Promoting computational thinking through project-based learning

This paper introduces project-based learning (PBL) features for developing technological, curricular, and pedagogical supports to engage students in computational thinking (CT) through modeling. CT is recognized as the collection of approaches that involve people in computational problem solving. CT supports students in deconstructing and reformulating a phenomenon such that it can be resolved using an information-processing agent (human or machine) to reach a scientifically appropriate explanation of a phenomenon. PBL allows students to learn by doing, to apply ideas, figure out how phenomena occur and solve challenging, compelling and complex problems. In doing so, students take part in authentic science practices similar to those of professionals in science or engineering, such as computational thinking. This paper includes 1) CT and its associated aspects, 2) The foundation of PBL, 3) PBL design features to support CT through modeling, and 4) a curriculum example and associated student models to illustrate how particular design features can be used for developing high school physical science materials, such as an evaporative cooling unit to promote the teaching and learning of CT.

KDExplainer: A Task-oriented Attention Model for Explaining Knowledge Distillation

Knowledge distillation (KD) has recently emerged as an efficacious scheme for learning compact deep neural networks (DNNs). Despite the promising results achieved, the rationale that interprets the behavior of KD has yet remained largely understudied. In this paper, we introduce a novel task-oriented attention model, termed as KDExplainer, to shed light on the working mechanism underlying the vanilla KD. At the heart of KDExplainer is a Hierarchical Mixture of Experts (HME), in which a multi-class classification is reformulated as a multi-task binary one. Through distilling knowledge from a free-form pre-trained DNN to KDExplainer, we observe that KD implicitly modulates the knowledge conflicts between different subtasks, and in reality has much more to offer than label smoothing. Based on such findings, we further introduce a portable tool, dubbed as virtual attention module (VAM), that can be seamlessly integrated with various DNNs to enhance their performance under KD. Experimental results demonstrate that with a negligible additional cost, student models equipped with VAM consistently outperform their non-VAM counterparts across different benchmarks. Furthermore, when combined with other KD methods, VAM remains competent in promoting results, even though it is only motivated by vanilla KD. The code is available at https:// github.com/zju-vipa/KDExplainer.

Effective training of convolutional neural networks for age estimation based on knowledge distillation

Age estimation from face images can be profitably employed in several applications, ranging from digital signage to social robotics, from business intelligence to access control. Only in recent years, the advent of deep learning allowed for the design of extremely accurate methods based on convolutional neural networks (CNNs) that achieve a remarkable performance in various face analysis tasks. However, these networks are not always applicable in real scenarios, due to both time and resource constraints that the most accurate approaches often do not meet. Moreover, in case of age estimation, there is the lack of a large and reliably annotated dataset for training deep neural networks. Within this context, we propose in this paper an effective training procedure of CNNs for age estimation based on knowledge distillation, able to allow smaller and simpler “student” models to be trained to match the predictions of a larger “teacher” model. We experimentally show that such student models are able to almost reach the performance of the teacher, obtaining high accuracy over the LFW+, LAP 2016 and Adience datasets, but being up to 15 times faster. Furthermore, we evaluate the performance of the student models in the presence of image corruptions, and we demonstrate that some of them are even more resilient to these corruptions than the teacher model.

Sampling redesign of soil penetration resistance in spatial t-Student models

Aim of study: To reduce the sample size in an agricultural area of 167.35 hectares, cultivated with soybean, to analyze the spatial dependence of soil penetration resistance (SPR) with outliers. Area of study: Cascavel, Brazil Material and methods: The reduction of sample size was made by the univariate effective sample size ( ) methodology, assuming that the t-Student model represents the probability distribution of SPR. Main results: The radius and the intensity of spatial dependence have an inverse relationship with the estimated value of the . For the depths of SPR with spatial dependence, the highest estimated value of the reduced the sample size by 40%. From the new sample size, the sampling redesign was performed. The accuracy indexes showed differences between the thematic maps with the original and reduced sampling designs. However, the lowest values of the standard error in the parameters of the spatial dependence structure evidenced that the new sampling design was appropriate. Besides, models of semivariance function were efficiently estimated, which allowed identifying the existence of spatial dependence in all depth of SPR.Research highlights: The sample size was reduced by 40%, allowing for lesser financial investments with data collection and laboratory analysis of soil samples in the next mappings in the agricultural area. The spatial t-Student model was able to reduce the influence of outliers in the spatial dependence structure.

Machine Learning-Based Student Modeling Methodology for Intelligent Tutoring Systems

Machine learning-based modeling technology has recently become a powerful technique and tool for developing models for explaining, predicting, and describing system/human behaviors. In developing intelligent education systems or technologies, some research has focused on applying unique machine learning algorithms to build the ad-hoc student models for specific educational systems. However, systematically developing the data-driven student models from the educational data collected over prior educational experiences remains a challenge. We proposed a systematic and comprehensive machine learning-based modeling methodology to develop high-performance predictive student models from the historical educational data to address this issue. This methodology addresses the fundamental modeling issues, from data processing, to modeling, to model deployment. The said methodology can help developing student models for intelligent educational systems. After a detailed description of the proposed machine learning-based methodology, we introduce its application to an intelligent navigation tutoring system. Using the historical data collected in intelligent navigation tutoring systems, we conduct large-scale experiments to build the student models for training systems. The preliminary results proved that the proposed methodology is useful and feasible in developing the high-performance models for various intelligent education systems.

Cross-layer knowledge distillation with KL divergence and offline ensemble for compressing deep neural network

Deep neural networks (DNN) have solved many tasks, including image classification, object detection, and semantic segmentation. However, when there are huge parameters and high level of computation associated with a DNN model, it becomes difficult to deploy on mobile devices. To address this difficulty, we propose an efficient compression method that can be split into three parts. First, we propose a cross-layer matrix to extract more features from the teacher's model. Second, we adopt Kullback Leibler (KL) Divergence in an offline environment to make the student model find a wider robust minimum. Finally, we propose the offline ensemble pre-trained teachers to teach a student model. To address dimension mismatch between teacher and student models, we adopt a $1\times 1$ convolution and two-stage knowledge distillation to release this constraint. We conducted experiments with VGG and ResNet models, using the CIFAR-100 dataset. With VGG-11 as the teacher's model and VGG-6 as the student's model, experimental results showed that the Top-1 accuracy increased by 3.57% with a $2.08\times$ compression rate and 3.5x computation rate. With ResNet-32 as the teacher's model and ResNet-8 as the student's model, experimental results showed that Top-1 accuracy increased by 4.38% with a $6.11\times$ compression rate and $5.27\times$ computation rate. In addition, we conducted experiments using the ImageNet $64\times 64$ dataset. With MobileNet-16 as the teacher's model and MobileNet-9 as the student's model, experimental results showed that the Top-1 accuracy increased by 3.98% with a $1.59\times$ compression rate and $2.05\times$ computation rate.

Role of Machine Learning in Modern Education and Teaching

Computers are being utilized in field in education for many years. In last few decades, research within the field of artificial intelligence (AI) is positively affecting educational application. Advanced machine learning and deep learning techniques could be used for extracting knowledgeable information from crude information. In this chapter, the authors have analysed the impact of artificial intelligence in the education domain. The authors will discuss how with the development of machine learning techniques in last few decades, machine learning models can anticipate student performance. By learning about every student, models can identify the shortcomings. Then the authors will propose different approaches to improve student performance. Teachers can also use this model to understand student perception levels in a better way so that they can modulate their lectures according to student perception levels.