Showing papers on "Active learning (machine learning) published in 2022"

Model-Based Machine Learning

Abstract: Several decades of research in the field of machine learning have resulted in a multitude of different algorithms for solving a broad range of problems. To tackle a new application, a researcher typically tries to map their problem onto one of these existing methods, often influenced by their familiarity with specific algorithms and by the availability of corresponding software implementations. In this study, we describe an alternative methodology for applying machine learning, in which a bespoke solution is formulated for each new application. The solution is expressed through a compact modelling language, and the corresponding custom machine learning code is then generated automatically. This model-based approach offers several major advantages, including the opportunity to create highly tailored models for specific scenarios, as well as rapid prototyping and comparison of a range of alternative models. Furthermore, newcomers to the field of machine learning do not have to learn about the huge range of traditional methods, but instead can focus their attention on understanding a single modelling environment. In this study, we show how probabilistic graphical models, coupled with efficient inference algorithms, provide a very flexible foundation for model-based machine learning, and we outline a large-scale commercial application of this framework involving tens of millions of users. We also describe the concept of probabilistic programming as a powerful software environment for model-based machine learning, and we discuss a specific probabilistic programming language called Infer.NET, which has been widely used in practical applications.

Employing Lexicalized Dependency Paths for Active Learning of Relation Extraction

Abstract: Active learning methods which present selected examples from the corpus for annotation provide more efficient learning of supervised relation extraction models, but they leave the developer in the unenviable role of a passive informant. To restore the developer’s proper role as a partner with the system, we must give the developer an ability to inspect the extraction model during development. We propose to make this possible through a representation based on lexicalized dependency paths (LDPs) coupled with an active learner for LDPs. We apply LDPs to both simulated and real active learning with ACE as evaluation and a year’s newswire for training and show that simulated active learning greatly reduces annotation cost while maintaining similar performance level of supervised learning, while real active learning yields comparable performance to the state-of-the-art using a small number of annotations.

Human-in-the-loop machine learning: a state of the art

Abstract: Abstract Researchers are defining new types of interactions between humans and machine learning algorithms generically called human-in-the-loop machine learning. Depending on who is in control of the learning process, we can identify: active learning, in which the system remains in control; interactive machine learning, in which there is a closer interaction between users and learning systems; and machine teaching, where human domain experts have control over the learning process. Aside from control, humans can also be involved in the learning process in other ways. In curriculum learning human domain experts try to impose some structure on the examples presented to improve the learning; in explainable AI the focus is on the ability of the model to explain to humans why a given solution was chosen. This collaboration between AI models and humans should not be limited only to the learning process; if we go further, we can see other terms that arise such as Usable and Useful AI. In this paper we review the state of the art of the techniques involved in the new forms of relationship between humans and ML algorithms. Our contribution is not merely listing the different approaches, but to provide definitions clarifying confusing, varied and sometimes contradictory terms; to elucidate and determine the boundaries between the different methods; and to correlate all the techniques searching for the connections and influences between them.

Hypothesis Learning in Automated Experiment: Application to Combinatorial Materials Libraries

Abstract: Machine learning is rapidly becoming an integral part of experimental physical discovery via automated and high-throughput synthesis, and active experiments in scattering and electron/probe microscopy. This, in turn, necessitates the development of active learning methods capable of exploring relevant parameter spaces with the smallest number of steps. Here, an active learning approach based on conavigation of the hypothesis and experimental spaces is introduced. This is realized by combining the structured Gaussian processes containing probabilistic models of the possible system's behaviors (hypotheses) with reinforcement learning policy refinement (discovery). This approach closely resembles classical human-driven physical discovery, when several alternative hypotheses realized via models with adjustable parameters are tested during an experiment. This approach is demonstrated for exploring concentration-induced phase transitions in combinatorial libraries of Sm-doped BiFeO3 using piezoresponse force microscopy, but it is straightforward to extend it to higher-dimensional parameter spaces and more complex physical problems once the experimental workflow and hypothesis generation are available.

AALpy: an active automata learning library

Abstract: Abstract AALpy is an extensible open-source Python library providing efficient implementations of active automata learning algorithms for deterministic, non-deterministic, and stochastic systems. We put a special focus on the conformance testing aspect in active automata learning, as well as on an intuitive and seamlessly integrated interface for learning automata characterizing real-world reactive systems. In this article, we present AALpy ’s core functionalities, illustrate its usage via examples, and evaluate its learning performance. Finally, we present selected case studies on learning models of various types of systems with AALpy .

