Portfolio Transformer for Attention-Based Asset Allocation

Traditional approaches to financial asset allocation start with returns forecasting followed by an optimization stage that decides the optimal asset weights. Any errors made during the forecasting step reduce the accuracy of the asset weightings, and hence the profitability of the overall portfolio. The Portfolio Transformer (PT) network, introduced here, circumvents the need to predict asset returns and instead directly optimizes the Sharpe ratio, a risk-adjusted performance metric widely used in practice. The PT is a novel end-to-end portfolio optimization framework, inspired by the numerous successes of attention mechanisms in natural language processing. With its full encoder-decoder architecture, specialized time encoding layers, and gating components, the PT has a high capacity to learn long-term dependencies among portfolio assets and hence can adapt more quickly to changing market conditions such as the COVID-19 pandemic. To demonstrate its robustness, the PT is compared against other algorithms, including the current LSTM-based state of the art, on three different datasets, with results showing that it offers the best risk-adjusted performance. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Drifts of Collaborative Online International Learning (COIL) Towards Pedagogical Innovation: A Foretelling Bibliometric Analysis

In the current global world, and also due to the CoVID-19 pandemic and the evolving digital technologies, education is changing faster and more dramatically now than at any time in history. In this context, education is drifting away from the traditional approaches and into the integration of international, intercultural and global dimensions in higher education, vocational education and training, as well as in secondary education. Thus, the purpose of this exploratory study is to unveil current trends and future needs of research and practice focused on Collaborative Online International Learning (COIL), to set conceptual foundations for further pedagogical innovation. Following the PRISMA 2020 statement, the corpus of the study integrates a final set of research studies (n = 135), published between 2019–2022. Considering the main key-terms of recent literature in the area, the collected data were analyzed by means of a bibliometric analysis, using VOSviewer's network and overlay visualization of most frequent terms. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Feasibility Study on Browser-Based Federated Machine Learning (FML) Architecture for Medical Application

Traditional approaches to Artificial Intelligence (AI) based medical image classification requires huge amounts of data sets to be stored in a centralized server for analysis and training. In medical applications, data privacy and ownership may pose a challenge. In addition, costs incurred by data transfer and cloud server may pose a challenge to implementing a large dataset. This work studies the feasibility of a decentralized, browser-based Artificial Intelligence (AI) federated machine learning (FML) architecture. The proposed work studies the feasibility of bringing training and inference to the browser, hence removing the need to transfer raw data to a centralized server. If feasible, the system allows practitioners to compress and upload their pre-trained model to the server instead of raw data. This allows medical practitioners to update the model without the need to reveal their raw data. A sandbox system was implemented by applying transfer learning on MobileNet V3 and was tested with chest X-ray image datasets from COVID-19, viral pneumonia, and normal patients to simulate medical usage environment. The training speed, model performance and inference speed were tested on a PC browser and mobile phone with various levels of network throttling and image degradation. © 2022 IEEE.

ARC-NLP at CheckThat! 2022: Contradiction for Harmful Tweet Detection

The target task of our team in CLEF2022 CheckThat! Lab challenge is Task-1C, harmful tweet detection. We propose a novel approach, called ARC-NLP-contra, which is a contradiction check approach by using the idea that harmful tweets contradict with the real-life facts in the scope of COVID-19 pandemic. Besides, we propose and examine two other models. The first model, called ARC-NLP-hc, is a traditional approach that utilizes hand-crafted tweet and user features. The second model, called ARC-NLP-pretrain, pretrains a Transformer-based language model by using COVID-related Turkish tweets. We compare the performances of these three models, and submit the highest performing model in the preliminary experiments to the challenge. We make submissions for Task-1A, 1B, 1C in Turkish and Task-1C in English. We have the winning solution for Task-1C, harmful tweet detection in Turkish, using ARC-NLP-contra that is our contradiction check approach. © 2022 Copyright for this paper by its authors.

IoT-enabled model for Digital Twin of Mental Stress (DTMS)

Stress has become one of the mental health adversaries of the COVID-19 pandemic. Several stressors like fear of infection, lockdown, and social distancing are commonly accountable for the stress. The existing stress prediction systems are less compatible to handle diversly changing stressors during COVID-19. The traditional approaches often use incomplete features from limited sources (e.g., only wearable sensor or user device) and static prediction techniques. The Edge Artificial Intelligence (Edge AI) employs machine learning to make data from these sources usable for decision making. Therefore, In this study, we propose a Digital Twin of Mental Stress (DTMS) model that employs IoT-based multimodal sensing and machine learning for mental stress prediction. We obtained 98% accuracy for four widely used Machine Learning(ML) algorithms Naïve Bayes(NB), Random Forest(RF), Multilayer Perceptron(MLP), and Decision Tree (DT). The optimal Digital Twin Features (DTF) could reduce the classification time. © 2021 IEEE.

