman's (see record 1992-04135-001) variant of the Rescorla-Wagner model, and A. B. Markman's (see record 1990-09047-001) coding of the LMSs network all failed to account for at least one of these results. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Base-rate effects in category learning: A comparison of parallel network and memory storage-retrieval models.

Exemplar-memory and adaptive network models were compared in application to category learning data, with special attention to base rate effects on learning and transfer performance. Subjects classified symptom charts of hypothetical patients into disease categories, with informative feedback on learning trials and with the feedback either given or withheld on test trials that followed each fourth of the learning series. The network model proved notably accurate and uniformly superior to the exemplar model in accounting for the detailed course of learning; both the parallel, interactive aspect of the network model and its particular learning algorithm contribute to this superiority. During learning, subjects' performance reflected both category base rates and feature (symptom) probabilities in a nearly optimal manner, a result predicted by both models, though more accurately by the network model. However, under some test conditions, the data showed substantial base-rate neglect, in agreement with M. A. Gluck and G. H. Bower (see record 1989-00340-001). (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

LMS rules and the inverse base-rate effect: Comment on Gluck and Bower (1988).

M. A. Gluck and G. H. Bower (see record 1989-00340-001) suggested that through the use of the Rescorla-Wagner learning rule, a connectionist network might be able to model the inverse base-rate phenomenon found by D. L. Medin and S. M. Edelson (see record 1988-31640-001). I prove that a network of the type that they proposed does not capture this effect. However, one can also prove that with additional assumptions about the encoding of features, the Rescorla-Wagner learning rule can be made to model the inverse base-rate effect. The importance of these assumptions and an outline of how they might be tested are then discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Instrumental learning within the spinal cord: I. Behavioral properties

Four experiments are reported that explore whether spinal neurons can support instrumental learning. During training, one group of spinal rats (master) received legshock whenever one hindlimb was extended. Another group (yoked) received legshock independent of leg position. Master, but not yoked, rats learned to maintain their leg in a flexed position, exhibiting progressively longer flexions as a function of training (Experiment 1). All subjects were then tested by applying controllable shock to the same leg (Experiment 2). Master rats reacquired the instrumental response more rapidly (positive transfer), whereas yoked rats failed to learn (a learned helplessness-like effect). Disrupting response-outcome contiguity by delaying the onset and offset of shock by 100 ms eliminated learning (Experiment 3). Experiment 4 showed that shock onset contributes more to learning than does shock offset.     

Component and pattern information in adaptive networks.

M. A. Gluck and G. H. Bower (see record 1989-00340-001) presented an adaptive network model of human classification in which associative weights are modified according to R. A. Rescorla and A. R. Wagner (1972) conditioning theory, a special case of the Widrow-Hoff/LMS learning rule. Presenting empirical data from a series of artificial medical classification tasks, we argued that the network model predicts results that were unanticipated, given prevailing alternative theories of category learning. We consider here some alternative interpretations of this data suggested by D. R. Shanks (see record 1990-27514-001) and argue that they are not sufficiently compelling when compared to the network model's treatment of the data from all the experiments presented in our earlier study. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Connectionism and human learning: Critique of Gluck and Bower (1988).

M. A. Gluck and G. H. Bower (see record 1989-00340-001) argued that their data supported a connectionist model of human learning. Their Ss judged the probabilities of two diseases, given a target symptom; these probabilities were in fact equal. Because one disease was more frequent than the other, the target symptom was a better predictor of the less frequent disease than of the other. The Ss rated the two probabilities as unequal. In addition, Gluck and Bower argued that a simple connectionist model would be able to simulate the judgment bias. They assumed that Ss estimated the probabilities of the two diseases across all trials on which the target symptom occurred, either alone or with other symptoms, rather than the probability of each disease given only the target symptom. In fact, it is possible that the Ss were estimating the latter, in which case the bias in probability judgments was justified. Last, even if the Ss were estimating the former, the connectionist model cannot account for the bias seen in the probability judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Assessment of the Rescorla-Wagner model.

