Large-scale Learning Research Articles

Measuring Equity-Promoting Behaviors in Digital Teaching Simulations: A Topic Modeling Approach

Diversity, equity, and inclusion (DEI) issues are urgent in education. We developed and evaluated a massive open online course ( N = 963) with embedded equity simulations that attempted to equip educators with equity teaching practices. Applying a structural topic model (STM)—a type of natural language processing (NLP)—we examined how participants with different equity attitudes responded in simulations. Over a sequence of four simulations, the simulation behavior of participants with less equitable beliefs converged to be more similar with the simulated behavior of participants with more equitable beliefs ( ES [effect size] = 1.08 SD). This finding was corroborated by overall changes in equity mindsets ( ES = 0.88 SD) and changed in self-reported equity-promoting practices ( ES = 0.32 SD). Digital simulations when combined with NLP offer a compelling approach to both teaching about DEI topics and formatively assessing learner behavior in large-scale learning environments.

A Rasch Model and Rating System for Continuous Responses Collected in Large-Scale Learning Systems.

An extension to a rating system for tracking the evolution of parameters over time using continuous variables is introduced. The proposed rating system assumes a distribution for the continuous responses, which is agnostic to the origin of the continuous scores and thus can be used for applications as varied as continuous scores obtained from language testing to scores derived from accuracy and response time from elementary arithmetic learning systems. Large-scale, high-stakes, online, anywhere anytime learning and testing inherently comes with a number of unique problems that require new psychometric solutions. These include (1) the cold start problem, (2) problem of change, and (3) the problem of personalization and adaptation. We outline how our proposed method addresses each of these problems. Three simulations are carried out to demonstrate the utility of the proposed rating system.

ЛІТЕРАТУРНЕ НАВЧАННЯ В "ЗОНІ НАЙБЛИЖЧОГО РОЗВИТКУ"

This article deals with the problem of organising students’ primary learning activities during Ukrainian literature lessons. By adopting L. Vygotsky’s zone of development concept, the author focuses on an important stage in students’ education during which they master specific kinds of work with literary material under the assistance of their teacher. The author bases his discussion on the methodologically established ways of analysing fictional and historical texts and highlights the operational mechanisms present within each kind of analysis that need to be mastered during secondary school. This helps him set minimal educational goals (micro-objectives) which need to be achieved in order to develop the whole system of students’ literature skills. This helps bring the problem of proximal development in line with a number of other problems which arise during literature study – those pertaining to the systemic andconsistent nature of the learning process, its goal-oriented character, the precision in determining didactic objectives, as well as its grounding in psychology. According to the author, planning a literature course needs to start with outlining the most large-scale learning goals, followed by elaborating them in as much detail as possible. The same approach, the author argues, should be used when developing the algorithm of the earning process, which should be viewed as a system of logical, consistent steps. Finally, the author of the article details the behavioural strategy of the teacher and of the students during these activities, emphasising the role of teacher’s instructions and demonstrations as well as the ability of students to follow and imitate them. The authorstresses that the success of learning largely depends on how effective students’ activities in the zone of development are – such effectiveness will allow students to develop primary knowledge and skills as well as to make them part of actual development, i.e. turn them into permanent skills. In this light, the author examines specific learning situations and the organisation of lessons and their learning stages.

RTop-k: A Statistical Estimation Approach to Distributed SGD

The large communication cost for exchanging gradients between different nodes significantly limits the scalability of distributed training for large-scale learning models. Motivated by this observation, there has been significant recent interest in techniques that reduce the communication cost of distributed Stochastic Gradient Descent (SGD), with gradient sparsification techniques such as top-k and random-k shown to be particularly effective. The same observation has also motivated a separate line of work in distributed statistical estimation theory focusing on the impact of communication constraints on the estimation efficiency of different statistical models. The primary goal of this paper is to connect these two research lines and demonstrate how statistical estimation models and their analysis can lead to new insights in the design of communication-efficient training techniques. We propose a simple statistical estimation model for the stochastic gradients which captures the sparsity and skewness of their distribution. The statistically optimal communication scheme arising from the analysis of this model leads to a new sparsification technique for SGD, which concatenates random-k and top-k, considered separately in the prior literature. We show through extensive experiments on both image and language domains with CIFAR-10, ImageNet, and Penn Treebank datasets that the concatenated application of these two sparsification methods consistently and significantly outperforms either method applied alone.

