Synergistic Learning Research Articles

HirMTL: Hierarchical Multi-Task Learning for dense scene understanding

In the realm of artificial intelligence, simultaneous multi-task learning is crucial, particularly for dense scene understanding. To address this, we introduce HirMTL, a novel hierarchical multi-task learning framework designed to enhance dense scene analysis. HirMTL is adept at facilitating interaction at the scale level, ensuring task-adaptive multi-scale feature fusion, and fostering task-level feature interchange. It leverages the inherent correlations between tasks to create a synergistic learning environment. Initially, HirMTL enables concurrent sharing and fine-tuning of features at the single-scale level. This is further extended by the Task-Adaptive Fusion module (TAF), which intelligently blends features across scales, specifically attuned to each task’s unique requirements. Complementing this, the Asymmetric Information Comparison Module (AICM) skillfully differentiates and processes both shared and unique features, significantly refining task-specific performance with enhanced accuracy. Our extensive experiments on various dense prediction tasks validate HirMTL’s exceptional capabilities, showcasing its superiority over existing multi-task learning models and underscoring the benefits of its hierarchical approach.

Fostering STEM Learning: Exploring the Integration of Design Thinking in Islamic STEM Education

This study investigates the integration of Design Thinking in STEM education within an Islamic context, focusing on thermal packaging as a learning module. The research examines differences in STEM Interest, Design Thinking skills, and Subject Knowledge across 7th and 8th-grade students, as well as the relationships between these constructs. A quantitative approach was employed, involving 64 preparatory students from an institute of religious studies in Qatar. Participants engaged in a Design Thinking-based STEM workshop on thermal packaging, followed by a comprehensive survey assessing the three key constructs. The Independent-Samples Mann-Whitney U Test revealed no significant differences between grade levels for STEM Interest (U = 424.500, p = .434), Design Thinking skills (U = 494.000, p = .843), or Subject Knowledge (U = 469.500, p = .873). However, bivariate correlation analysis showed strong positive relationships among the constructs for both grade levels. For 7th-grade students, STEM Interest and Design Thinking Skills were strongly correlated (r = 0.856), while 8th-grade students showed a robust correlation between Design Thinking Skills and Subject Knowledge (r = 0.884). These findings suggest that integrating Design Thinking approaches in STEM education within an Islamic context can effectively enhance students' interest, skills, and knowledge simultaneously. The study highlights the potential of Design Thinking-based STEM workshops to create a synergistic learning environment that aligns with Islamic principles and values while preparing students for contemporary challenges. The research contributes to the growing body of literature on culturally responsive STEM education and provides insights for educators seeking to bridge traditional Islamic teachings with modern technological advancements.

Interactive attack-defense for generalized person re-identification

Generalized Person Re-Identification (GReID) aims to develop a model capable of robust generalization across unseen target domains, even with training on a limited set of observed domains. Recently, methods based on the Attack-Defense mechanism are emerging as a prevailing technology to this issue, which treats domain transformation as a type of attack and enhances the model’s generalization performance on the target domain by equipping it with a defense module. However, a significant limitation of most existing approaches is their inability to effectively model complex domain transformations, largely due to the separation of attack and defense components. To overcome this limitation, we introduce an innovative Interactive Attack-Defense (IAD) mechanism for GReID. The core of IAD is the interactive learning of two models: an attack model and a defense model. The attack model dynamically generates directional attack information responsive to the current state of the defense model, while the defense model is designed to derive generalizable representations by utilizing a variety of attack samples. The training approach involves a dual process: for the attack model, the aim is to increase the challenge for the defense model in countering the attack; conversely, for the defense model, the focus is on minimizing the effects instigated by the attack model. This interactive framework allows for mutual learning between attack and defense, creating a synergistic learning environment. Our diverse experiments across datasets confirm IAD’s effectiveness, consistently surpassing current state-of-the-art methods, and using MSMT17 as the target domain in different protocols resulted in a notable 13.4% improvement in GReID task average Rank-1 accuracy. Code is available at: https://github.com/lhf12278/IAD.