The Rescorla-Wagner model has been the most influential theory of associative learning to emerge from the study of animal behavior over the last 25 years. Recently, equivalence to this model has become a benchmark in assessing connectionist models, with such equivalence often achieved by incorporating the Widrow-Hoff delta rule. This article presents the Rescorla-Wagner model's basic assumptions, reviews some of the model's predictive successes and failures, relates the failures to the model's assumptions, and discusses the model's heuristic value. It is concluded that the model has had a positive influence on the study of simple associative learning by stimulating research and contributing to new model development. However, this benefit should neither lead to the model being regarded as inherently "correct" nor imply that its predictions can be profitably used to assess other models. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Stimulus preexposure, comparison, and changes in the associability of common stimulus features

In four experiments, rats received preexposure either to both of two compound flavours (AX and BX), or to just one (BX). Experiment 1 demonstrated a perceptual learning effect, showing that, for animals given preexposure to both flavours, an aversion conditioned to AX generalized only poorly to BX. Subsequent experiments assessed the properties of the common feature, X. Experiment 3 showed that the two preexposure treatments did not differ in the extent to which they produced habituation of the neophobia evoked by X. Experiment 2 showed that conditioning to X proceeded more rapidly in subjects given preexposure to both AX and BX than in subjects preexposed to BX alone. In Experiment 4, a similar effect was found when the elements of the compounds were presented serially. It is concluded that the perceptual learning effect of Experiment 1 occurs in spite of the fact that preexposure to two stimuli tends to maintain the associability of their common elements.     

面向匹配决策问题的漏整合神经元稀疏ESN网络

为了对匹配决策问题进行建模与预测,提出了一种具有更多神经生理学特征的稀疏回声状态网络(ESN),并基于在线监督学习方法对网络进行训练.为了评估网络的匹配决策性能,设计了三组测试数据集对网络性能进行测试,并提出了一种基于网络期望输出与实际输出序列最大相关系数的评价方法.仿真结果表明,新模型只需要较少的训练时间即可获得较好的决策性能,且对发放时间间隔、平移和网络噪声具有较好的鲁棒性.     

The effect of frequency of judgement and the type of trials on covariation learning.

This study showed that accuracy of the estimated relationship between a fictitious symptom and a disease depends on the interaction between the frequency of judgment and the last trial type. This effect appeared both in positive and zero contingencies (Experiment 1), and judgments were less accurate as frequency increased (Experiment 2). The effect can be explained neither by interference of previous judgments or memory demands (Experiment 3), nor by the perceptual characteristics of the stimuli (Experiments 4 and 5), and instructions intended to alter processing strategies do not produce any reliable effect. The interaction between frequency and trial type on covariation judgment is not predicted by any model (either statistical or associative) currently used to explain performance in covariation detection. The authors propose a belief-revision model to explain this effect as an important response mode variable on covariation learning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Cue competition in human categorization: Contingency or the Rescorla-Wagner Learning Rule? Comment on Shanks (1991).

D. R. Shanks (1991) reported experiments that show selective-learning effects in a categorization task, and presented simulations of his data using a connectionist network model implementing the Rescorla-Wagner (R-W) theory of animal conditioning. He concluded that his results (1) support the application of the R-W theory to account for human categorization, and (2) contradict a particular variant of contingency-based theories of categorization. These conclusions are examined. It is shown that the asymptotic weights produced by the R-W model actually predict systematic deviations from the observed human learning data. Shanks claimed that his simulations provided good qualitative fits to the observed data when the weights in the networks were allowed to reach their asymptote. However, analytic derivations of the asymptotic weights reveal that the final weights obtained in Shanks's Simulations 1 and 2 do not correspond to the actual asymptotic weights, apparently because the networks were not in fact run to asymptote. It is also shown that a contingency-based theory that incorporates the notion of focal sets can provide a more adequate explanation of cue competition than does the R-W model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Preferred hearing-aid frequency responses in simulated listening environments