Peer Led Team Learning in a Foundational IPE Curriculum

BACKGROUND The Peer Led Team Learning (PLTL) instructional model utilizes Peer Leaders, advanced students who mentor and guide student teams to collaborate on applied course concepts. PURPOSE To apply a modified PLTL model in the university’s foundational, longitudinal, competency-based interprofessional education (IPE) curriculum. METHODS Twelve Peer Leaders were selected, trained, and deployed as facilitators for interprofessional teams of students during the IPE curriculum’s first three large-scale learning events. Peer Leaders completed an evaluation of training, a facilitation skills survey, and participated in a semi-structured focus group interview process. RESULTS After participating in the PLTL program, Peer Leaders reported increased confidence in their interprofessional knowledge and facilitation skills. The primary challenge for Peer Leaders in facilitating teams was lack of student engagement (n=7, 58%). CONCLUSION PLTL is a feasible model for IPE settings. It has the potential to both increase facilitator capacity in interprofessional learning activities and have a positive impact on Peer Leaders.

Multiview Active Learning for Scene Classification with High-Level Semantic-Based Hypothesis Generation

Multiview active learning (MVAL) is a technique which can result in a large decrease in the size of the version space than traditional active learning and has great potential applications in large-scale data analysis. This paper made research on MVAL-based scene classification for helping the computer accurately understand diverse and complex environments macroscopically, which has been widely used in many fields such as image retrieval and autonomous driving. The main contribution of this paper is that different high-level image semantics are used for replacing the traditional low-level features to generate more independent and diverse hypotheses in MVAL. First, our algorithm uses different object detectors to achieve local object responses in the scenes. Furthermore, we design a cascaded online LDA model for mining the theme semantic of an image. The experimental results demonstrate that our proposed theme modeling strategy fits the large-scale data learning, and our MVAL algorithm with both high-level semantic views can achieve significant improvement in the scene classification than traditional active learning-based algorithms.

Variance Counterbalancing for Stochastic Large-scale Learning

Stochastic Gradient Descent (SGD) is perhaps the most frequently used method for large scale training. A common example is training a neural network over a large data set, which amounts to minimizing the corresponding mean squared error (MSE). Since the convergence of SGD is rather slow, acceleration techniques based on the notion of “Mini-Batches” have been developed. All of them however, mimicking SGD, impose diminishing step-sizes as a means to inhibit large variations in the MSE objective. In this article, we introduce random sets of mini-batches instead of individual mini-batches. We employ an objective function that minimizes the average MSE and its variance over these sets, eliminating so the need for the systematic step size reduction. This approach permits the use of state-of-the-art optimization methods, far more efficient than the gradient descent, and yields a significant performance enhancement.

Educational Assessment of the Post‐Pandemic Age: Chinese Experiences and Trends Based on Large‐Scale Online Learning

Owing to the break‐out of the COVID‐19 pandemic, students have to take more online learning than offline, and large‐scale education assessment programs have to be suspended or postponed. How could education assessment adapt to large‐scale online learning? How could the effect and safety of online assessment be improved? What role should formative assessment play in student admissions? How could different assessment results be linked? Reflections on and trends of the Chinese experiences are presented in this article. Based on cross‐cultural comparison research, measures to be recommended are as follows: reviewing previous theories, improving existing methods continuously, and developing assessment techniques innovatively according to new application scenarios.

Abstract IA04: Clinical challenges in oncologic imaging: AI support from image analysis to integrated diagnostics

