Exemplar-based Learning Research Articles

Why do we need to employ exemplars in moral education? Insights from recent advances in research on artificial intelligence

ABSTRACT In this paper, I examine why moral exemplars are useful and even necessary in moral education despite several critiques. To support my point, I review recent AI research demonstrating that exemplar-based learning is superior to rule-based learning in model performance in training neural networks, such as large language models. I particularly focus on why education aiming at promoting the development of multifaceted moral functioning can be done effectively by using exemplars, which is like exemplar-based learning in AI model training. Furthermore, I discuss the potential limitations and issues related to exemplar-applied moral education with findings from recent AI research raising concerns about model biases and toxic outcomes. I propose ways to address the concerns regarding employing moral exemplars as well. As remedies, I suggest that autonomy-supporting deliberative and reflective learning should be utilized. Furthermore, based on the discussion, I examine how macroscopic socio-cultural aspects influence the effectiveness of exemplar-applied education.

Exemplar learners and rule learners: Stable tendencies or malleable preferences?

When learning new concepts, students tend to use either exemplar-based learning strategies (e.g., memorizing specific examples) or rule-based learning strategies (e.g., abstracting general rules). Prior research suggests that participants' strategy choices during learning depend on individuals' preexisting learning tendencies, with some people being exemplar learners and others rule learners. Yet, strategy choices are also influenced by how the study materials are taught (rule-focused or exemplar-focused). The present study examined how these two factors interact using an alphanumeric symbol addition task. We examined whether exemplar learners would switch to using rule-based strategies when given rule-focused training and if rule learners would fail to learn the rule when given exemplar-focused training. We found that both rule and exemplar learners used a rule-based strategy after a rule-focused training and neither group learned the rule after an exemplar-focused training. Our results suggest that individuals can be shaped to adopt either rule-based or exemplar-based strategies during learning, regardless of their inherent learning tendencies. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Exemplar MobileNetV2-Based Artificial Intelligence for Robust and Accurate Diagnosis of Multiple Sclerosis.

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system that prominently affects young adults due to its debilitating nature. The pathogenesis of the disease is focused on the inflammation and neurodegeneration processes. Inflammation is associated with relapses, while neurodegeneration emerges in the progressive stages of the disease. Magnetic resonance imaging (MRI) is commonly used for the diagnosis of MS, and guidelines such as the McDonald criteria are available. MRI is an essential tool to demonstrate the spatial distribution and changes over time in the disease. This study discusses the use of image processing techniques for the diagnosis of MS and specifically combines the MobileNetV2 network with exemplar-based learning, IMrMr feature selection, and K-Nearest Neighbors (KNN) classification methods. Experiments conducted on two different datasets (Dataset 1 and Dataset 2) demonstrate that these methods provide high accuracy in diagnosing MS. Dataset 1 comprises 128 patients with 706 MRI images, 131 MS patients with 667 MRI images, and 150 healthy control subjects with 1373 MRI images. Dataset 2 includes an MS group with 650 MRI images and a healthy control group with 676 MRI images. The results of the study include 10-fold cross-validation results performed on different image sections (Axial, Sagittal, and Hybrid) for Dataset 1. Accuracy rates of 99.76% for Axial, 99.48% for Sagittal, and 98.02% for Hybrid sections were achieved. Furthermore, 100% accuracy was achieved on Dataset 2. In conclusion, this study demonstrates the effective use of powerful image processing methods such as the MobileNetV2 network and exemplar-based learning for the diagnosis of MS. These findings suggest that these methods can be further developed in future research and offer significant potential for clinical applications in the diagnosis and monitoring of MS.

Camera-independent color constancy by scene semantics

Current learning-based color constancy methods are typically employed to find camera-specific illuminant mappings. Consequently, these methods exhibit poor generalization to images captured by varying cameras. In this paper, we present a camera-independent learning method based on Scene Semantics, and we call it CISS. Inspired by the camera-independent property of gray-based methods, CISS does not directly estimate camera-specific illuminant by training model as most learning methods do. Instead, the model's output is transformed into camera-independent scene statistics related to gray-based assumptions to avoid being affected by camera variations. Based on these estimated scene statistics, illuminant can be calculated indirectly. To estimate scene statistics accurately, CISS designs illuminant-invariance scene semantics features as input to the model. Then, the model estimates scene statistics for each input image in terms of scene semantics with exemplar-based learning. Experiments show that, on several public datasets, CISS is able to outperform present methods for multi-cameras color constancy, and is flexible enough to be well generalized to the unseen camera without fine-tuning by additional images.

