Amalgamation Of Knowledge Research Articles

Using drawing to learn about figurative language among tertiary EFL English literature students

Visual art has gained popularity in education and highlighted remarkable results in terms of improving the learning experience. This has triggered a response to investigate how visual art, particularly drawing, can be applied in an educational setting. Nevertheless, there is still a missing link on how drawing functions as a tool for tertiary English as a Foreign Language (EFL) English Literature students to learn about, specifically, figurative language. This paper employs an action research method to investigate how drawing using soft pastels is used to scaffold learning about figurative language at the same time, study its benefits. Using selected literary texts involving mainly eleven tertiary EFL English Literature students, this action research qualitative study incorporates Richard Mayer’s Cognitive Theory of Multimedia Learning in three repeated cycles. This investigation may be construed as a means to transition from conventional pedagogical methodologies into contemporary paradigms, functioning as a facilitative instrument for the amalgamation of knowledge about learning literature while enhancing confidence, higher-order cognitive skills, and expressive capabilities. It allows educators to adapt their approaches, stimulating increased reflection and participation. The findings delineate two primary strategies employed in elucidating the efficacy of incorporating drawing as a pedagogical tool for English Literature students, i.e., mining to trigger thoughts and engaging with senses to assist EFL students in learning about figurative language. It can be asserted that integrating drawing empowers individuals to formulate idiosyncratic responses by exploring concepts, articulating emotions, validating responses, and discerning the profound dimensions of literary texts.

Transformation of Mandatory Physiotherapy Internship into A Structured Competency-Based Program: A Pilot Study.

Health professions education is challenging in terms of developing ways to construct and assess the amalgamation of knowledge, skills and attitudes that result in novice graduates to be self-aware and confident to practice without supervision. Currently, the Physiotherapy internship program in India does not have a comprehensive competency-based framework. With the current batch of interns facing the wrath of the COVID-19 pandemic during their undergraduate training, it is a matter of concern to ensure they have developed the required competencies to handle patients independently. This is a single-group interventional study using convenience sampling. A competency framework inclusive of a comprehensive assessment and training module was developed and introduced to 27 physiotherapy interns. Every intern underwent 10 training sessions along with 8 competency assessments during 6 months of internship training. A pre-post indigenous questionnaire was used to assess their self-perceived competency along with feedback questionnaire taken at the end of the module. Implementation of the competency framework revealed a statistically significant difference in the self-perceived competency (pre=84.36±10.98, post=98.55±8.74, p≤0.000). Having uniformity in assessment techniques among the faculty, being exposed to challenging cases during assessment, having training module at the beginning of the internship were some of the suggestions given by the participants. Competency based education offers to be an effective technique in health professional program. Incorporating a competency-based training would help the learners to understand their strengths and weaknesses that would go a long way to develop competent health-care professionals. Developing such framework in curriculum will emphasize standardization of learning outcomes, thereby resulting in quality education, further enhancing patient care, and improving the health outcomes worldwide.

Amalgamating Multi-Task Models with Heterogeneous Architectures

Multi-task learning (MTL) is essential for real-world applications that handle multiple tasks simultaneously, such as selfdriving cars. MTL methods improve the performance of all tasks by utilizing information across tasks to learn a robust shared representation. However, acquiring sufficient labeled data tends to be extremely expensive, especially when having to support many tasks. Recently, Knowledge Amalgamation (KA) has emerged as an effective strategy for addressing the lack of labels by instead learning directly from pretrained models (teachers). KA learns one unified multi-task student that masters all tasks across all teachers. Existing KA for MTL works are limited to teachers with identical architectures, and thus propose layer-to-layer based approaches. Unfortunately, in practice, teachers may have heterogeneous architectures; their layers may not be aligned and their dimensionalities or scales may be incompatible. Amalgamating multi-task teachers with heterogeneous architectures remains an open problem. For this, we design Versatile Common Feature Consolidator (VENUS), the first solution to this problem. VENUS fuses knowledge from the shared representations of each teacher into one unified generalized representation for all tasks. Specifically, we design the Feature Consolidator network that leverages an array of teacher-specific trainable adaptors. These adaptors enable the student to learn from multiple teachers, even if they have incompatible learned representations. We demonstrate that VENUS outperforms five alternative methods on numerous benchmark datasets across a broad spectrum of experiments.