Abstract The automated analysis of large-scale clinical datasets poses new compelling challenges for data-driven and model-based computational methods in medical imaging. As a matter of fact, the amount of heterogeneous biomedical data is considerably increasing due to the advances in medical imaging acquisition modalities and high-throughput technologies for multi-omics. In addition, electronic health records can be properly integrated to support personalized screening and diagnosis. In such a context, artificial intelligence (AI) is revolutionizing cancer image analysis by relying on sophisticated machine learning and computational intelligence techniques. Therefore, cutting-edge AI methods can enable the shift from organization-centric to patient-centric models, leading to effective multi-institutional health care services in terms of both clinical outcomes and costs. Computerized oncologic image analysis is encouraging the transition from largely qualitative image interpretation to quantitative assessment through automated methods aiming at early detection as well as lesion characterization. Nevertheless, several challenges and opportunities exist: • Reproducible and reliable segmentation methods are required to cope with time-consuming and error-prone manual delineation in laborious human decision-making tasks. As a matter of fact, accurate operator-independent segmentation procedures can improve the robustness of radiomics analyses. • Accurate computer-assisted diagnosis, associated with proper data curation, can reduce the risk of overlooking the diagnosis in a clinical environment. • Prognostic/predictive biomarker discovery by capturing the underlying tumor heterogeneity. • Quantification and monitoring of intra-/intertumoral heterogeneity during the course of the disease, by a proper integration of imaging, clinical, and molecular data on a patient-by-patient level • From a health-economics perspective, increasing costs of highly specific oncologic treatments (including immunotherapies) require accurate patient selection strategies, as well as better biomarkers of treatment response. Currently, many state-of-the-art methods based on deep learning have been achieving outstanding performance. The key to their success is the strong learning ability of fully supervised ML models and the availability of large-scale labeled datasets that include precise annotations. Unfortunately, in biomedical research, collecting such accurate annotations is an expensive process due to the need for domain experts’ knowledge. The generation of large mineable imaging datasets might mitigate this challenge by overcoming data paucity and heterogeneity issues. However, along with the availability of samples, data quality and diversity should be considered by collecting and preparing harmonized datasets. Therefore, the generalization abilities in multi-institutional studies can be improved by exploiting transfer learning and domain adaptation techniques. Finally, the explainability and safety of clinical decision support systems should be guaranteed, by avoiding “black-box” AI models exhibiting unpredictable behaviors. In conclusion, precision oncology should be directed towards the development of integrated radiogenomics paradigms that provide robust computational tools for investigating cancer biology as well as its implications for predicting treatment response. This solution allows for large-scale data collection (from multiple institutions) and continuous learning, by dealing with cyber-security and privacy issues. At present, the ultimate challenge is bridging the gap between AI and clinical practice, by firstly performing well-validated clinical research studies. This step is vital for the final translation and deployment of these advanced AI approaches in precision oncology. Citation Format: Evis Sala. Clinical challenges in oncologic imaging: AI support from image analysis to integrated diagnostics [abstract]. In: Proceedings of the AACR Special Conference on Advancing Precision Medicine Drug Development: Incorporation of Real-World Data and Other Novel Strategies; Jan 9-12, 2020; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(12_Suppl_1):Abstract nr IA04.

Large-Scale Data Learning Method for Anomaly Detection using Machine Learning for Monitoring Vibration in Vehicle Equipment

Large-Scale Data Learning Method for Anomaly Detection using Machine Learning for Monitoring Vibration in Vehicle Equipment

Augmenting the effect of professional development on effective instruction through professional communities

ABSTRACT This study addresses gaps in the research about the effect of professional development (PD) quality on effective instruction by taking a broad perspective when defining professional learning of teachers and by providing a theory-based explanation for effective instruction. Using a large-scale data set, Teaching and Learning International Survey (TALIS) 2013, the interrelationships between the quality of PD experience, professional communities at the school, and effective instruction were investigated. This theory testing study showed that the effect of PD events on effective instruction was stronger with the spillover effect of professional communities than with its sole effect.

Qsparse-Local-SGD: Distributed SGD With Quantization, Sparsification, and Local Computations

Communication bottleneck has been identified as a significant issue in distributed optimization of large-scale learning models. Recently, several approaches to mitigate this problem have been proposed, including different forms of gradient compression or computing local models and mixing them iteratively. In this paper, we propose Qsparse-local-SGD algorithm, which combines aggressive sparsification with quantization and local computation along with error compensation, by keeping track of the difference between the true and compressed gradients. We propose both synchronous and asynchronous implementations of Qsparse-local-SGD . We analyze convergence for Qsparse-local-SGD in the distributed setting for smooth non-convex and convex objective functions. We demonstrate that Qsparse-local-SGD converges at the same rate as vanilla distributed SGD for many important classes of sparsifiers and quantizers. We use Qsparse-local-SGD to train ResNet-50 on ImageNet and show that it results in significant savings over the state-of-the-art, in the number of bits transmitted to reach target accuracy.