RUBubbles as a novel tool to study categorization learning

Grouping objects into discrete categories affects how we perceive the world and represents a crucial element of cognition. Categorization is a widespread phenomenon that has been thoroughly studied. However, investigating categorization learning poses several requirements on the stimulus set in order to control which stimulus feature is used and to prevent mere stimulus–response associations or rote learning. Previous studies have used a wide variety of both naturalistic and artificial categories, the latter having several advantages such as better control and more direct manipulation of stimulus features. We developed a novel stimulus type to study categorization learning, which allows a high degree of customization at low computational costs and can thus be used to generate large stimulus sets very quickly. ‘RUBubbles’ are designed as visual artificial category stimuli that consist of an arbitrary number of colored spheres arranged in 3D space. They are generated using custom MATLAB code in which several stimulus parameters can be adjusted and controlled separately, such as number of spheres, position in 3D-space, sphere size, and color. Various algorithms for RUBubble generation can be combined with distinct behavioral training protocols to investigate different characteristics and strategies of categorization learning, such as prototype- vs. exemplar-based learning, different abstraction levels, or the categorization of a sensory continuum and category exceptions. All necessary MATLAB code is freely available as open-source code and can be customized or expanded depending on individual needs. RUBubble stimuli can be controlled purely programmatically or via a graphical user interface without MATLAB license or programming experience.Graphical abstract

Exemplar-based learning probably requires learning abstractions: A commentary on Ambridge (2020)

Exemplar-based learning requires (i) a segmentation procedure for identifying the units of past experiences that a present experience can be compared to, and (ii) a similarity function for comparing these past experiences to the present experience. This article argues that for a learner to learn a language these two mechanisms will require abstractions such as linguistic features. Moreover, because the segmentation procedure will vary depending on the language, a radical exemplar theory is likely to require linguistic abstractions for learning.

Individual and age differences in block-by-block dynamics of category learning strategies.

The present experiment examined individual and age differences in the dynamics of category learning strategies. Participants learned categories determined by a disjunctive rule with relational features through a feedback training procedure. During training, participants responded to strategy probes following each block to provide online assessment of the extent to which rule- and exemplar-based strategies were used throughout the training period. We introduced this measure as an alternative to model-based approaches to assessing individual differences in strategy use during training. Following training, participants classified ambiguous transfer objects that were assumed to distinguish between earlier use of rule- and exemplar-based learning strategies. We included this measure to obtain a relatively objective index of strategy use during training. Next, participants provided global ratings of their use of rule- and exemplar-based strategies during training. Results showed that strategy preferences expressed on the final training block predicted categorisation of ambiguous transfer objects better than global strategy reports. In addition, we utilised the block-by-block strategy reports to investigate the dynamics of learners' strategy preferences over the course of training. The findings revealed greater fluidity in strategy preferences for both younger and older adults than has been previously documented in the category learning literature. The novel block-by-block strategy reports in conjunction with the transfer-based approach allowed for a more nuanced examination of individual and age differences in strategy use and categorisation performance.

Unsupervised Exemplar-Based Learning for Improved Document Image Classification

Many recent state-of-the-art approaches for document image classification are based on supervised feature learning that requires a large amount of labeled training data. In real-world problem of document image classification, the available amount of labeled data is limited and scarce while a large amount of unlabeled data is often available at almost no cost. In this paper, we present an approach for learning visual features for document analysis in an unsupervised way, which improves the document image classification performance without increasing the amount of annotated data. The proposed approach trains a neural network model on an auxiliary task in which every training example is associated with a different label (exemplar) and expanded to multiple images through a data augmentation technique. Thus, the learned model, which is trained in an unsupervised way, is used to boost the document classification performance. In fact, this learned model has proved to be consistently efficient in two different settings: i) as an unsupervised feature extractor to represent document images for an unsupervised classification task (i.e., clustering); and ii) in the parameters initialization of a supervised classification task trained with a small amount of annotated data. We perform experiments on the Tobacco-3482 dataset and demonstrate the capability of our approach to improve i) the unsupervised classification accuracy up to 2.4%; and ii) the supervised classification accuracy by 1.5% without any extra data or by 5% when using 3000 additional not annotated samples.

Stress Elevates Frontal Midline Theta in Feedback-based Category Learning of Exceptions.

Adapting behavior based on category knowledge is a fundamental cognitive function, which can be achieved via different learning strategies relying on different systems in the brain. Whereas the learning of typical category members has been linked to implicit, prototype abstraction learning, which relies predominantly on prefrontal areas, the learning of exceptions is associated with explicit, exemplar-based learning, which has been linked to the hippocampus. Stress is known to foster implicit learning strategies at the expense of explicit learning. Procedural, prefrontal learning and cognitive control processes are reflected in frontal midline theta (4-8 Hz) oscillations during feedback processing. In the current study, we examined the effect of acute stress on feedback-based category learning of typical category members and exceptions and the oscillatory correlates of feedback processing in the EEG. A computational modeling procedure was applied to estimate the use of abstraction and exemplar strategies during category learning. We tested healthy, male participants who underwent either the socially evaluated cold pressor test or a nonstressful control procedure before they learned to categorize typical members and exceptions based on feedback. The groups did not differ significantly in their categorization accuracy or use of categorization strategies. In the EEG, however, stressed participants revealed elevated theta power specifically during the learning of exceptions, whereas the theta power during the learning of typical members did not differ between the groups. Elevated frontal theta power may reflect an increased involvement of medial prefrontal areas in the learning of exceptions under stress.