A Synapse-Threshold Synergistic Learning Approach for Spiking Neural Networks

Spiking neural networks (SNNs) have demonstrated excellent capabilities in various intelligent scenarios. Most existing methods for training SNNs are based on the concept of synaptic plasticity; however, learning in the realistic brain also utilizes intrinsic non-synaptic mechanisms of neurons. The spike threshold of biological neurons is a critical intrinsic neuronal feature that exhibits rich dynamics on a millisecond timescale and has been proposed as an underlying mechanism that facilitates neural information processing. In this study, we develop a novel synergistic learning approach that involves simultaneously training synaptic weights and spike thresholds in SNNs. SNNs trained with synapse-threshold synergistic learning (STL-SNNs) achieve significantly superior performance on various static and neuromorphic datasets than SNNs trained with two degenerated single-learning models. During training, the synergistic learning approach optimizes neural thresholds, providing the network with stable signal transmission via appropriate firing rates. Further analysis indicates that STL-SNNs are robust to noisy data and exhibit low energy consumption for deep network structures. Additionally, the performance of STL-SNN can be further improved by introducing a generalized joint decision framework. Overall, our findings indicate that biologically plausible synergies between synaptic and intrinsic non-synaptic mechanisms may provide a promising approach for developing highly efficient SNN learning methods.

A Synergic Model for STEAM Learning in Sciences (First High School Stage)

The article presents information about what the synergistic learning model is when studying complex systems such as the education system. Information is also presented on what STEAM education is, as a more developed version of STEM education. The results of tasks included in external assessment in natural sciences – biology and targeted education for the first high school stage are presented, meeting the set conditions for two groups of students intensively studying a foreign language and without intensive foreign language study. The sample is representative of the country, but not representative of the regions, and includes 2,682 students. From the test, only individual tasks that correspond to the synergistic model are included, and the length of the tests for the three years is 15 tasks – 13 with a selectable answer and 2 with a free answer. The results of the single trial test show that the tasks have good statistical characteristics. There is a statistically significant difference in student achievement in favor of those with intensive foreign language study.

Evolving Dual-Threshold Bienenstock-Cooper-Munro Learning Rules in Echo State Networks.

The strengthening and the weakening of synaptic strength in existing Bienenstock-Cooper-Munro (BCM) learning rule are determined by a long-term potentiation (LTP) sliding modification threshold and the afferent synaptic activities. However, synaptic long-term depression (LTD) even affects low-active synapses during the induction of synaptic plasticity, which may lead to information loss. Biological experiments have found another LTD threshold that can induce either potentiation or depression or no change, even at the activated synapses. In addition, existing BCM learning rules can only select a set of fixed rule parameters, which is biologically implausible and practically inflexible to learn the structural information of input signals. In this article, an evolved dual-threshold BCM learning rule is proposed to regulate the reservoir internal connection weights of the echo-state-network (ESN), which can contribute to alleviating information loss and enhancing learning performance by introducing different optimal LTD thresholds for different postsynaptic neurons. Our experimental results show that the evolved dual-threshold BCM learning rule can result in the synergistic learning of different plasticity rules, effectively improving the learning performance of an ESN in comparison with existing neural plasticity learning rules and some state-of-the-art ESN variants on three widely used benchmark tasks and the prediction of an esterification process.

A Learning Ecology Perspective of Energy Literacy among Youth: A Case Study from Alabama High Schools

Developing energy literacy and pathways for youth to clean energy careers are vital for sustainable development, aligning with various Sustainable Development Goals (SDGs), from climate action to quality education. Despite the increasing focus on nurturing energy-literate youth, there is a lack of comprehensive insights into how students develop their energy literacy within diverse learning environments. This study addresses the research question of what factors across various learning environments play a significant role in the development of energy literacy among youth. To address the research question, we develop a conceptual framework for understanding the development of energy literacy among high school students based on a learning ecology perspective. Focusing on students’ energy literacy, encompassing information search, attitudes, behaviors, and knowledge of energy, we examined the influence of school-based clean energy program participation, virtual information sources, and interactions with peers and family. Furthermore, we conducted an empirical analysis to identify factors associated with energy literacy among high school students from Alabama high schools. The student sample was intentionally recruited from case study high schools in Alabama that implemented a clean energy education pilot project. This sample includes both program participants and non-participants who completed pre- and post-program surveys, resulting in a total of 189 survey responses. Findings indicate a positive association between the school-based clean energy program and a significantly higher increase in energy-related information searches among program participants compared with non-participants. The virtual learning setting, especially video platforms, was also significantly associated with students’ energy-related information searches and positive attitudes toward energy. Findings also indicate that family relationships and pre-existing STEM interests significantly determine youth energy literacy. These findings provide valuable insights for the development and expansion of future energy education programs. If the energy education program incorporates these crucial learning factors and establishes an interconnected learning environment, the convergence of multiple learning aspects within the program can foster a synergistic learning ecosystem for youth energy literacy.