Accelerating incremental gradient optimization with curvature information

This paper studies an acceleration technique for incremental aggregated gradient ({\sf IAG}) method through the use of \emph{curvature} information for solving strongly convex finite sum optimization problems. These optimization problems of interest arise in large-scale learning applications. Our technique utilizes a curvature-aided gradient tracking step to produce accurate gradient estimates incrementally using Hessian information. We propose and analyze two methods utilizing the new technique, the curvature-aided IAG ({\sf CIAG}) method and the accelerated CIAG ({\sf A-CIAG}) method, which are analogous to gradient method and Nesterov's accelerated gradient method, respectively. Setting $\kappa$ to be the condition number of the objective function, we prove the $R$ linear convergence rates of $1 - \frac{4c_0 \kappa}{(\kappa+1)^2}$ for the {\sf CIAG} method, and $1 - \sqrt{\frac{c_1}{2\kappa}}$ for the {\sf A-CIAG} method, where $c_0,c_1 \leq 1$ are constants inversely proportional to the distance between the initial point and the optimal solution. When the initial iterate is close to the optimal solution, the $R$ linear convergence rates match with the gradient and accelerated gradient method, albeit {\sf CIAG} and {\sf A-CIAG} operate in an incremental setting with strictly lower computation complexity. Numerical experiments confirm our findings. The source codes used for this paper can be found on \url{http://github.com/hoitowai/ciag/}.

Video Based Person Re Identification Methods, Datasets, and Deep Learning

Video Analytics applications like security and surveillance face a critical problem of person re-identification abbreviated as re-ID. The last decade witnessed the emergence of large-scale datasets and deep learning methods to use these huge data volumes. Most current re-ID methods are classified into either image-based or video-based re-ID. Matching persons across multiple camera views have attracted lots of recent research attention. Feature representation and metric learning are major issues for person re-identification. The focus of re-ID work is now shifting towards developing end-to-end re-Id and tracking systems for practical use with dynamic datasets. Most previous works contributed to the significant progress of person re-identification on still images using image retrieval models. This survey considers the more informative and challenging video-based person re-ID problem, pedestrian re-ID in particular. Publicly available datasets and codes are listed as a part of this work. Current trends which include open re-identification systems, use of discriminative features and deep learning is marching towards new applications in security and surveillance, typically for tracking.

Recognition of Large-Scale ncRNA Data Using a Novel Multitask Cross-Learning 0-Order TSK Fuzzy Classifier

Recognizing noncoding ribonucleic acid (ncRNA) data is helpful in realizing the regulation of tumor formation and certain aspects of life mechanisms, such as growth, differentiation, development, and immunity. However, the scale of ncRNA data is usually very large. Using machine learning (ML) methods to automatically analyze these data can obtain more precise results than manually analyzing these data, but the traditional ML algorithms can process only small-scale training data. To solve this problem, a novel multitask cross-learning 0-order Takagi–Sugeno–Kang fuzzy classifier (MT-CL-0-TSK-FC) is proposed that uses a multitask cross-learning mechanism to solve the large-scale learning problem of ncRNA data. In addition, the proposed MT-CL-0-TSK-FC method naturally inherits the interpretability of traditional fuzzy systems and eventually generates an interpretable rulesbased database to recognize the ncRNA data. The experimental results indicate that the proposed MT-CL-0TSK-FC method has a faster running time and better classification accuracy than traditional ML methods.

Stochastic trust region inexact Newton method for large-scale machine learning

Nowadays stochastic approximation methods are one of the major research direction to deal with the large-scale machine learning problems. From stochastic first order methods, now the focus is shifting to stochastic second order methods due to their faster convergence and availability of computing resources. In this paper, we have proposed a novel Stochastic Trust RegiOn Inexact Newton method, called as STRON, to solve large-scale learning problems which uses conjugate gradient (CG) to inexactly solve trust region subproblem. The method uses progressive subsampling in the calculation of gradient and Hessian values to take the advantage of both, stochastic and full-batch regimes. We have extended STRON using existing variance reduction techniques to deal with the noisy gradients and using preconditioned conjugate gradient (PCG) as subproblem solver, and empirically proved that they do not work as expected, for the large-scale learning problems. Finally, our empirical results prove efficacy of the proposed method against existing methods with bench marked datasets.