Hybrid non-parametric particle swarm optimization and its stability analysis

As a population-based random search optimization technique, particle swarm optimization (PSO) has become an important branch of swarm intelligence (SI). The tuning of parameters in PSO has attracted the attention of many researchers. This study proposes an alternative technology called hybrid non-parametric PSO (HNPPSO) algorithm. Other SI operations, including a multi-crossover operation, a vertical crossover, and an exemplar-based learning strategy, are combined with the proposed algorithm to balance the global and local search capabilities. The first- and second-order stability analyses conducted for the present study showed that the particle positions are expected to converge at a fixed point in the search space and that the variance of the particle positions converges to zero. In the experiments, the proposed algorithm was compared with 10 other advanced PSO techniques using 40 widely used benchmark functions. The experimental results indicated that the proposed algorithm yields better solution accuracy and convergence speed than the other PSO techniques. The proposed algorithm significantly outperformed the other PSO approaches in terms of convergence speed.

Exemplar-Based Learning Classifier System with Dynamic Matching Range for Imbalanced Data

In this paper, we propose a method to improve ECS-DMR which enables appropriate output for imbalanced data sets. In order to control generalization of LCS in imbalanced data set, we propose a method of applying imbalance ratio of data set to a sigmoid function, and then, appropriately update the matching range. In comparison with our previous work (ECS-DMR), the proposed method can control the generalization of the appropriate matching range automatically to extract the exemplars that cover the given problem space, wchich consists of imbalanced data set. From the experimental results, it is suggested that the proposed method provides stable performance to imbalanced data set. The effect of the proposed method using the sigmoid function considering the data balance is shown.

Learning representative exemplars using one-class Gaussian process regression

An exemplar is an observation that represents a group of similar observations. Exemplars from data are examined to divide entire heterogeneous data into several homogeneous subgroups, wherein each subgroup is represented by an exemplar. With its inherent sparsity, an exemplar-based learning model provides a parsimonious model to represent or cluster large-scale data. A novel exemplar learning method using one-class Gaussian process (GP) regression is proposed in this study. The proposed method constructs data distribution support from one-class GP regression using automatic relevance determination prior and heterogeneous GP noise. Exemplars that correspond to the basis vectors of the constructed support function are then automatically located during the training process. The proposed method is applied to various data sets to examine its operability, characteristics of data representation, and cluster analysis. The exemplars of some real data generated by the proposed method are also reported.

Children's Form-Meaning Connections to Verb Phrases and Exemplar-Based Learning in Japanese Elementary School

Children's Form-Meaning Connections to Verb Phrases and Exemplar-Based Learning in Japanese Elementary School

Text-Guided Attention Model for Image Captioning

Visual attention plays an important role to understand images and demonstrates its effectiveness in generating natural language descriptions of images. On the other hand, recent studies show that language associated with an image can steer visual attention in the scene during our cognitive process. Inspired by this, we introduce a text-guided attention model for image captioning, which learns to drive visual attention using associated captions. For this model, we propose an exemplar-based learning approach that retrieves from training data associated captions with each image, and use them to learn attention on visual features. Our attention model enables to describe a detailed state of scenes by distinguishing small or confusable objects effectively. We validate our model on MS-COCO Captioning benchmark and achieve the state-of-the-art performance in standard metrics.

Theories or fragments?

Lake et al. argue persuasively that modelling human-like intelligence requires flexible, compositional representations in order to embody world knowledge. But human knowledge is too sparse and self-contradictory to be embedded in "intuitive theories." We argue, instead, that knowledge is grounded in exemplar-based learning and generalization, combined with high flexible generalization, a viewpoint compatible both with non-parametric Bayesian modelling and with sub-symbolic methods such as neural networks.

