Active learning for hierarchical multi-label classification

This paper proposes a hierarchical multi-label classification method based on SVM with composite kernel for solving gene function prediction. The hierarchical multi-label classification problem is resolved into a set of binary classification tasks. A composite kernel based SVM (ck-SVM) is introduced to deal with the binary classification tasks. In estimation procedure of ck-SVM, a supervised clustering with over-sampling strategy is introduced for solving imbalance dataset learning problem and improve classification performance. Experimental results on benchmark datasets demonstrate that the proposed method improves the classification performance efficiently.

Metallocene and Single Site Polymers (#1624)

In this work, we extend the sequential multitask learning model called Resource Allocating Network for Multi-Task Pattern Recognition (RAN-MTPR) by introducing the following new learning functions: multi-label recognition, semi-supervised task learning and active learning. The extended RAN-MTPR can learn a training data with multiple class labels, can handle a semi-supervised setting for task learning, and can actively request class labels for unsure inputs. We evaluate the performance of the extended RAN-MTPR, and we know that the above three functions work well to enhance the generalization performance for pattern recognition problems.

This study examines the application of active learning in the Law major of the University of Costa Rica, Guanacaste campus, during 2022. This research has an exploratory scope and follows a qualitative approach. One of its main results reflects the identification of ten types of active methods applied in the courses. Professors consider that this approach contributes to promoting a comprehensive and in-depth learning of the major. It also enables a better link between theory and practice, which encourages the development of subject-specific and soft skills. Students hold a positive view towards using active learning methods in class since they believe them to provide better learning conditions. It can be concluded that the frequent use of active learning methods in Law courses shows a noticeable integration of this approach to the pedagogic culture of the major. Both faculty and students view the educational process as a complex phenomenon integrating concepts, practice, and emotions. Thus, they perceive these methods as ideal teaching instruments. Therefore, implementing active learning in the Law major is aligned with the theoretical principles about this topic as well as responds to the current educational requirements.

There is a growing interest for specialized classrooms, termed active learning classrooms (ALC), which are designed to facilitate the use of active learning methods and information and communication technologies (ICT) by students. Thanks to pioneering studies such as SCALE-UP, there is a better understanding of the benefits of these classrooms and the pedagogy taking place in them. Teachers accustomed to traditional classes have to change many aspects of their pedagogy in order to reap the benefits of the ALCs, however. The purpose of this research is to gain a better understanding of the adoption process of an ALC by teachers and how its adoption modify teaching preferences and practices. Relying on an in-depth case study methodology founded on interviews and questionnaires about the adoption of innovations (CBAM), Approaches to Teaching Inventory, technopedagogical competencies and collaborative, competitive or individual teaching preferences, this article describes the cases of two teachers who used an ALC over a three-semester period. The results show that the teachers develop their courses quickly, with an emphasis on the active learning aspects of their pedagogy rather than on ICT integration, and that there are a lot of personal and management concerns. When the pedagogical changes are stabilized, the teachers retained their personal concerns about the innovation and were highly motivated to collaborate with other ALC users. Finally, apparently minor increases in student-centred teaching approaches result in significant pedagogical changes when they are studied qualitatively. These changes did not lead to a reduction in teacher-centred teaching approaches, suggesting that a significant portion of teacher-directed activities remain.

In machine learning,active learningrefers to algorithms that autonomously select the data points from which they will learn. There are many data mining applications in which large amounts of unlabeled data are readily available, but labels (e.g., human annotations or results coming from complex experiments) are costly to obtain. In such scenarios, an active learning algorithm aims at identifying data points that, if labeled and used for training, would most improve the learned model. Labels are then obtained only for the most promising data points. This speeds up learning and reduces labeling costs. Support vector machine (SVM) classifiers are particularly well‐suited for active learning due to their convenient mathematical properties. They perform linear classification, typically in a kernel‐induced feature space, which makes expressing the distance of a data point from the decision boundary straightforward. Furthermore, heuristics can efficiently help estimate how strongly learning from a data point influences the current model. This information can be used to actively select training samples. After a brief introduction to the active learning problem, we discuss different query strategies for selecting informative data points and review how these strategies give rise to different variants of active learning withSVMs.This article is categorized under:Technologies > Machine Learning

In this paper, we present a hierarchical multi-label classification system for visual concepts detection and image annotation. Hierarchical multi-label classification (HMLC) is a variant of classification where an instance may belong to multiple classes at the same time and these classes/labels are organized in a hierarchy. The system is composed of two parts: feature extraction and classification/annotation. The feature extraction part provides global and local descriptions of the images. These descriptions are then used to learn a classifier and to annotate an image with the corresponding concepts. To this end, we use predictive clustering trees (PCTs), which are able to classify target concepts that are organized in a hierarchy. Our approach to HMLC exploits the annotation hierarchy by building a single predictive clustering tree that can simultaneously predict all of the labels used to annotate an image. Moreover, we constructed ensembles (random forests) of PCTs, to improve the predictive performance. We tested our system on the image database from the ImageCLEF@ICPR 2010 photo annotation task. The extensive experiments conducted on the benchmark database show that our system has very high predictive performance and can be easily scaled to large number of visual concepts and large amounts of data.