Cross-factory Polarizer Sheet Surface Defect Inspection System Based on Multi-Teacher Knowledge Amalgamation

Cross-factory Polarizer Sheet Surface Defect Inspection System Based on Multi-Teacher Knowledge Amalgamation

Federated knowledge amalgamation with unbiased semantic attributes under cloud–edge collaboration for heterogeneous fault diagnosis

Federated fault diagnosis attracts increasing attention in industrial cloud–edge collaboration scenarios due to the consideration of privacy and security issues. However, statistical heterogeneity caused by non-independent and identically distributed (non-IID) fault data is a ubiquitous problem. Previous methods designed for this problem mainly apply weighted averaging or knowledge distillation to integrate biased local knowledge from clients, ignoring the discrepancy between knowledge extracted from non-IID data solely. This is prone to result in slow and unstable model convergence and global model drift. In this article, a Federated Knowledge Amalgamation (FedKA) framework is proposed to alleviate the knowledge discrepancy caused by statistical heterogeneity. To mitigate the knowledge discrepancy, we first reveal and mine unbiased public knowledge for industrial fault diagnosis, i.e., fault attributes, which describe the fault characterization and mechanism without changing with data distribution. Then, a generative mapping between unbiased fault attributes and fault data is established for clients to incorporate unbiased knowledge into the diagnostic model. Moreover, to mitigate knowledge discrepancy and overcome the global model drift, a knowledge amalgamation algorithm is designed. Through the proposed layer-wise alignment, the multi-level fault semantic knowledge that stems from fault attributes is finely exploited and aggregated. Additionally, unlike the common fixed weighting strategy, we explore an adaptive weighting mechanism through the sample-wise diagnostic confidence to further alleviate the global model drift. Extensive experiments on a real thermal power plant process demonstrate the superiority of the proposed method for federated fault diagnosis.

Knowledge Amalgamation for Multi-Label Classification via Label Dependency Transfer

Multi-label classification (MLC), which assigns multiple labels to each instance, is crucial to domains from computer vision to text mining. Conventional methods for MLC require huge amounts of labeled data to capture complex dependencies between labels. However, such labeled datasets are expensive, or even impossible, to acquire. Worse yet, these pre-trained MLC models can only be used for the particular label set covered in the training data. Despite this severe limitation, few methods exist for expanding the set of labels predicted by pre-trained models. Instead, we acquire vast amounts of new labeled data and retrain a new model from scratch. Here, we propose combining the knowledge from multiple pre-trained models (teachers) to train a new student model that covers the union of the labels predicted by this set of teachers. This student supports a broader label set than any one of its teachers without using labeled data. We call this new problem knowledge amalgamation for multi-label classification. Our new method, Adaptive KNowledge Transfer (ANT), trains a student by learning from each teacher’s partial knowledge of label dependencies to infer the global dependencies between all labels across the teachers. We show that ANT succeeds in unifying label dependencies among teachers, outperforming five state-of-the-art methods on eight real-world datasets.

Federated selective aggregation for on-device knowledge amalgamation

In this paper, we explore a new knowledge amalgamation problem, termed Federated Selective Aggregation for on-device knowledge amalgamation (FedSA). FedSA aims to train an on-device student model for a new task with the help of several decentralized teachers whose pre-training tasks and data are different and agnostic. Our motivation for investigating such a problem setup stems from a recent dilemma of model sharing. Due to privacy, security, or intellectual property issues, the pre-trained models are, however, not able to be shared, and the resources of devices are usually limited. The proposed FedSA offers a solution to this dilemma and makes it one step further, again, the method can be employed on low-power and resource-limited devices. To this end, we proposed a dedicated strategy for handling knowledge amalgamation. Specifically, our student-training process is driven by a novel saliency-based approach that adaptively selects teachers as the participants and integrates their representative capabilities into the student. To evaluate the effectiveness of FedSA, we conduct experiments on both single-task and multi-task settings. Experimental results demonstrate that FedSA effectively amalgamates knowledge from decentralized models and achieves competitive performance to centralized baselines.

The evolution of women ophthalmologists in India.

The evolution of women ophthalmologists in India.