A high-performance turnkey system for customer lifetime value prediction in retail brands

Customer lifetime value (CLV) modeling underpins modern marketing analytics, enabling the development of tailored customer relationship management strategies based on the predicted future value of their customers. As part of Amperity’s enterprise customer data platform (CDP), we deploy and maintain a CLV prediction system that caters to a rapidly growing list of brands across various industries, purchase behaviors, and scales. Given the impracticality of developing bespoke models for each brand, our solution must be adaptive, generalizable, and high-performing ”out of the box”. Furthermore, our platform demands daily prediction updates to facilitate prompt marketing decisions. This paper introduces a turnkey CLV prediction system that achieves state-of-the-art performance across a diverse set of brands. This system has several contributions: 1) the use of encodings and embeddings to incorporate signals from high-cardinality data; 2) a multi-stage churn-CLV modeling framework that augments additional flexibility in adjusting churn probabilities, subsequently reducing CLV prediction errors while maintaining a synergistic learning process; 3) a feature-weighted ensemble of both generative and discriminative models to accommodate diverse underlying purchase patterns. Empirical results show that our enhanced model consistently surpasses benchmark performances for twelve retail brands across six evaluation intervals from June 2020 to September 2022.

Improving domain generalization performance for medical image segmentation via random feature augmentation.

Deep convolutional neural networks (DCNNs) have shown remarkable performance in medical image segmentation tasks. However, medical images frequently exhibit distribution discrepancies due to variations in scanner vendors, operators, and image quality, which pose significant challenges to the robustness of trained models when applied to unseen clinical data. To address this issue, domain generalization methods have been developed to enhance the generalization ability of DCNNs. Feature space-based data augmentation methods have been proven effective in improving domain generalization, but they often rely on prior knowledge or assumptions, which can limit the diversity of source domain data. In this study, we propose a novel random feature augmentation (RFA) method to diversify source domain data at the feature level without prior knowledge. Specifically, our RFA method perturbs domain-specific information while preserving domain-invariant information, thereby adequately diversifying the source domain data. Furthermore, we propose a dual-branches invariant synergistic learning strategy to capture domain-invariant information from the augmented features of RFA, enabling DCNNs to learn a more generalized representation. We evaluate our proposed method on two challenging medical image segmentation tasks, optic cup/disc segmentation on fundus images and prostate segmentation on MRI images. Extensive experimental results demonstrate the superior performance of our method over state-of-the-art domain generalization methods.

Optimalisasikan Peran Mahasiswa Dalam Pencegahan Stunting Melalui Inovatif Makanan Sehat Gimbab Isi Sayuran Dan Telur

Real work lectures are a synergistic learning program as a form of student service to society. The real work lecture raised the theme of stunting prevention. One of the real work college programs carried out is making innovative food that is healthy and easy for all ages, especially toddlers, to enjoy. An innovative food with basic ingredients of rice, vegetables and eggs is kimbap. Stunting prevention outreach and demonstrations were carried out once at the Village Pendopo targeting toddlers and mothers. The result of this activity is the formation of local kimbap products that toddlers like.