Learning by Doing from Afar: Mobilizing the Internet of Things to Enrich Remote Learning

The consequences of the COVID-19 pandemic have introduced acute challenges to traditional approaches of engi-neering education. Without direct and regular access to physical equipment in laboratories, the opportunities for engineering students to practice 'learning by doing' has been greatly reduced or curtailed. This paper presents our attempt to demonstrate a simple, efficient, and effective method of approximating the hands-on experience of manipulating lab equipment by using an 'Internet of Things' (IoT) approach, which involves networked physical objects and the use of embedded software, sensors, and other technologies to facilitate control and exchange of data over the internet. While a few other researchers have also started experimenting with IoT in test labs, our approach differs sharply in some key ways. Instead of building entirely new IoT-integrated equipment for a given lesson, our approach is to adapt traditional lab equipment for remote use by making a number of small changes, which keep things simple to operate and cost-effective. As a result, students can connect to the laboratory facilities remotely and conduct experiments and gather data on their own. In this initial project, the target lab was a lesson on 'controls' consisting of solenoid valves as controllable actuators, high voltage controllable switch, and a series of sensors and web-cameras as monitoring facilities. Using a desktop remote control app, students make a connection to the university laboratory computers. The preliminary results provide a proof of concept for remotely control laboratories for undergraduate students in engineering that allow them to perform experiments in a fixable yet reliable manner. The implications for methods of enhancing remote learning extend beyond pandemic conditions © 2021 IEEE.

Evaluation of Academic Performance of Students Using Fuzzy Logic

The entire education system has undergone numerous changes to stand unhindered during the current COVID-19 pandemic. All over the world, the educational system has changed its teaching and learning methods. One of its important aspects, evaluating the students' overall performance has become a complex task with these changing patterns. The traditional approach of evaluation may not be a best fit anymore since multiple factors are required to make an all-inclusive, multifaceted decision to keep up with the upgrades in evaluation schemes and patterns. Also, Universities and educational institutes understood the importance of skill based learning and major changes are being made in the curriculum, which in turn need cognitive approach to evaluate the students' performance. Hence, we have proposed, designed and implemented a solution, a fuzzy logic-based model. This model, while showing the difference between the traditional approach and the inference system, will enable the educational institutes not only to evaluate a students' performance but also to understand the students in a comprehensive manner. © 2021 IEEE.

Work-in-Progress: Incorporating Computational Thinking Instruction into K-12 Using 3D Weather

The 3D Weather Analysis and Visualization project (3D Weather) improves middle school students' computational thinking skills. 3D Weather focuses on evaluating three-dimensional data using the full data volume instead of the traditional approach using two-dimensional cross-sectional slices. We use basic meteorology as a contextual framework to introduce and provide the data for the computational thinking exercises. However, instead of directly developing lessons for the classroom, we are using a “teach the teacher” model. We provide a two-week summer professional development for middle school teachers in the state. During the program, the teachers are taught introductory weather science, the IDV visualization software basics, and how to obtain free weather data from the National Oceanic and Atmospheric Administration (NOAA). Restrictions due to the COVID19 pandemic required modifications to our planned initial year activities, but we were able to pilot and obtain feedback to improve the program. Our plans for the 2021-2022 school year include offering our full summer professional development workshop, observing teachers in their classrooms while they implement meteorology lessons with computational thinking, and collecting data from both teachers and students measuring attitudes towards meteorology and computational thinking as well as changes in 3D visualization abilities. © American Society for Engineering Education, 2021

Integrative role of traditional and modern technologies to combat COVID-19.

INTRODUCTION: With the development of various branches of sciences, we will be able to resolve different clinical aspects of various diseases better. The convergence of these sciences can potentially tackle the new corona crisis. AREAS COVERED: In this review, we attempted to explore and describe various scientific branches studying COVID-19. We have reviewed the literature focusing on the prevention, diagnosis, and treatment of COVID-19. The primary databases targeted were Science Direct, Scopus and PubMed. The most relevant reports from the recent two decades were collected utilizing keywords including SARS-CoV, MERS-CoV, COVID-19, epidemiology, therapeutics and diagnosis. EXPERT OPINION: Based on this literature review, both traditional and emerging approaches are vital for the prevention, diagnosis and treatment of COVID-19. The traditional sciences play an essential role in the preventive and supportive care of corona infection, and modern technologies appear to be useful in the development of precise diagnosis and powerful treatment approaches for this disease. Indeed, the integration of these sciences will help us to fight COVID-19 disease more efficiently.