Importance of interprofessional collaboration in teaching diabetic foot self-management: Perceptions of health care professionals

BackgroundDiabetes and associated diabetic foot ulcers require coordinated management, including several health care professional (HCPs). Therefore, an interprofessional (IP) team-based approach is essential for effectively managing and educating the population on diabetic foot self-management strategies. However, the perceptions of the HCPs related to the importance of IP teamwork in diabetic foot care and their readiness to work in an IP team are less explored. MethodsThis qualitative study aimed to investigate the HCPs' perception on the importance of IP collaboration in diabetic foot care and their readiness to work as part of an IP team in teaching patients’ effective diabetic foot self-management techniques. The study includes seven HCPs involved in diabetic foot management patient education and treatment. In-depth interviews using an interview guide and a thematic analysis of the interview transcripts were undertaken. ResultsEvery participant was strongly motivated to be an IP team member. The following themes were identified that supported IP team-based approach in diabetic foot care: ‘patient-centric practices,’ ‘comprehensive care,’ ‘teamwork and coordination for improved patient outcome,’ ‘integrated approach,’ ‘professional knowledge amalgamation,’ ‘time-management,’ ‘education in a favorable environment,’ ‘constant motivation and support through educational modules,’ and ‘commitment and policy reforms.’ ConclusionsThe HCPs' were aware about the benefits of IP team-based approach in diabetic foot care. They further expressed their willingness to work as part of an IP team and suggested appropriate teaching methods for diabetic foot self-management. This manuscript thus attempts to create an awareness about the importance of IP approach in teaching the proper practices of diabetic foot care. The problems faced in implementing an IP approach in diabetic foot care and the possible solutions are also discussed.

Assessment Pedagogical Content Knowledge (APCK): A Model for Implementing Science Classroom Assessments

Shulman’s Pedagogical Content Knowledge (PCK), developed in the mid-1980s has undergone a number of revisions, with the aim of building upon the concept. The current paper introduces a new theoretical concept named Assessment Pedagogical Content Knowledge (APCK), for implementing formative assessment during science teaching in order to aid students’ learning. The new concept, APCK, is an amalgam of knowledge that draws on how to combine assessment with pedagogy. It involves a combination of knowledge of assessment, the skills of assessment and the contents that are taught. The appropriate use of APCK inures to the benefit of students’ science learning. The study uses both qualitative and quantitative approaches to show that integration of assessment with instruction has proved valuable for science learning. The feedback, which teachers give to their students during teaching are used by students to improve their learning of science. This shows that a good grasp of the concept will enable teachers to use APCK formatively when teaching science and their students will improve their science learning. Hence, APCK holds a great potential, especially for science teachers, to develop their skills of interweaving instruction with assessments to promote learning. This study then proposes that APCK of science teachers can be identified through their knowledge and practice of the inter-relationships between pedagogy and assessments. Keywords: APCK, assessments, curriculum, formative assessment, purposes of assessment DOI: 10.7176/JEP/14-6-03 Publication date: February 28 th 2023

Diabetic wounds and artificial intelligence: A mini-review.

Diabetic wound takes longer time to heal due to micro and macro-vascular ailment. This longer healing time can lead to infections and other health complications. Foot ulcers are one of the most common diabetic wounds. These are one of the leading cause of amputations. Medical science is continuously striving for improving quality of human life. A recent trend of amalgamation of knowledge, efforts and technological advancement of medical science experts and artificial intelligence researchers, has made tremendous success in diagnosis, prognosis and treatment of a variety of diseases. Diabetic wounds are no exception, as artificial intelligence experts are putting their research efforts to apply latest technological advancements in the field to help medical care personnel to deal with diabetic wounds in more effective manner. The presented study reviews the diagnostic and treatment research under the umbrella of Artificial Intelligence and computational science, for diabetic wound healing. Framework for diabetic wound assessment using artificial intelligence is presented. Moreover, this review is focused on existing and potential contribution of artificial intelligence to improve medical services for diabetic wound patients. The article also discusses the future directions for the betterment of the field that can lead to facilitate both, clinician and patients.

Knowledge Amalgamation for Object Detection With Transformers.