Synergetic learning for unknown nonlinear [formula omitted] control using neural networks

The well-known H∞ control design gives robustness to a controller by rejecting perturbations from the external environment, which is difficult to do for completely unknown affine nonlinear systems. Accordingly, the immediate objective of this paper is to develop an on-line real-time synergetic learning algorithm, so that a data-driven H∞ controller can be received. By converting the H∞ control problem into a two-player zero-sum game, a model-free Hamilton–Jacobi–Isaacs equation (MF-HJIE) is first derived using off-policy reinforcement learning, followed by a proof of equivalence between the MF-HJIE and the conventional HJIE. Next, by applying the temporal difference to the MF-HJIE, a synergetic evolutionary rule with experience replay is designed to learn the optimal value function, the optimal control, and the worst perturbation, that can be performed on-line and in real-time along the system state trajectory. It is proven that the synergistic learning system constructed by the system plant and the evolutionary rule is uniformly ultimately bounded. Finally, simulation results on an F16 aircraft system and a nonlinear system back up the tractability of the proposed method.

Turning the Wheel through the Ages

Methods for reliably generating alternating current via placing metal coils in a rotating magnetic field were discovered by two individuals almost simultaneously: Italian engineer and physicist, Galileo Ferraris, and Austrian-American engineer, Nikola Tesla. While Tesla attained the patent for the AC-generator, Ferraris, who published the discovery a few months prior, is credited with having paved the way for the modern-day alternator, now used ubiquitously.The origins of rotational energy conversion can be traced back to sugar plantations in the 1600s. At that time, the milling process was initially sustained through slave-driven turning of a wheel, which was geared to large wooden rollers that would squeeze sugar cane. The Caribbean and South American variant of this system was known as a trapiche, which employed three geared rollers. Over time, slaves were replaced with oxen and horses, which were eventually supplanted by water wheels and finally by steam engines. The water wheels implemented for sugar milling ultimately paved the path for creating electrical currents. Water turbines for hydroelectric power subsequently took off. Principles of steam engine operation were adapted for electrical power as well, ultimately culminating in the use of nuclear plants to generate “wheel” turning for electricity.Electricity and magnetism, as well as interactions between them, are frequently discussed in introductory undergraduate physics. However, practical applications of these principles are often omitted from lectures. If taught under an experiential environment, these concepts can be introduced at even the middle school level, taught by first year college students taking basic college physics - a synergistic learning environment. With this in mind, we developed a middle school kit consisting of simple 3D-printed components, recycled DVDs and four bearings for support of 4 mm nail shafts. Enameled wire (30 AWG) is used to prepare the wire-wound coils. A large rubber band is used as the pulley-belt. The above represents the generator component. The power use component is assembled on an accompanying bread board. Power processing is demonstrated using diodes for rectification to DC. The DC power is used to recharge a 3V, 40 mAh lithium-ion coin cell. AC power from the generator can also be directly applied to load resistors. Students will be able to see the effects of varying the load by changes in the time of the free rotation of the generator wheels. Students will experience the charge and discharge of batteries with safe voltages and currents.Middle school students will record data from the generator and conditioning systems and draw conclusions on worksheets that accompany the hardware.

Educational Robotics: Development of computational thinking in collaborative online learning.

In the previous study the work experience on organization of teaching Robotics to secondary school students at school lessons and in study groups was introduced. This study which was conducted within 2019 and 2021 covered the period of distant learning caused by COVID-19 pandemic and even post-pandemic period, when a part of school students continued learning online. The study deals with the problem of developing school students' computational thinking in online learning. We consider computational thinking as a set of cognitive skills of solving educational and cognitive problems. The research questions raised were aimed at solving the problem of the influence of Educational Robotics on developing computational thinking. During the research we have found out that due to the adaptability of robots, Educational Robotics, the development of individual learning programs, and the arrangement of collaborative online learning are instruments and a solution to the problem of developing computational thinking. The main components of computational thinking, which were studied within those 3 years, are the following: algorithmic thinking, ability to program, and efficiency in team work. The influence of the learning strategy we chose enabled us to determine the level of computational thinking and its dependence on learning Robotics. We used statistical criteria in order to summarize the results of our research. The statistics provided suggests progress in the indicator tracked. Based on the experimental data received we approximated reliability (R2) and relevant exponential equation (trend lines). The research we carried out also has led to the general conclusion that Educational Robotics helps to create synergistic learning environment for stimulating students' motivation, collaboration, self-efficacy and creativity.

