Types Of Learning Problems Research Articles

Selective consolidation of learning and memory via recall-gated plasticity.

In a variety of species and behavioral contexts, learning and memory formation recruits two neural systems, with initial plasticity in one system being consolidated into the other over time. Moreover, consolidation is known to be selective; that is, some experiences are more likely to be consolidated into long-term memory than others. Here, we propose and analyze a model that captures common computational principles underlying such phenomena. The key component of this model is a mechanism by which a long-term learning and memory system prioritizes the storage of synaptic changes that are consistent with prior updates to the short-term system. This mechanism, which we refer to as recall-gated consolidation, has the effect of shielding long-term memory from spurious synaptic changes, enabling it to focus on reliable signals in the environment. We describe neural circuit implementations of this model for different types of learning problems, including supervised learning, reinforcement learning, and autoassociative memory storage. These implementations involve synaptic plasticity rules modulated by factors such as prediction accuracy, decision confidence, or familiarity. We then develop an analytical theory of the learning and memory performance of the model, in comparison to alternatives relying only on synapse-local consolidation mechanisms. We find that recall-gated consolidation provides significant advantages, substantially amplifying the signal-to-noise ratio with which memories can be stored in noisy environments. We show that recall-gated consolidation gives rise to a number of phenomena that are present in behavioral learning paradigms, including spaced learning effects, task-dependent rates of consolidation, and differing neural representations in short- and long-term pathways.

Pastoral care of children with specific learning disabilities

This thesis deals with the issue of pastoring children with specific learning disabilities. It introduces the manifestations of specific learning disabilities and their basic classification, taking into account other types of learning problems and their framework characteristics. The uniqueness of pastoral work with children with specific learning disabilities is presented and examples of pastoral work are given. Emphasis is placed on creating a supportive environment with emphasis on the proclamation of God’s Word. The Man was created in the image of God, endowed with God’s love, and saved to eternal life in Jesus Christ. This message (the gospel) is also for the Man of today. However, it is also necessary to point out that each of us has different gifts, different abilities and different possibilities to receive this unique message. Therefore, when proclaiming the love of God, it is necessary to approach each person individually and to adapt the proclamation to the possibilities of the recipient of faith. Particularly with children, there is a need for a strong individualization of the approach, since the child’s discovery of God and the child’s journey to faith are very fragile.

Invariance encoding in sliced-Wasserstein space for image classification with limited training data

Deep convolutional neural networks (CNNs) are broadly considered to be state-of-the-art generic end-to-end image classification systems. However, they are known to underperform when training data are limited and thus require data augmentation strategies that render the method computationally expensive and not always effective. Rather than using a data augmentation strategy to encode invariances as typically done in machine learning, here we propose to mathematically augment a nearest subspace classification model in sliced-Wasserstein space by exploiting certain mathematical properties of the Radon Cumulative Distribution Transform (R-CDT), a recently introduced image transform. We demonstrate that for a particular type of learning problem, our mathematical solution has advantages over data augmentation with deep CNNs in terms of classification accuracy and computational complexity, and is particularly effective under a limited training data setting. The method is simple, effective, computationally efficient, non-iterative, and requires no parameters to be tuned. Python code implementing our method is available at https://github.com/rohdelab/mathematical_augmentation. Our method is integrated as a part of the software package PyTransKit, which is available at https://github.com/rohdelab/PyTransKit.

Thermodynamic efficiency of learning a rule in neural networks

Biological systems have to build models from their sensory input data that allow them to efficiently process previously unseen inputs. Here, we study a neural network learning a binary classification rule for these inputs from examples provided by a teacher. We analyse the ability of the network to apply the rule to new inputs, that is to generalise from past experience. Using stochastic thermodynamics, we show that the thermodynamic costs of the learning process provide an upper bound on the amount of information that the network is able to learn from its teacher for both batch and online learning. This allows us to introduce a thermodynamic efficiency of learning. We analytically compute the dynamics and the efficiency of a noisy neural network performing online learning in the thermodynamic limit. In particular, we analyse three popular learning algorithms, namely Hebbian, Perceptron and AdaTron learning. Our work extends the methods of stochastic thermodynamics to a new type of learning problem and might form a suitable basis for investigating the thermodynamics of decision-making.