Tracking Performance of Online Stochastic Learners

The utilization of online stochastic algorithms is popular in large-scale learning settings due to their ability to compute updates on the fly, without the need to store and process data in large batches. When a constant step-size is used, these algorithms also have the ability to adapt to drifts in problem parameters, such as data or model properties, and track the optimal solution with reasonable accuracy. Building on analogies with the study of adaptive filters, we establish a link between steady-state performance derived under stationarity assumptions and the tracking performance of online learners under random walk models. The link allows us to infer the tracking performance from steady-state expressions directly and almost by inspection.

Problematika Internalisasi Pendidikan Karakter di Indonesia

The purpose of this study is to analyze the problems of internalizing character education in Indonesia using several methods, including meta-analysis. the way to formulate the problem of this research begins with conducting observations, then tracing the results of other relevant research to be analyzed. The research data consisted of six journal articles and four theses that had been tried. The process of data analysis is done by descriptive qualitative method. The results of the analysis show that there is a problem in internalizing character education in Indonesia because the current problem of internalizing character education is (a) students' lack of interest in practicing the nation's character and culture (b) The internalisation system of character values ​​in schools is not continued by the environment outside the school (c) Crisis exemplary ( d) Foreign culture that is not in accordance with the culture and character of the Indonesian nation is entered into a large scale (e) Learning programs that are less relevant to the times. The expected solution to the internalization of character education is (a) conceptual approach (b) systematic / habituation approach (c) modeling / exemplary approach. The educational model for character development in adolescents is integrated in school regulations, learning and extracurricular activities, namely scout extracurricular activities. Based on the results of the analysis it can be concluded that an effective character education model is built from a school climate that is conducive to the development of positive character both through learning models at school.

Deep Reinforcement Learning for User Association and Resource Allocation in Heterogeneous Cellular Networks

Heterogeneous cellular networks can offload the mobile traffic and reduce the deployment costs, which have been considered to be a promising technique in the next-generation wireless network. Due to the non-convex and combinatorial characteristics, it is challenging to obtain an optimal strategy for the joint user association and resource allocation issue. In this paper, a reinforcement learning (RL) approach is proposed to achieve the maximum long-term overall network utility while guaranteeing the quality of service requirements of user equipments (UEs) in the downlink of heterogeneous cellular networks. A distributed optimization method based on multi-agent RL is developed. Moreover, to solve the computationally expensive problem with the large action space, multi-agent deep RL method is proposed. Specifically, the state, action and reward function are defined for UEs, and dueling double deep Q-network (D3QN) strategy is introduced to obtain the nearly optimal policy. Through message passing, the distributed UEs can obtain the global state space with a small communication overhead. With the double-Q strategy and dueling architecture, D3QN can rapidly converge to a subgame perfect Nash equilibrium. Simulation results demonstrate that D3QN achieves the better performance than other RL approaches in solving large-scale learning problems.

Near Optimal Coded Data Shuffling for Distributed Learning

Data shuffling between distributed cluster of nodes is one of the critical steps in implementing large-scale learning algorithms. Randomly shuffling the data-set among a cluster of workers allows different nodes to obtain fresh data assignments at each learning epoch. This process has been shown to provide improvements in the learning process (via testing and training error). However, the statistical benefits of distributed data shuffling come at the cost of extra communication overhead from the master node to worker nodes, and can act as one of the major bottlenecks in the overall time for computation. There has been significant recent interest in devising approaches to minimize this communication overhead. One approach is to provision for extra storage at the computing nodes. The other emerging approach is to leverage coded communication to minimize the overall communication overhead. The focus of this work is to understand the fundamental tradeoff between the amount of storage and the communication overhead for distributed data shuffling. In this paper, we first present an information theoretic formulation for the data shuffling problem, accounting for the underlying problem parameters (number of workers, $K$ , number of data points, $N$ , and available storage, and $S$ per node). We then present an information theoretic lower bound on the communication overhead for data shuffling as a function of these parameters. We next present a novel coded communication scheme and show that the resulting communication overhead of the proposed scheme is within a multiplicative factor of at most $\frac {K}{K-1}$ from the lower bound (which is upper bounded by 2 for $K\geq 2$ ). Furthermore, we present new results towards closing this gap through a novel coded communication scheme, which we call the aligned coded shuffling. This scheme is inspired by the ideas of coded shuffling and interference alignment. In particular, we show that the aligned scheme achieves the optimal storage vs communication trade-off for $K , and further reduces the maximum multiplicative gap down to $\frac {K-\frac {1}{3}}{K-1}$ , for $K\geq 5$ .