Out of sight, out of mind: Categorization learning and normal aging

The present combined EEG and eye tracking study examined the process of categorization learning at different age ranges and aimed to investigate to which degree categorization learning is mediated by visual attention and perceptual strategies. Seventeen young subjects and ten elderly subjects had to perform a visual categorization task with two abstract categories. Each category consisted of prototypical stimuli and an exception. The categorization of prototypical stimuli was learned very early during the experiment, while the learning of exceptions was delayed. The categorization of exceptions was accompanied by higher P150, P250 and P300 amplitudes. In contrast to younger subjects, elderly subjects had problems in the categorization of exceptions, but showed an intact categorization performance for prototypical stimuli. Moreover, elderly subjects showed higher fixation rates for important stimulus features and higher P150 amplitudes, which were positively correlated with the categorization performances. These results indicate that elderly subjects compensate for cognitive decline through enhanced perceptual and attentional processing of individual stimulus features. Additionally, a computational approach has been applied and showed a transition away from purely abstraction-based learning to an exemplar-based learning in the middle block for both groups. However, the calculated models provide a better fit for younger subjects than for elderly subjects. The current study demonstrates that human categorization learning is based on early abstraction-based processing followed by an exemplar-memorization stage. This strategy combination facilitates the learning of real world categories with a nuanced category structure. In addition, the present study suggests that categorization learning is affected by normal aging and modulated by perceptual processing and visual attention.

Exemplar-Based Color Constancy and Multiple Illumination.

Exemplar-based learning or, equally, nearest neighbor methods have recently gained interest from researchers in a variety of computer science domains because of the prevalence of large amounts of accessible data and storage capacity. In computer vision, these types of technique have been successful in several problems such as scene recognition, shape matching, image parsing, character recognition, and object detection. Applying the concept of exemplar-based learning to the problem of color constancy seems odd at first glance since, in the first place, similar nearest neighbor images are not usually affected by precisely similar illuminants and, in the second place, gathering a dataset consisting of all possible real-world images, including indoor and outdoor scenes and for all possible illuminant colors and intensities, is indeed impossible. In this paper, we instead focus on surfaces in the image and address the color constancy problem by unsupervised learning of an appropriate model for each training surface in training images. We find nearest neighbor models for each surface in a test image and estimate its illumination based on comparing the statistics of pixels belonging to nearest neighbor surfaces and the target surface. The final illumination estimation results from combining these estimated illuminants over surfaces to generate a unique estimate. We show that it performs very well, for standard datasets, compared to current color constancy algorithms, including when learning based on one image dataset is applied to tests from a different dataset. The proposed method has the advantage of overcoming multi-illuminant situations, which is not possible for most current methods since they assume the color of the illuminant is constant all over the image. We show a technique to overcome the multiple illuminant situation using the proposed method and test our technique on images with two distinct sources of illumination using a multiple-illuminant color constancy dataset. The concept proposed here is a completely new approach to the color constancy problem and provides a simple learning-based framework.

Computing phonological generalization over real speech exemplars

Though it has attracted growing attention from phonologists and phoneticians, Exemplar Theory (e.g. Bybee, 2001) has hitherto lacked an explicit production model that can apply to speech signals. An adequate model must be able to generalize; but this presents the problem of how to generate an output that generalizes over a collection of unique, variable-length signals. Rather than resorting to a priori phonological units such as phones, we adopt a dynamic programming approach, using an optimization criterion that is sensitive to the frequency of similar subsequences within other exemplars: the Phonological Exemplar-Based Learning System. We show that PEBLS displays pattern-entrenchment behaviour, central to Exemplar Theory's account of phonologization.

Lexical frequency and exemplar-based learning effects in language acquisition: evidence from sentential complements

Usage-based approaches to language acquisition argue that children acquire the grammar of their target language using general-cognitive learning principles. The current paper reports on an experiment that tested a central assumption of the usage-based approach: argument structure patterns are connected to high frequency verbs that facilitate acquisition. Sixty children ( N = 60) aged 4- and 6-years participated in a sentence recall/lexical priming experiment that manipulated the frequency with which the target verbs occurred in the finite sentential complement construction in English. The results showed that the children performed better on sentences that contained high frequency verbs. Furthermore, the children’s performance suggested that their knowledge of finite sentential complements relies most heavily on one particular verb – think, supporting arguments made by Goldberg [Goldberg, A.E., 2006. Constructions at Work: The Nature of Generalization in Language. Oxford University Press, Oxford], who argued that skewed input facilitates language learning.

An MRF-Based DeInterlacing Algorithm With Exemplar-Based Refinement

In this paper, we propose an MRF-based deinterlacing algorithm that combines the benefits of rule-based algorithms such as motion-adaptation, edge-directed interpolation, and motion compensation, with those of an MRF formulation. MRF-based interpolation and enhancement algorithms are typically formulated as an optimization over pixel intensities or colors, which can make them relatively slow. In comparison, our MRF-based deinterlacing algorithm uses interpolation functions as labels.We use seven interpolants (three spatial, three temporal, and one for motion compensation). The core dynamic programming algorithm is, therefore, sped up greatly over the direct use of intensity as labels. We also show how an exemplar-based learning algorithm can be used to refine the output of our MRF-based algorithm. The training set can be augmented with exemplars from static regions of the same video, as a form of "self-learning."