The relationship between Kolb's learning style and technological problem solving styles of middle school students

본 연구의 목적은 학습양식에 따라 중학생의 학습유형을 분류하고, 학습유형과 기술적 문제해결 성향과의 관계를 밝히는 데 있었다. 이를 위해 A중학교의 학생 344명을 대상으로 하여 학습양식과 기술적 문제해결 성향을 검사하여 분석하였다. SPSS 12.0 통계프로그램으로 일원분산분석(one-way ANOVA)과 Pearson 적률상관관계 분석을 실시하였다. 연구의 결과는 다음과 같다. 첫째, Kolb의 학습양식에 따라 학습유형을 분류한 결과 수렴형(19.77%), 확산형(22.09%), 동화형(30.23%), 조절형(27.91%)의 분포를 보였다. 이 결과를 통하여 학습에 참여하는 학습자들을 네 가지의 다른 학습유형으로 구분할 수 있다는 것을 확인할 수 있었다. 둘째, 학습유형 간 기술적 문제해결 성향 점수를 비교한 결과 각 유형 별로 유의한 차이가 발견되었다. 특히 동화형과 수렴형은 모든 요소에서 유의미한 차이가 있음이 검증되었다. 셋째, 학습유형 별 기술적 문제해결 성향의 점수는 동화형의 점수가 가장 높았으며 확산형, 조절형, 수렴형의 순으로 이어졌다. 이러한 순서는 각 하위요소에서 약간의 변동은 있으나 비슷한 경향성을 보였다. 동화형 학습자는 가장 높은 기술적 문제해결 성향을 가지고 있고 수렴형 학습자는 상대적으로 가장 낮은 기술적 문제해결 성향을 가지고 있다는 것을 알 수 있었다. 넷째, 구체적 경험을 통하여 학습할 때 수렴형, 확산형 학습자는 낮은 기술적 문제해결 성향을 보이고, 동화형 학습자는 높은 기술적 문제해결 성향을 보였다. 또한 활동적 실험을 통하여 학습할 때 수렴형, 확산형, 조절형 학습자는 높은 기술적 문제해결 성향을 보이고, 동화형 학습자는 낮은 기술적 문제해결 성향을 보였다. 추후 학습유형의 차이를 고려한 다양한 교수⋅학습 방법이 연구되어야 하며, 학습유형과 기술적 문제해결력 등 인지적 특성과의 관련에 관한 후속연구가 필요하다. 또한, 학습유형 이외에 학습자 개인차를 결정짓는 다양한 요인과 기술적 문제해결의 관계를 구명하려는 노력이 필요하다.The purposes of this study were to classify middle school students learning styles according to their learning styles and to clarify the relationship between learning styles and their styles to solve technological problems. For this purpose, 344 students of A junior high school were examined and analyzed for their learning style and technological problem solving styles. One-way analysis of variance (ANOVA) and Pearson s correlation analysis were performed with SPSS 12.0 statistical program. The results of the study are as follows; First, according to Kolb s learning style, there were Converging Type (19.77%), Diverging Type (22.09%), Assimilating Type (30.23%) and Accommodating Type (27.91%). The results show that learners participating in learning can be classified into four different types of learning. Second, there was a significant difference between the types of learning problems. Especially, Assimilating Type and Converging Type were proved to have significant difference in all factors. Third, the score of the tendency toward solving the technological problems according to the learning type was highest in the Assimilating Type, followed by the Diverging Type, the Accommodating Type and the Converging Type. Fourth, when learners learn through specific experiences, Converging and Diverging type learners showed low tendency to solve technological problems, and Assimilated Learners showed high tendency to solve technological problems. There is a need to study the teaching and learning methods considering differences in learning styles, and further research is needed on the relationship between cognitive characteristics and technological problem solving ability.

Conservation Abilities, Visuospatial Skills, and Numerosity Processing Speed: Association With Math Achievement and Math Difficulties in Elementary School Children

The aim of the present study was to investigate the associations between elementary school children’s mathematical achievement and their conservation abilities, visuospatial skills, and numerosity processing speed. We also assessed differences in these abilities between children with different types of learning problems. In Study 1 (N = 229), we investigated second to fourth graders and in Study 2 (N = 120), third and fourth graders. Analyses revealed significant contributions of numerosity processing speed and visuospatial skills to math achievement beyond IQ. Conservation abilities were predictive in Study 1 only. Children with math difficulties showed lower visuospatial skills and conservation abilities than children with typical achievement levels and children with reading and/or spelling difficulties, whereas children with combined difficulties explicitly showed low conservation abilities. These findings provide further evidence for the relations between children’s math skills and their visuospatial skills, conservation abilities, and processing speed and contribute to the understanding of deficits that are specific to mathematical difficulties.

Model-based transductive learning of the kernel matrix

This paper addresses the problem of transductive learning of the kernel matrix from a probabilistic perspective. We define the kernel matrix as a Wishart process prior and construct a hierarchical generative model for kernel matrix learning. Specifically, we consider the target kernel matrix as a random matrix following the Wishart distribution with a positive definite parameter matrix and a degree of freedom. This parameter matrix, in turn, has the inverted Wishart distribution (with a positive definite hyperparameter matrix) as its conjugate prior and the degree of freedom is equal to the dimensionality of the feature space induced by the target kernel. Resorting to a missing data problem, we devise an expectation-maximization (EM) algorithm to infer the missing data, parameter matrix and feature dimensionality in a maximum a posteriori (MAP) manner. Using different settings for the target kernel and hyperparameter matrices, our model can be applied to different types of learning problems. In particular, we consider its application in a semi-supervised learning setting and present two classification methods. Classification experiments are reported on some benchmark data sets with encouraging results. In addition, we also devise the EM algorithm for kernel matrix completion.