Blended Learning and Student-centered Active Learning Environment: a Case Study with STEM Undergraduate Students

Abstract: Abstract This paper describes an embedded case study of “blended” teaching integrated with traditional lessons in a Student-Centered Active Learning Environment and social activities on the platform. The didactic phenomena were designed by creating learning environments, artifacts, and teaching/learning sequences in authentic educational contexts. We aim at improving the task design of a mathematics lesson with an impact on students’ performance in mathematics. Quantitative results show considerable benefits in the evolution of the use and coordination of several systems of semiotic representation. As a result, a better predisposition to the study of the subject seems to appear; moreover, the satisfaction test shows the achievement of alternative teaching methodologies for most of the students.

Conjunction of Active and Semi-Supervised Learning for Wireline Logs-Based Automatic Lithology Identification

Abstract: The sample labeling is the basis for wireline logs-based automatic lithology identification (WLALI). Due to the paucity of core data, the expert-annotation accounts for most labeling workload. In the traditional workflow of WLALI, the samples for labeling are passively selected. The labeled samples thus fail to reflect the overall distribution and feature of whole data set, leading to the difficulty in training the prediction model with robust generalization performance. To solve this problem, a novel workflow of hybrid active and semi-supervised learning-based WLALI has been proposed. First, we alter the conventional labeling process to active selection mode by employing the approach of active learning. In such way the labeled samples are representative of whole data features. Second, we propose a novel active learning algorithm based on density difference of Gaussian probability (DDGP) to obtain a better performance of sample query. Third, the semi-supervised learning method is introduced to combine the active learning algorithm, attempting to minimize the involvement of hand-crafted labeling under the target performance. In the evaluation, the proposed workflow not only has achieved favorable result in the uppercase image dataset, but also the practical application shows great promise. When compared with the workflow of random sampling + semi-supervised learning that is similar to the traditional mode, the average F1-score of proposed workflow of DDGP + semi-supervised learning can increase by approximately 3.47%.

Medical students’ self-regulation of learning in a blended learning environment: a systematic scoping review

Abstract: ABSTRACT Background Medical curricula are constantly evolving in response to the needs of society, accrediting bodies and developments in education and technology. The integration of blended learning modalities has challenged traditional methods of teaching, offering new prospects in the delivery of medical education. The purpose of this review is to explore how medical students adapt their learning behaviours in a Blended Learning environment to become more independent and self-regulated, in addition to highlighting potential avenues to enhance the curriculum and support student learning. Methods Using the approach described by Levac et al. (2010), which builds on Arksey and O’Malley’s framework, we conducted a literature search of the following databases: MEDLINE (Ovid), ERIC, EBSCO, SCOPUS and Google Scholar, utilising key terms and variants of “medical student’, ‘self-regulated learning’ and ‘blended learning’. The search yielded 305 studies which were further charted and screened according to the Joanna Briggs Institute. Results Forty-four studies were identified and selected for inclusion in this review. After full analysis of these studies, underpinned by Self-regulation theory, five major concepts associated with students’ learning behaviours in a Blended Learning environment were identified: Scaffolding of instructional guidance may support self-regulated learning; Self-regulated learning enhances academic performance; Self-regulated Learning improves study habits through resource selection; Blended learning drives student motivation and autonomy; and the Cognitive apprenticeship approach supports Self-regulated learning. Conclusion This review uncovers medical students’ learning behaviours within a Blended learning environment which is important to consider for curricular adaptations and student support.

How Useful Is Image-Based Active Learning for Plant Organ Segmentation?

Abstract: Training deep learning models typically requires a huge amount of labeled data which is expensive to acquire, especially in dense prediction tasks such as semantic segmentation. Moreover, plant phenotyping datasets pose additional challenges of heavy occlusion and varied lighting conditions which makes annotations more time-consuming to obtain. Active learning helps in reducing the annotation cost by selecting samples for labeling which are most informative to the model, thus improving model performance with fewer annotations. Active learning for semantic segmentation has been well studied on datasets such as PASCAL VOC and Cityscapes. However, its effectiveness on plant datasets has not received much importance. To bridge this gap, we empirically study and benchmark the effectiveness of four uncertainty-based active learning strategies on three natural plant organ segmentation datasets. We also study their behaviour in response to variations in training configurations in terms of augmentations used, the scale of training images, active learning batch sizes, and train-validation set splits.