Knowledge amalgamation (KA) is a novel deep model reusing task aiming to transfer knowledge from several well-trained teachers to a multi-talented and compact student. Currently, most of these approaches are tailored for convolutional neural networks (CNNs). However, there is a tendency that Transformers, with a completely different architecture, are starting to challenge the domination of CNNs in many computer vision tasks. Nevertheless, directly applying the previous KA methods to Transformers leads to severe performance degradation. In this work, we explore a more effective KA scheme for Transformer-based object detection models. Specifically, considering the architecture characteristics of Transformers, we propose to dissolve the KA into two aspects: sequence-level amalgamation (SA) and task-level amalgamation (TA). In particular, a hint is generated within the sequence-level amalgamation by concatenating teacher sequences instead of redundantly aggregating them to a fixed-size one as previous KA approaches. Besides, the student learns heterogeneous detection tasks through soft targets with efficiency in the task-level amalgamation. Extensive experiments on PASCAL VOC and COCO have unfolded that the sequence-level amalgamation significantly boosts the performance of students, while the previous methods impair the students. Moreover, the Transformer-based students excel in learning amalgamated knowledge, as they have mastered heterogeneous detection tasks rapidly and achieved superior or at least comparable performance to those of the teachers in their specializations.

Generalized Voice Spoofing Detection via Integral Knowledge Amalgamation

Most of the voice spoofing detection methods are designed for specific kinds of spoofing attacks, synthetic or replay. In practice, however, there is no prior information about these two kinds of spoofing attacks. To this end, this paper proposes a generalized voice spoofing detection method based on integral knowledge amalgamation to detect jointly synthetic attacks and replay attacks. Two amalgamation mechanisms, feature amalgamation and structure amalgamation, are designed from different perspectives, so that the model can generalize better and run fast. Specifically, the feature amalgamation transfers the high-level sematic knowledge from two teacher models to the compact model. The structure amalgamation employs the adversarial learning to ensure the global structure consistency of two teacher models and a student model. In addition, the feature matching loss is introduced to capture the distinctive features of synthetic attacks and replay attacks. We conduct extensive experiments on logical access (LA) scenario and physical access (PA) scenario of ASVspoof 2019 dataset to verify the validity of the proposed method. The experimental results show that compared with the most advanced generalized voice spoofing detection methods, the proposed method achieves a comparable or even better performance. In particular, our method gains the state-of-the-art detection capability on LA scenario. Moreover, our method achieves similar or even outstanding detectability when compared with specialized anti-spoofing methods.

Hierarchical knowledge amalgamation with dual discriminative feature alignment

Heterogeneous Knowledge Amalgamation (HKA) algorithms attempt to learn a versatile and lightweight student neural network from multiple pre-trained heterogeneous teachers. They encourage the student not only to produce the same prediction as the teachers but also to imitate each teacher’s features separately in a learned Common Feature Space (CFS) by using Maximum Mean Discrepancy (MMD). However, there is no theoretical guarantee of the Out-of-Distribution robustness of teacher models in CFS, which can cause an overlap of feature representations when mapping unknown category samples. Furthermore, global alignment MMD can easily result in a negative transfer without considering class-level alignment and the relationships among all teachers. To overcome these issues, we propose a Dual Discriminative Feature Alignment (DDFA) framework, consisting of a Discriminative Centroid Clustering Strategy (DCCS) and a Joint Group Feature Alignment method (JGFA). DCCS promotes the class-separability of the teachers’ features to alleviate the overlap issue. Meanwhile, JGFA decouples the complex discrepancy among teachers and the student at both category and group levels, extending MMD to align the features discriminatively. We test our model on a list of benchmarks and demonstrate that the learned student is robust and even outperforms its teachers in most cases.

Morele vorming en filosofie opnieuw met elkaar verbinden

Studenten van de Universiteit van Tilburg onderscheiden zich van die van andere universiteiten doordat er in hun opleiding veel aandacht wordt besteed aan karaktervorming. Op welke manier kunnen docenten Studenten helpen om hun karakter te vormen? Kern van mijn betoog is dat morele vorming hierbij een belangrijke rol speelt en dat filosofiedocenten bij uitstek ertoe doen om de morele vorming plaats te laten vinden. Morele vorming en filosofie worden door mij opnieuw met elkaar verbonden, in de zin dat een verbinding via de empirie wordt gemaakt. Ik zal voortbouwend op Martha Nussbaum morele vorming plaatsen in een context van verschillende manieren van ethiek leren; de pedagogische taak van de docent schetsen; voortbouwend op John Dewey en Hannah Arendt empirisch onderzochte kenmerken geven van effectieve filosofielessen; en voortbouwend op Pierre Hadot het goede als oefening die het karakter vormt voorstellen. Zo kunnen we, ondersteund door onderzoek uit faculteiten waar karaktervorming tot de kern behoort, filosofie en theologie, werken aan de morele vorming van studenten. Niet alleen studenten en hun docenten zullen hiervan profiteren, maar ook de gemeenschap rond de studenten (bijvoorbeeld leeftijdgenoten en familie) en hun beroepspraktijken (pastoraat, justitie, defensie, geestelijke verzorging en onderwijs).