Cognitive processing and developmental changes in finnish school students: Comparison between normal and language impaired (dysphasic) students

Course of cognitive development was investigated in a sample of normal school students (N=163) in the middle part of Finland by giving the students tests of four cognitive functions that were categorized according to the PASS theory of intelligence (planning, attention, simultaneous and successive processing). Subsequently, a group of students in special education (N=60, language impaired school students) was also studied for possible deviations in cognitive development. The age of the students varied from 7 to 19 years. Three main goals were (1) to study developmental trends of cognitive functions across three different age groups of school students and in two samples, (2) to identify distinct cognitive subgroups and profiles among the students, and, (3) to compare the cognitive profiles of normal and language impaired (dysphasic) students. First, in a cross-sectional design, it was seen linear cognitive trends for all PASS tasks, however so, that for the language impaired group the cognitive development was not as salient as for the normal students. Especially successive processing did not show developmental trend in the sample of language impaired (dysphasic) students. Secondly, a cluster analysis procedure following a factor analysis demonstrated two very distinct cognitive subgroups among the normal and dysphasic students. The subgroups in both separate samples differed from each other in successive and simultaneous processes. Additionally, the two subgroups of the normal students differed from each other also in planning. Thirdly, compared to the normal students, the dysphasic highly functioning subgroup did worse than normals in nonverbal successive ordering, but did better in verbal successive task. Contrary to expectations, the deviation was not on the verbal but instead on a nonverbal task. Need for extending PASS research into other types of learning problems (such as arithmetics, severe learning disability) is articulated. Pragmatically, observed minor differences in cognitive structure between the two student samples refer to differential remediation and skill training.

Effects of Dynamic Testing on the Classification of Learning Disabilities: The Predictive and Discriminant Validity of the Swanson-Cognitive Processing Test (S-CPT)

This study assessed the criterion-related and discriminant validity of the Swanson-Cognitive Processing Test (S-CPT). The S-CPT is an individually administered standardized battery of information processing tasks that assess performance under initial (static) and dynamic testing (testing-the-limits) conditions. In Experiment 1, children's ( N = 61) performance on the S-CPT under initial and dynamic testing conditions was compared to the WISC-R in predicting performance on Reading and Mathematics subtests of the WRAT-R. Gain scores established during dynamic testing conditions were the best predictors of reading and mathematics performance (14% and 26% of the variance, respectively). In Experiment 2, 156 children with reading deficits and 35 skilled readers were compared on the S-CPT under static and dynamic testing conditions. As expected, skilled readers performed better than learning-disabled readers across all subtests, but the results were qualified when subgroups were analyzed. The cluster analysis indicated that approximately 25% of the sample with learning disabilities was misdiagnosed, and better reflected slow learners (poor readers) or instructionally responsive children. The results support the notion that dynamic assessment contributes significant variance in the prediction of academic achievement and that such assessment procedures discriminate the processing potential of children with reading disabilities from children with other types of learning problems.

Strategy instruction and attributional beliefs in West Germany and the United States: Do teachers foster metacognitive development?

Previous research has shown German children to be more strategic on sort-recall memory tasks than their American age-mates, and to show fewer effort-related attributions. We conducted this study to determine if those differences are due to systematic differences in the strategy instruction and attributional beliefs of German and U.S. teachers, and to explore metacognitive instructional practices in the two countries. Teachers responded to a questionnaire that inquired about their use of strategy instructions, fostering of reflective thinking in pupils, sources of children's learning problems, and modeling of metacognitive skills such as monitoring. The second part of the questionnaire asked about the reasons underlying children's academic successes and failures. German teachers reported more instruction of task-specific strategies, while American teachers showed more effort-related attributions. The types of strategies instructed and types of learning problems most frequently described varied across the two countries, and also according to how many years the teachers had taught. Results were discussed regarding their implications for metacognitive developmental theory, particularly regarding culture and other environmental influences on the development of controlled processing.

Learning disabilities and special education: a critique.