The classification problems described in the Machine Learning literature usually relate to the classification of data in which each example is associated to a class belonging to a finite set of classes, all at the same level. However, there are classification issues, of a hierarchical nature, where the classes can be either subclasses or super classes of other classes. In many hierarchical problems, one or more examples may be associated with more than one class simultaneously. These problems are known as hierarchical multi-label classification (HMC) problems. In this work, the ML-KNN algorithm was used to predict hierarchical multi-label problems, in order to determine the number of classes that can be assigned to an example. Through the experiments performed on 10 protein function databases and the statistical analysis of the results, it can be shown that the adaptations performed in the ML-KNN algorithm brought significant performance improvements based on the hierarchical precision and recall metrics Hierarchical.

The study aims to theoretically substantiate the opportunities to use active learning methods in training future specialists in the tourism industry for leading excursions. The authors conclude that the forms and methods of active learning that can be utilized in the training of future tourism industry specialists for leading excursions include: practical lessons in the form of excursions and research lessons in the urban environment, individual types of lectures (dialogue lecture-conversation, lecture-discussion, problem-based lecture, lecture with an analysis of specific professional cases) and seminars-discussion, as well as independent work in the form of individual research assignments. The priority form of active learning is considered to be practical lessons in the form of excursions and research lessons organized immediately in the urban environment.

Teachers need to follow adult learning principles and students need to have discussion and problem-solving skills for the use of active learning methods in classroom. We aimed to find out whether MBBS freshers had group discussion and problem-solving skills. A free group discussion session for problem-solving using adult learning principles was held during the foundation course for MBBS freshers. The task for group discussion was to give suggestions to improve the situation indicated by the given general medical professional problem. Qualitative deductive thematic analysis was used to analyze suggestions which could be grouped under the themes of seven roles of CanMEDS Physician Competency Framework. The results of deductive thematic analysis of suggestions of student groups are described under seven themes of the roles of the CanMEDS Physician Competency Framework. MBBS freshers showed excellent group discussion, plenary presentation, and problem-solving skills for the given problem. They also had relevant general knowledge about medical profession. Their problem-oriented thinking skills were facilitated using adult learning principles. It is recommended that the use of group discussion for problem-solving with facilitation by using adult learning principles should be emphasized in undergraduate medical education to promote the use of active learning methods in classroom.

Cost-sensitive active learning through statistical methods

Deep hiearchical multi-label classification applied to chest X-ray abnormality taxonomies.

Active machine learning is an experimental design approach that puts machine learning models in the driver seat of data acquisition and automated optimization. Introduced to drug discovery approximately 15 years ago, a handful of impressive studies have revealed the potential of active machine learning to guide molecular discovery and optimization. Most recently, researchers have shown that active learning can also be applied to datasets retrospectively, enabling automated data curation to train powerful machine learning models on small datasets. This chapter reviews the key findings from these active learning studies and summarizes remaining challenges and future opportunities of active learning as an adaptive learning technology in the context of drug discovery.

In this study, a hierarchical multi-label classification method called HFFNet-2d is proposed for the automatic classification of scanning electron microscope (SEM) images of metal failure. The method combines the advantages of convolutional neural networks (CNN) and Vision Transformers (ViT) to effectively realize hierarchical feature extraction and classification of SEM images of fracture morphologies, enabling accurate identification of metal failure mechanisms at different scales. The dataset of high-quality SEM images in this work is sourced from reputable materials science publications for its comprehensive coverage of various failure modes and its suitability for training and validating the hierarchical multi-label classification model. The HFFNet-2d model can achieve a high accuracy of 97.71% in the first-level classification and 92.62% in the second-level sub-category identification. This performance surpasses the human experts on the same task. To ensure that the model predictions are sufficiently reliable, a multi-level gradcam algorithm is also introduced for checking the regions of interest of the Hierarchical model at two levels and the comparisons are made with human experts. It is anticipated that the optimization and extension of HFFNet-2d are conducive in diverse material systems and application scenarios to accelerate the intelligent process of material development and failure analysis, ultimately supporting the design of reliable and high-performance engineering materials.

Hierarchical multi-label text classification (HMTC) has become rather challenging when it requires handling large sets of closely related categories. We present a novel software that provides a hybrid embedding-based text representation for HMTC, shortened as HE-HMTC. It made full use of categories’ structure and their labels semantics to enrich the text representation, and therefore attempted to improve the classification performance of the text. In addition, HE-HMTC facilitates improving the accuracy of HMTC tasks. Moreover, our HE-HMTC can easily be generalized in other hierarchical classification tasks and achieve superior performance.