This study was designed to determine if an adaptive strategy could be used to select frequency/gain characteristics that would be considered appropriate across a variety of listening environments. In Experiment I, the test-retest reliability of the paired comparison procedure used in Experiment II was assessed in quiet for nine subjects and in speech noise for six subjects. For both conditions, results revealed mean standard deviations of < 3 dB from 200 through 4000 Hz. In Experiment II, four subjects selected frequency/gain characteristics for five different listening environments (quiet, speech noise, quiet conference room, reverberant lecture hall, and reverberant lecture hall in noise). In general, subjects did not tend to select different frequency/gain characteristics across the five simulated environments used in this study. When differences in frequency responses were observed, they tended to be alterations in overall gain rather than changes in relative frequency response. Findings support additional evaluation in more diverse listening environments, possibly with systems that incorporate nonlinear signal processing.     

What is matched in direct matching? Intention attribution modulates motor priming.

Converging evidence has shown that action observation and execution are tightly linked. The observation of an action directly activates an equivalent internal motor representation in the observer (direct matching). However, whether direct matching is primarily driven by basic perceptual features of the observed movement or is influenced by more abstract interpretative processes is an open question. A series of behavioral experiments tested whether direct matching, as measured by motor priming, can be modulated by inferred action goals and attributed intentions. Experiment 1 tested whether observing an unsuccessful attempt to execute an action is sufficient to produce a motor-priming effect. Experiment 2 tested alternative perceptual explanations for the observed findings. Experiment 3 investigated whether the attribution of intention modulates motor priming by comparing motor-priming effects during observation of intended and unintended movements. Experiment 4 tested whether participants' interpretation of the movement as triggered by an external source or the actor's intention modulates the motor-priming effect by a pure instructional manipulation. Our findings support a model in which direct matching can be top-down modulated by the observer's interpretation of the observed movement as intended or not. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The generation effect and the modeling of associations in memory

The search of associative memory (SAM) model of Gillund and Shiffrin (1984) was applied to data of two experiments that examined the generation effect (Slamecka & Graf, 1978). Subjects studied a list of related word pairs, in which they either read both words in the pair or generated the right- hand response term using the left-hand stimulus term plus the response word fragment as generation cues. Experiment 1 manipulated encoding condition within subjects and used an incidental learning procedure. Experiment 2 manipulated encoding condition between subjects and used an intentional learning procedure. Memory was tested with recognition, cued recall, and free recall. A higher order association model gave a better and more parsimonious fit to the results than did an item-level association model. The relationship between various versions of SAM and current accounts of the generation effect are discussed, particulary the two-factor theory of Hirshman and Bjork (1988).     

Peer-model attributes and children's achievement behaviors.

In two experiments, we investigated how attributes of peer models influenced achievement behaviors among children who had experienced difficulties learning mathematical skills in school. In Experiment 1, children (M?=?10.6 years) observed either a same- or opposite-sex peer model demonstrating rapid (mastery model) or gradual (coping model) acquisition of fraction skills. Observing a coping model led to higher self-efficacy, skill, and training performance. In Experiment 2, children (M?=?10.9 years) observed either one or three same-sex peer models demonstrating mastery or coping behaviors while solving fractions. Children in the single-coping-model, multiple-coping-model, and multiple-mastery-model conditions demonstrated higher self-efficacy, skill, and training performance, compared with subjects who observed a single mastery model. In both studies, children who observed coping models judged themselves more similar in competence to the models than did subjects who observed mastery models. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Acquisition of automatic imitation is sensitive to sensorimotor contingency.

The associative sequence learning model proposes that the development of the mirror system depends on the same mechanisms of associative learning that mediate Pavlovian and instrumental conditioning. To test this model, two experiments used the reduction of automatic imitation through incompatible sensorimotor training to assess whether mirror system plasticity is sensitive to contingency (i.e., the extent to which activation of one representation predicts activation of another). In Experiment 1, residual automatic imitation was measured following incompatible training in which the action stimulus was a perfect predictor of the response (contingent) or not at all predictive of the response (noncontingent). A contingency effect was observed: There was less automatic imitation indicative of more learning in the contingent group. Experiment 2 replicated this contingency effect and showed that, as predicted by associative learning theory, it can be abolished by signaling trials in which the response occurs in the absence of an action stimulus. These findings support the view that mirror system development depends on associative learning and indicate that this learning is not purely Hebbian. If this is correct, associative learning theory could be used to explain, predict, and intervene in mirror system development. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Problem structure and the use of base-rate information from experience.