For those enmeshed in the day-by-day problems of the field, the trees often obscure the forest. Those familiar with the career of James Gallagher know that he has not only managed to keep the forest in view but has created a vision of how it should grow and evolve. Thus, he was a clear choice for bringing an insider's perspective to the difficult task of critiquing the papers in this series. In offering his views, he begins by noting the consensus among respondents to the future survey with regard to four key problems: (a) the unsatisfactory nature of the status quo, (b) the heterogeneity among those labeled as LD, (c) the need for greater emphasis on differential treatment, and (d) concern over political factors shaping professional decisions. He then explores the classification problem, drawing implications for resolving the conflict between etiology and treatment and for improving research and training. Recognizing that naming problems and identifying possible solutions is not enough, Gallagher concludes his discussion with a five-step plan outlining specific actions that could be taken immediately to influence current public policy related to the differentiation and treatment of the different types of learning problems presently diagnosed as LD. Both his analysis and the call for immediate action underscore the current state of the field's development and directions for future advancement.—Howard S. Adelman and Linda Taylor, Guest Editors

Memory for verbal and nonverbal stimuli in learning disability subgroups: Analysis by selective reminding

Memory for verbal and nonverbal stimuli was evaluated using selective reminding procedures in normal achieving children and four groups of disabled learners: (1) reading-spelling disabled (R-S); (2) reading-spelling-arithmetic disabled (R-S-A); (3) spelling-arithmetic disabled (S-A); and (4) arithmetic disabled (A). Each child received two analogous free-list memory tasks, one for verbal material (animal names) and the other for nonverbal material (random dot patterns). These tasks were administered using selective reminding procedures that permit separation of storage and retrieval aspects of memory by reminding children only of those words not recalled on previous trials. Results revealed that relative to controls, the A and S-A children had significantly lower storage and retrieval scores on the nonverbal task, but did not differ on the verbal task; the R-S children differed only on retrieval scores from the verbal task; and the R-S-A children on retrieval scores on the verbal task and storage and retrieval scores on the nonverbal task. Thus, results indicate that the memory performance of disabled learners varies according to (1) the type of learning problem (arithmetic vs reading), (2) the nature of the stimuli (verbal vs nonverbal), and (3) the aspect of memory being assessed (storage vs retrieval). This study provides external validation for the classification of disabled learners according to patterns of academic achievement, demonstrating a useful procedure for dealing with the intrasubject variability characteristic of disabled learners.

The Dimension of Inattention Among Children with School Problems

Within a population of children referred for school-related problems, youngsters with significant attention deficits were compared with children having other types of learning problems. Using data from parent and teacher questionnaires as well as results of multidisciplinary team assessments, each patient was assigned either to a group with significant attention deficits or to one with learning problems, but fewer, if any, problems with attention. Disagreement between observation sources was demonstrated. Children qualifying for the significant attention deficit group were characterized by a greater likelihood of having behavioral problems during the toddler and preschool years. They also had a higher prevalence of minor neurologic signs and difficulty on tests of language development. Their current behavioral and adjustment problems seemed to be more severe and widespread than those of the youngsters with whom they were compared. No major intergroup differences were observed with regard to age at referral, socioeconomic status, family history of learning and behavior problems, perinatal health risks, or psychosocial difficulties. There was considerable historic and symptomatic overlap between the two groups, suggesting that the clinical picture characteristic of significant attention deficits is relatively nonspecific and is either a primary or secondary finding in a large proportion of a heterogeneous population of children experiencing difficulties in school.

Learning processes in physician education: integrating psyche and soma.

There was a wide range of difference in the way fifty-five third-year medical students wrote about a hypothetical medical patient with abdominal complaints and mild schizophrenic traits. The written exercise served as a projective test. Twenty-five wrote from a singularly psychological orientation; eighteen mentioned possible somatic disease in terms of dichotomy, psychic vs. somatic; twelve wrote in terms that allowed for the possible coexistence of psychic and somatic diseases. In each of these groups there were a few who demonstrated a relatively extreme attitude. The latter tend to represent a type of learning problem involving compartmentalized thinking. The three such students who dichotomized also demonstrated a tendency for strong ambivalence, vigorous likes and dislikes. One of the students who made a strong dichotomy between psyche and soma is the subject of a more intensive study of his learning processes. Hypotheses are suggested to explain how the physician learns to regard the “whole man” in his practice of medicine. An integration of psyche and soma must pervade the medical curriculum to enhance professional development. The development of a well-integrated image of the whole man is an emotional as well as an intellectual learning task, especially if “the whole man” orientation is to take deep root in the physician and in his everyday practice of medicine.

Learning problems in psychoanalytic supervision: diagnosis and management

An attempt has been made to present a general conceptual framework within which to consider the various types of learning problems that may occur in the psychoanalytic supervisory situation. The classification offered is somewhat artificial since many of the learning problems encountered in students derive from more than one type of difficulty. A general scheme of management has also been suggested, but probably no two supervisors would be in complete agreement regarding all of the details of the management of the educational needs of the student. But the opportunities for individuality and creativity in the individual teacher are what make education the lively and challenging experience that it can be.