Adopting A New Hybrid Force Model: A Survey During Covid-19 In Indian Higher Education

This quantitative study proposes and validates the Hybrid force model by analysing the challenges and opportunities in online education during COVID-19 through an online survey from students and teachers of Indian higher education institutions. The proposed model considers the learner as a system of study and talks about the amalgamation of knowledge, human force (teacher), and techno-logical force for better online learning opportunities and experience. The results show Google Meet, Zoom, and Microsoft Team as the main online platforms for imparting the classes while almost 80% of the students prefer a non-traditional teaching method (online and hybrid). The results highlight that more than 88% of the students prefer to have a personal teaching assistant in their learning trajectory.

Semi-Supervised Knowledge Amalgamation for Sequence Classification

Sequence classification is essential for domains from medical diagnosis to online advertising. In these settings, data are typically proprietary, and annotations are expensive to acquire. Often times, so few annotations are available that training a robust model from scratch is impractical. Recently, knowledge amalgamation (KA) has emerged as a promising strategy for training models without this hard-to-come-by labeled training dataset. To achieve this, KA methods combine the knowledge of multiple pre-trained teacher models (trained on different classification tasks and proprietary datasets) into one student model that becomes an expert on the union of all teachers’ classes. However, we demonstrate that the state-of-the-art solutions fail in the presence of overconfident teachers, which make confident but incorrect predictions for instances from classes upon which they were not trained. Additionally, to-date no work has explored KA for sequence models. Therefore, we propose and then solve the open problem of semi-supervised KA for sequence classification (SKA). Our SKA approach first learns to estimate how trustworthy each teacher is for a given instance, then rescales the predicted probabilities from all teachers to supervise a student model. Our solution overcomes overconfident teachers through careful use of a very small amount of labeled instances. We demonstrate that this approach beats eight state-of-the-art alternatives on four real-world datasets by on average 15% in accuracy with as little as 2% of training data being annotated.

Making Marvels: Science and Splendor at the Courts of Europe ed. by Wolfram Koeppe (review).

Reviewed by: Making Marvels: Science and Splendor at the Courts of Europe ed. by Wolfram Koeppe Mark A. Meadow (bio) Making Marvels: Science and Splendor at the Courts of Europe Edited by Wolfram Koeppe. New Haven: The Metropolitan Museum of Art, New York, 2019. Pp. 308. This exhibition catalog, richly illustrated and furnished with nine brief but significant essays, accompanied the exhibition of the same name held at the Metropolitan Museum of Art in New York, 2019–2020. The focus of Making Marvels is technology, as revealed in the 154 entries, which include Renaissance and Baroque mathematical and musical instruments, automata, tools, finely crafted jewelry, furniture, turned ivory, cups, and other vessels. The title itself, which gives an equal emphasis to the production, or "making," of these objects as well as to their reception, the "marvel" that they elicited from their original owners and from scholars and museum-goers today, makes the book's double imperative evident—to demonstrate that these objects exemplify an important period of technological innovation and are visually stunning artifacts worthy of our aesthetic attention. The paired terms "science" and "splendor" in the subtitle similarly signal a conceptual duality at work in Making Marvels: on the one hand, the book advances a claim that the creation and use of these objects and instruments generated knowledge, and on the other, it acknowledges that they also served as resplendent signs of princely power and prestige. The catalog is divided into four sections, each containing one or more essays accompanied by catalog entries. In the first section, "Setting the Standard: Forging a Culture of Magnificence," Wolfram Koeppe, the exhibition curator and the catalog's editor, frames this exhibition of technology within the contexts of the "marvelous" and courtly "magnificence." Many of these objects certainly were showpieces that served to glorify their collectors, and many were clearly designed as clever amusements, but as Koeppe and the other contributors to the catalog make clear, this is only one facet of their value, which should also be measured in the advancement of technologies, including metallurgy and metalworking, alchemy, glass-making, horology, mensuration, and many others. The second section, "The Kunstkammer: A Haven of Splendor & Study," is devoted to examining the Kunstkammer (the German term for museums in the sixteenth and seventeenth centuries, often called "curiosity cabinets" in English) as a locus of technology. In his contribution, Dirk Syndram provides a brief history of princely Kunstkammern, particularly those in German-speaking territories. Pamela H. Smith's essay offers some important insights on the lessons to be drawn from the exhibition as a whole; especially salient is her observation that the exigencies of courtly statecraft and display required the amalgamation of knowledge and practice from which modern [End Page 1241] science proceeded. Paulus Rainer takes up the place of foreign rarities in Kunstkammern, correctly noting their perceived utility in advancing knowledge of the world. Florian Thaddäus Bayer's essay is devoted to the Esterházy Kunstkammer, the only early modern collection to survive intact in its original location. The third section, "Princely Education & Entertainment," shifts attention from the collections in general to particular classes of objects, each exemplifying the meeting of pure and applied knowledge that typifies courtly interest in technology. Peter Plassmeyer provides a concise history of the sponsorship of European princely courts as patrons and consumers of useful mathematical instruments such as astrolabes, sundials, and astronomical clocks. Finely crafted artifacts made of ivory turned on the lathe are the topic of Noam Andrews's essay, which he rightly explains as instantiations not just of craft skills but also of applied geometry. Lathework was a particular passion of prince-practitioners, as was alchemy, the subject of Ana Matisse Donefer-Hickie's essay. As she summarizes, the quest for transmuting base metals into gold was certainly a motivation for princes, but alchemy was also embraced for its practicality in producing medicines and other useful chemical products. The catalog culminates with a section on "Clocks & Automata: The Art of Technological Development," which contains a single essay by Koeppe. As the pinnacle of period craft skills and technology, these were Kunstkammer treasures and the stars of Making Marvels. More emphatically than in his introduction, here...