Investigating the Association between Algorithmic Thinking and Performance in Environmental Study

Presently, computational thinking (CT) is considered necessary for adapting to the future. Concurrently, the COVID-19 pandemic has accentuated the demand for strengthening Environmental Education as a means to improve sustainability and stimulate environmental protection and public health. Having in mind that CT does not concern only technocrats but also applies in solving everyday problems, we introduce the novel idea of the synergistic learning of CT and Environmental Study. Thus, our research aim is to explore the correlation between algorithmic thinking (AT), as a fundamental CT competency, and educational achievements in the Environmental Study course during the early primary school years. Towards this end, we implemented a quantitative research study, employing an innovative assessment framework we propose. The adoption of cluster sampling eventuated in a sample of 435 students. The exploitation of ordinal logistic regression analysis and machine learning method validated the correlation of the two fields and pointed out that AT levels constitute a predictive factor for performance in the Environmental Study course and vice versa. These results support the novel idea of concurrently cultivating environmental consciousness and CT and build a robust base for future studies that will focus on providing an ecological reflection on CT activities.

Learning Decomposed Representations for Treatment Effect Estimation

In observational studies, confounder separation and balancing are the fundamental problems of treatment effect estimation. Most of the previous methods focused on addressing the problem of confounder balancing by treating all observed pre-treatment variables as confounders, ignoring confounder separation. In general, not all the observed pre-treatment variables are confounders that refer to the common causes of the treatment and the outcome, some variables only contribute to the treatment (i.e., instrumental variables) and some only contribute to the outcome (i.e., adjustment variables). Balancing those non-confounders, including instrumental variables and adjustment variables, would generate additional bias for treatment effect estimation. By modeling the different causal relations among observed pre-treatment variables, treatment variables and outcome variables, we propose a synergistic learning framework to i) separate confounders by learning decomposed representations of both confounders and non-confounders, ii) balance confounder with sample re-weighting technique, and simultaneously iii) estimate the treatment effect in observational studies via counterfactual inference. Empirical results on synthetic and real-world datasets demonstrate that the proposed method can precisely decompose confounders and achieve a more precise estimation of treatment effect than baselines.

On Differentiation Strategies in the EFL Mixed-Ability Classroom: Towards Promoting the Synergistic Learning Environment

On Differentiation Strategies in the EFL Mixed-Ability Classroom: Towards Promoting the Synergistic Learning Environment

Perceptual learning as a result of concerted changes in prefrontal and visual cortex

Our perceptual ability remarkably improves with training. Some studies on visual perceptual learning have shown refined neural representation of the trained stimulus in the visual cortex, whereas others have exclusively argued for improved readout and decision-making processes in the frontoparietal cortex. The mixed results have rendered the underlying neural mechanisms puzzling and hotly debated. By simultaneously recording from monkey visual area V4 and ventrolateral prefrontal cortex (PFC) implanted with microelectrode arrays, we dissected learning-induced cortical changes over the course of training the monkeys in a global form detection task. Decoding analysis dissociated two distinct components of neuronal population codes that were progressively and markedly enhanced in both V4 and PFC. One component was closely related to the target stimulus feature and was subject to task-dependent top-down modulation; it emerged earlier in V4 than PFC and its enhancement was specific to the trained configuration of the target stimulus. The other component of the neural code was entirely related to the animal's behavioral choice; it emerged earlier in PFC than V4 and its enhancement completely generalized to an untrained stimulus configuration. These results implicate two concurrent and synergistic learning processes: a perceptual process that is specific to the details of the trained stimulus feature and a cognitive process that is dependent on the total amount of learning experience in the trained task. When combined, these two learning processes were well predictive of the animal's learning behavior.

Preface

We are pleased to declare that we had really successful three days of 2021 the 5th Interntional Conference on Sustainable Energy Engineering (ICSEE 2021), which supposed to be held in Singapore on February 26-28, 2021, but was changed to be held online in Zoom due to the COVID-19 pandemic. The conference supported by Shanghai Jiatong University, China and Pusan National University, Korea. Again, the online conference was of great success, totally beyond anybody's expectation, having a diverse group of expertise and backgrounds. Not to menition that participants are geographically diverse. We are glad to confirm that such diverse expertise and backgrounds were connected, forming a coherent conference environment and creating synergistic learning and sharing.List of Conference Committees, Advisory Committees, Conference General Chair, Conference Program Chairs, Conference Program Co-chairs, Steering Chair, Conference Technical Committees are available in this pdf.