Learning for Constrained Optimization: Identifying Optimal Active Constraint Sets

Abstract: In many engineered systems, optimization is used for decision making at time scales ranging from real-time operation to long-term planning. This process often involves solving similar optimization problems over and over again with slightly modified input parameters, often under tight latency requirements. We consider the problem of using the information available through this repeated solution process to learn important characteristics of the optimal solution as a function of the input parameters. Our proposed method is based on learning relevant sets of active constraints, from which the optimal solution can be obtained efficiently. Using active sets as features preserves information about the physics of the system, enables interpretable results, accounts for relevant safety constraints, and is easy to represent and encode. However, the total number of active sets is also very large, as it grows exponentially with system size. The key contribution of this paper is a streaming algorithm that learns the relevant active sets from training samples consisting of the input parameters and the corresponding optimal solution, without any restrictions on the problem type, problem structure or probability distribution of the input parameters. The algorithm comes with theoretical performance guarantees and is shown to converge fast for problem instances with a small number of relevant active sets. It can thus be used to establish simultaneously learn the relevant active sets and the practicability of the learning method. Through case studies in optimal power flow, supply chain planning, and shortest path routing, we demonstrate that often only a few active sets are relevant in practice, suggesting that active sets provide an appropriate level of abstraction for a learning algorithm to target.

Uncertainty-aware molecular dynamics from Bayesian active learning for phase transformations and thermal transport in SiC

Abstract: Abstract Machine learning interatomic force fields are promising for combining high computational efficiency and accuracy in modeling quantum interactions and simulating atomistic dynamics. Active learning methods have been recently developed to train force fields efficiently and automatically. Among them, Bayesian active learning utilizes principled uncertainty quantification to make data acquisition decisions. In this work, we present a general Bayesian active learning workflow, where the force field is constructed from a sparse Gaussian process regression model based on atomic cluster expansion descriptors. To circumvent the high computational cost of the sparse Gaussian process uncertainty calculation, we formulate a high-performance approximate mapping of the uncertainty and demonstrate a speedup of several orders of magnitude. We demonstrate the autonomous active learning workflow by training a Bayesian force field model for silicon carbide (SiC) polymorphs in only a few days of computer time and show that pressure-induced phase transformations are accurately captured. The resulting model exhibits close agreement with both ab initio calculations and experimental measurements, and outperforms existing empirical models on vibrational and thermal properties. The active learning workflow readily generalizes to a wide range of material systems and accelerates their computational understanding.

Future of the Flipped Classroom in Chemistry Education: Recognizing the Value of Independent Preclass Learning and Promoting Deeper Understanding of Chemical Ways of Thinking During In-Person Instruction

Abstract: Studies on the efficacy of the flipped classroom structure have become more prevalent in the chemistry education research literature over the past 10 years, and the body of research that has been compiled appears to indicate flipped classroom structures positively impact student learning outcomes relative to “teaching as usual” comparison groups. Though the definition of a flipped classroom is almost universally defined as a class structure in which some portion of traditional in-person didactic lecture is transformed to an independent preclass learning mode, with the traditional lecture being replaced by active learning exercises, the exact nature of the preclass and in-person leaning activities varies quite widely among the studies reported in the literature. Furthermore, though educational studies typically acknowledge that preclass learning modules can reduce cognitive load for students, researchers often emphasize increased active learning in class as the motivation for adopting the flipped classroom. This commentary will highlight the variety of instructional practices reported to be used during the in-person phase of flipped classrooms and how well-designed prelecture activities also contribute to meaningful learning. This will lead to the proposition that the flipped classroom should not be considered a teaching best practice in and of itself, but rather an in-person/independent hybrid learning scaffold that supports other evidence-based instructional practices. Chemistry education researchers and practitioners are encouraged to focus their efforts on optimizing the flipped classroom for chemistry-specific learning objectives, with the ongoing challenge being to promote both skill-based learning and deeper conceptual understanding of chemical ways of thinking.

A Review on Machine Learning Styles in Computer Vision—Techniques and Future Directions

Abstract: Computer applications have considerably shifted from single data processing to machine learning in recent years due to the accessibility and availability of massive volumes of data obtained through the internet and various sources. Machine learning is automating human assistance by training an algorithm on relevant data. Supervised, Unsupervised, and Reinforcement Learning are the three fundamental categories of machine learning techniques. In this paper, we have discussed the different learning styles used in the field of Computer vision, Deep Learning, Neural networks, and machine learning. Some of the most recent applications of machine learning in computer vision include object identification, object classification, and extracting usable information from images, graphic documents, and videos. Some machine learning techniques frequently include zero-shot learning, active learning, contrastive learning, self-supervised learning, life-long learning, semi-supervised learning, ensemble learning, sequential learning, and multi-view learning used in computer vision until now. There is a lack of systematic reviews about all learning styles. This paper presents literature analysis of how different machine learning styles evolved in the field of Artificial Intelligence (AI) for computer vision. This research examines and evaluates machine learning applications in computer vision and future forecasting. This paper will be helpful for researchers working with learning styles as it gives a deep insight into future directions.