Determinants of Innovative Performance: The Case of an Emerging Country SMEs

Purpose: Malaysia aspires to vie with other developed countries and achieve similar status shortly. Thus, to generate a more knowledge-intensive and innovation-driven economy is crucial for the country. Considering the significance that an Innovation Economy is designed from an amalgamation of knowledge, technology, entrepreneurship, and innovation to hasten productivity, which is the core of economic growth; this study analyzed the relationship of entrepreneurial attitude orientation (EAO), market orientation, and entrepreneurial competencies to competitive intelligence and its impact on the innovative performance of SMEs in Malaysia. Approach/Methodology/Design: This study adopted a cross-sectional research design and collected data from small, medium enterprises from 13 states and two federal territories of Malaysia. Findings: The findings revealed that market orientation has a positive impact on entrepreneurial competencies. It is also found that EAO constructs, in particular conceptual, strategic, and technical competencies, have a significant relationship with competitive intelligence and thereby have an impact on the Malaysian SMEs. Likewise, competitive intelligence positively influences the innovative performance. Practical Implications: Apart from contributing to the body of existing literature on innovative firm performance, the study gives a guideline to the SMEs, government agencies, and universities to create the database and designing their training and development programs to cultivate the sense of innovativeness among Malaysians regardless of age, race, and education background. Originality/Value: This study contributes by supporting that conceptual, strategic, and technical competencies have a significant relationship with competitive intelligence and is achievable to be used by SMEs.

Amalgamating Knowledge towards Comprehensive Classification

With the rapid development of deep learning, there have been an unprecedentedly large number of trained deep network models available online. Reusing such trained models can significantly reduce the cost of training the new models from scratch, if not infeasible at all as the annotations used for the training original networks are often unavailable to public. We propose in this paper to study a new model-reusing task, which we term as knowledge amalgamation. Given multiple trained teacher networks, each of which specializes in a different classification problem, the goal of knowledge amalgamation is to learn a lightweight student model capable of handling the comprehensive classification. We assume no other annotations except the outputs from the teacher models are available, and thus focus on extracting and amalgamating knowledge from the multiple teachers. To this end, we propose a pilot two-step strategy to tackle the knowledge amalgamation task, by learning first the compact feature representations from teachers and then the network parameters in a layer-wise manner so as to build the student model. We apply this approach to four public datasets and obtain very encouraging results: even without any human annotation, the obtained student model is competent to handle the comprehensive classification task and in most cases outperforms the teachers in individual sub-tasks.