Preface

Welcome to the Proceedings of the 3rd International Conference on Environmental Prevention and Pollution Control Technologies (EPPCT 2021), which was supposed to be held in Zhuhai, China during January 15 to 17, 2021, but was changed to be held online due to the COVID-19 pandemic. The online conference was of great success, totally beyond anybody’s expectation, having a diverse group of expertise and backgrounds. Not to mention that participants are geographically diverse. We are glad to confirm that such diverse expertise and backgrounds were connected, forming a coherent conference environment and creating synergistic learning and sharing.The meeting focused on the research fields of “environmental prevention” and “environmental pollution prevention and control”. The primary goal of the conference is to provide an exchange platform for experts and scholars in “environmental prevention” and “pollution prevention” as well as enterprise developers, to promote the exchange of new technologies for environmental pollution prevention and control at home and abroad, to explore the development path of ecological civilization construction, and to cope with new opportunities and challenges.This scientific event brings together more than 160 national and international researchers. During the conference, the conference model was divided into three sessions, including oral presentations, keynote speeches, and online Q&A discussion. In the first part, some scholars, whose submissions were selected as the excellent papers, were given about 5-10 minutes to perform their oral presentations one by one. Then in the second part, keynote speakers were each allocated 30-45 minutes to hold their speeches.In the second part, we invited two professors as our keynote speakers. Prof. Shanquan Wang, from School of Environmental Science and Engineering, Sun Yat-Sen University, performed a speech: Spatial distribution, bioconversion and ecological risk of halogenated persistent organic pollutants in contaminated urban rivers: a nationwide study in China. His study provided unprecedented insights into contamination, conversion and ecological risk of PCBs and PBDEs in nationwide polluted urban rivers in China. And then we had Prof. Tao Zhu from China University of Mining & Technology (Beijing). His research area is air pollution control and solid waste treatment and resource. Their insightful speeches had triggered heated discussion in the third session of the conference. Every participant praised this conference for disseminating useful and insightful knowledge.The proceedings are a compilation of the accepted papers and represent an interesting outcome of the conference. Topics include but are not limited to the following areas: Air pollution and treatment, Groundwater Issues, Soil Pollution and Treatment and more related topics. All the papers have been through rigorous review and process to meet the requirements of International publication standard.We would like to acknowledge all of those who supported EPPCT 2021. The help and contribution of each individual and institution was instrumental in the success of the conference. In particular, we would like to thank the organizing committee for its valuable inputs in shaping the conference program and reviewing the submitted papers.Committee of EPPCT 2021List of Committee member are available in this Pdf.

Synergistic learning of lung lobe segmentation and hierarchical multi-instance classification for automated severity assessment of COVID-19 in CT images

Understanding chest CT imaging of the coronavirus disease 2019 (COVID-19) will help detect infections early and assess the disease progression. Especially, automated severity assessment of COVID-19 in CT images plays an essential role in identifying cases that are in great need of intensive clinical care. However, it is often challenging to accurately assess the severity of this disease in CT images, due to variable infection regions in the lungs, similar imaging biomarkers, and large inter-case variations. To this end, we propose a synergistic learning framework for automated severity assessment of COVID-19 in 3D CT images, by jointly performing lung lobe segmentation and multi-instance classification. Considering that only a few infection regions in a CT image are related to the severity assessment, we first represent each input image by a bag that contains a set of 2D image patches (with each cropped from a specific slice). A multi-task multi-instance deep network (called M2UNet) is then developed to assess the severity of COVID-19 patients and also segment the lung lobe simultaneously. Our M2UNet consists of a patch-level encoder, a segmentation sub-network for lung lobe segmentation, and a classification sub-network for severity assessment (with a unique hierarchical multi-instance learning strategy). Here, the context information provided by segmentation can be implicitly employed to improve the performance of severity assessment. Extensive experiments were performed on a real COVID-19 CT image dataset consisting of 666 chest CT images, with results suggesting the effectiveness of our proposed method compared to several state-of-the-art methods.