Inclusive and active pedagogies reduce academic outcome gaps and improve long-term performance

Abstract: We assessed the impacts of the implementation of inclusive and active pedagogical approaches in an introductory biology sequence at a large, public research university in the northeast United States. We compared academic performance between these sections with other sections of the same course where didactic approaches were used over a five-year period. We also compared this five-year period (2014–2018) with the previous five years of the same courses. Additionally, we also tracked the academic performance of the students from the sections where active learning and inclusive teaching were used, as well as the more conventionally taught (lecture-based) sections in future, mandatory biology courses. We found that the inclusively taught section of the first semester of introductory biology increased the odds of students earning higher grades in that particular section. The active learning section in the second semester narrowed the ethnic performance gap when compared to similar sections, both historically and those run concurrently. Finally, students who matriculated into the inclusively taught section of biology in the first semester followed by the active learning section in the second semester of introductory biology performed better in 200-level biology courses than students who had zero semesters of either active or inclusive pedagogy in their introductory year. Our results suggest that active and inclusive pedagogies hold great promise for improving academic performance when compared to didactic approaches, however, questions remain on the most appropriate ways for capturing the impact of inclusive approaches. Implications for institutional approaches and policy are also discussed.

Deep active learning models for imbalanced image classification

Abstract: Active learning can query valuable samples in an unlabeled sample pool for annotation, thus building a more informative labeled dataset and reducing the annotation cost. However, traditional active learning methods are not effective in the task of imbalanced image classification for ignoring the distribution bias. In this study, we propose a Balanced Active Learning (BAL) method for imbalanced image classification. BAL estimates the probability of a sample belonging to minority or majority classes and compensates the annotation query for the minority classes, thus alleviating the class imbalance in the selected training samples. Experiments on three imbalanced image classification datasets, imbalanced CIFAR-10, ISIC2020, and Caltech256, showed that BAL achieved new state-of-the-art performance of active learning in a variety of classification tasks and different types of imbalance.

Disciplinary variations in student perceptions of active learning classrooms

Abstract: The use of active learning classrooms (ALCs) has attracted considerable attention in higher education research in the past two decades. Researchers have reported the positive effects of ALCs on student learning. However, most of the published studies on the topic have been based on just one or a handful of academic disciplines. In this work, we conducted a large-scale empirical study across different academic disciplines, aiming to uncover whether there was any disciplinary variation in student perceptions about their learning experiences when ALCs were adopted as the key learning environment. During the four-year period of study, more than 30,000 students’ quantitative responses from 550 undergraduate courses across different disciplines were collected. Independent-samples t -test results revealed that when ALCs were used for courses in the sciences, technologies, arts, and humanities, students perceived themselves to have had significantly better experiences in three aspects of learning: the encouragement of innovation and creativity, course design, and support for critical thinking. Students enrolled in courses under the disciplinary area of the study of societies also reported a slight benefit from being taught in ALCs. In contrast, these three aspects of learning were not reported to be enhanced by ALCs among students who were studying business organization courses. Nonetheless, students studying courses across all of the disciplines felt that the course content was less difficult when ALCs were used. The results of this work suggest that student perceptions of their learning experiences and the level of positive effects on ALCs are indeed varied across academic disciplines.

The Effectiveness of the Hybrid Learning Materials with the Application of Problem Based Learning Model (Hybryd-PBL) to Improve Learning Outcomes during the COVID-19 Pandemic

Abstract: Study aimed to see the effectiveness of the hybrid learning materials application with Problem based learning (PBL) model to improve students’ outcome in evaluation course in learning outcomes at the Business Education Study Program, State University of Medan. The study used a quasi-experimental method of the Posttest control group design. Data collection was carried out by conducting tests after class. The sample selected in this study were all students of the Business Education study program in evaluation course in learning outcomes. The results showed that the application of hybrid learning with the PBL model was effective to improve learning outcomes in evaluation course in learning outcomes . The results of this study also found that the independence and creativity of student learning was higher by applying hybrid learning with problem-based learning models compared to the control class. The results of this study contributed to the lecturers' evaluation of learning outcomes in improving the quality of learning, which has been traditional, by applying hybrid learning teaching materials with problem-based learning models, especially during the current Covid 19 pandemic and to answer the demands of learning in the industrial revolution era. 4.0.