This article is concerned with the use of base-rate information that is derived from experience in classifying examples of a category. The basic task involved simulated medical decision making in which participants learned to diagnose hypothetical diseases on the basis of symptom information. Alternative diseases differed in their relative frequency or base rates of occurrence. In five experiments initial learning was followed by a series of transfer tests designed to index the use of base-rate information. On these tests, patterns of symptoms were presented that suggested more than one disease and were therefore ambiguous. The results reveal a consistent but complex pattern. Depending on the category structure and the nature of the ambiguous tests, participants use base-rate information appropriately, ignore base-rate information, or use base-rate information inappropriately (predict that the rare disease is more likely to be present). To our knowledge, no current categorization model predicts this pattern of results. To account for these results, a new model is described incorporating the ideas of property or symptom competition and context-sensitive retrieval. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Altered probabilistic learning and response biases in schizophrenia: Behavioral evidence and neurocomputational modeling.

Objective: Patients with schizophrenia (SZ) show reinforcement learning impairments related to both the gradual/procedural acquisition of reward contingencies, and the ability to use trial-to-trial feedback to make rapid behavioral adjustments. Method: We used neurocomputational modeling to develop plausible mechanistic hypotheses explaining reinforcement learning impairments in individuals with SZ. We tested the model with a novel Go/NoGo learning task in which subjects had to learn to respond or withhold responses when presented with different stimuli associated with different probabilities of gains or losses in points. We analyzed data from 34 patients and 23 matched controls, characterizing positive- and negative-feedback-driven learning in both a training phase and a test phase. Results: Consistent with simulations from a computational model of aberrant dopamine input to the basal ganglia patients, patients with SZ showed an overall increased rate of responding in the training phase, together with reduced response-time acceleration to frequently rewarded stimuli across training blocks, and a reduced relative preference for frequently rewarded training stimuli in the test phase. Patients did not differ from controls on measures of procedural negative-feedback-driven learning, although patients with SZ exhibited deficits in trial-to-trial adjustments to negative feedback, with these measures correlating with negative symptom severity. Conclusions: These findings support the hypothesis that patients with SZ have a deficit in procedural “Go” learning, linked to abnormalities in DA transmission at D1-type receptors, despite a “Go bias” (increased response rate), potentially related to excessive tonic dopamine. Deficits in trial-to-trial reinforcement learning were limited to a subset of patients with SZ with severe negative symptoms, putatively stemming from prefrontal cortical dysfunction. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning from text with diagrams: Promoting mental model development and inference generation.

Two experiments investigated learning outcomes and comprehension processes when students learned about the heart and circulatory system using (a) text only, (b) text with simplified diagrams designed to highlight important structural relations, or (c) text with more detailed diagrams reflecting a more accurate representation. Experiment 1 found that both types of diagrams supported mental model development, but simplified diagrams best supported factual learning. Experiment 2 replicated learning effects from Experiment 1 and tested the influence of diagrams on novices' comprehension processes. Protocol analyses indicated that both types of diagrams supported inference generation and reduced comprehension errors, but simplified diagrams most strongly supported information integration during learning. Visual representations appear to be most effective when they are designed to support the cognitive processes necessary for deep comprehension. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning about environmental geometry: An associative model.

K. Cheng (1986) suggested that learning the geometry of enclosing surfaces takes place in a geometric module blind to other spatial information. Failures to find blocking or overshadowing of geometry learning by features near a goal seem consistent with this view. The authors present an operant model in which learning spatial features competes with geometry learning, as in the Rescorla-Wagner model. Relative total associative strength of cues at a location determines choice of that location and thus the frequencies of reward paired with each cue. The model shows how competitive learning of local features and geometry can appear to result in potentiation, blocking, or independence, depending on enclosure shape and kind of features. The model reproduces numerous findings from dry arenas and water mazes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)