Investigating Characteristics of Learning Environments During the COVID-19 Pandemic: A Systematic Review

Abstract: Dramatic change in learning environments during the COVID-19 pandemic highlighted the significance of virtual learning and led to more interactive learning environments. Quick adoption of online and social interactive learning in many universities around the world raised challenges and emphasized the importance of investigating different learning environments. This paper investigates the accelerated transition in education from traditional learning environments through online learning environments to social innovative learning environments, and the latest trends of this change. The stages of transition were divided into three parts: before, during, and after the COVID-19 pandemic, which was the reason for this accelerated change. Features and characteristics of each stage of transition were analyzed and discussed, based on the following factors: edu-space and classrooms, the learning and teaching process, curricular choices, information and communication technology applications, students’ and educators’ perceptions, edu-approaches, and knowledge transformation. A systematic review approach was used to investigate learning environments based on the literature reviews of previous publications. Analysis of these features revealed the main characteristics and differences in each stage. New trends in online learning environments and social innovative learning environments were identified including cloud platforms, massive open online courses, digital learning management systems, open educational resources, open educational practices, m-learning, and social network applications. Finally, this study makes two recommendations: 1) the adoption of online learning environments and social innovative learning environment applications to continue the e-learning process during the pandemic, and 2) the enhanced usage of online learning environments and social innovative learning environment applications in the future by educational institutions and governments.

Improving Computational Efficiency of Machine Learning Modeling of Nonlinear Processes Using Sensitivity Analysis and Active Learning

Abstract: In this work, we develop a model reduction method using sensitivity analysis and active learning to improve the computational efficiency of machine learning modeling of nonlinear processes. Specifically, sensitivity analysis is first used to identify important connections between model outputs and inputs. Subsequently, active learning is used to enrich the training set by iteratively identifying the training data that most efficiently improve model performance. Reduced-order recurrent neural networks (RNN) using the important input features obtained from sensitivity analysis are developed to approximate the nonlinear system, and are incorporated within model predictive control (MPC) to stabilize the nonlinear system at the steady-state. Finally, the effectiveness of the proposed machine learning modeling approach using sensitivity analysis and active learning and machine-learning-based predictive control scheme are demonstrated using a reactor-reactor-separator process example.

Supporting students’ active learning with a computer based tool

Abstract: In recent years, the use of information technology to promote active learning in higher education has raised great interest. Teachers are continuously challenged to identify new research-informed approaches and educational practices for supporting students to actively learn and apply their knowledge. The present study tests the effects on students’ learning outcomes of an ad hoc developed learning tool (QLearn) which integrates three active learning strategies, previously empirically validated in face-to-face educational contexts. By using the QLearn software, students can generate questions, explain and develop answers, receive feedback from teacher and test their knowledge. Using a quasi-experimental design, we analyzed whether, in various course settings and instructional contexts, the students who use QLearn, as a support in their learning process, demonstrate a different learning performance compared to students who learn the same content by using their preferred learning strategies. The interventions were offered on a voluntary basis and implied participants from different fields (computer science, psychology) and different study levels (undergraduate and master’s level). The results showed that some groups of our participants significantly benefits from the use of QLearn platform. The outcomes of the present research advanced our understanding of the efficiency of technology-sustained learning in educational contexts and offer a promising strategy for facilitating the active involvement of students in the learning process.

Towards Active Evidence-Based Learning in Engineering Education: A Systematic Literature Review of PBL, PjBL, and CBL

Abstract: Implementing active learning methods in engineering education is becoming the new norm and is seen as a prerequisite to prepare future engineers not only for their professional life, but also to tackle global issues. Teachers at higher education institutions are expected and encouraged to introduce their students to active learning experiences, such as problem-, project-, and more recently, challenge-based learning. Teachers have to shift from more traditional teacher-centered education to becoming instructional designers of student-centered education. However, instructional designers (especially novice) often interpret and adapt even well-established methods, such as problem-based learning and project-based learning, such that the intended value thereof risks being weakened. When it comes to more recent educational settings or frameworks, such as challenge-based learning, the practices are not well established yet, so there might be even more experimentation with implementation, especially drawing inspiration from other active learning methods. By conducting a systematic literature analysis of research on problem-based learning, project-based learning, and challenge-based learning, the present paper aims to shed more light on the different steps of instructional design in implementing the three methods. Based on the analysis and synthesis of empirical findings, the paper explores the instructional design stages according to the ADDIE (analysis, design, development, implementation, and evaluation) model and provides recommendations for teacher practitioners.