Showing 1–50 of 152 results for author: Islam, M A

A Dataset for Research on Water Sustainability

Authors: Pranjol Sen Gupta, Md Rajib Hossen, Pengfei Li, Shaolei Ren, Mohammad A. Islam

Abstract: Freshwater scarcity is a global problem that requires collective efforts across all industry sectors. Nevertheless, a lack of access to operational water footprint data bars many applications from exploring optimization opportunities hidden within the temporal and spatial variations. To break this barrier into research in water sustainability, we build a dataset for operation direct water usage in… ▽ More Freshwater scarcity is a global problem that requires collective efforts across all industry sectors. Nevertheless, a lack of access to operational water footprint data bars many applications from exploring optimization opportunities hidden within the temporal and spatial variations. To break this barrier into research in water sustainability, we build a dataset for operation direct water usage in the cooling systems and indirect water embedded in electricity generation. Our dataset consists of the hourly water efficiency of major U.S. cities and states from 2019 to 2023. We also offer cooling system models that capture the impact of weather on water efficiency. We present a preliminary analysis of our dataset and discuss three potential applications that can benefit from it. Our dataset is publicly available at Open Science Framework (OSF) △ Less

Submitted 23 May, 2024; originally announced May 2024.

Comments: Accepted by ACM e-Energy 2024

MapCoder: Multi-Agent Code Generation for Competitive Problem Solving

Authors: Md. Ashraful Islam, Mohammed Eunus Ali, Md Rizwan Parvez

Abstract: Code synthesis, which requires a deep understanding of complex natural language problem descriptions, generation of code instructions for complex algorithms and data structures, and the successful execution of comprehensive unit tests, presents a significant challenge. While large language models (LLMs) demonstrate impressive proficiency in natural language processing, their performance in code ge… ▽ More Code synthesis, which requires a deep understanding of complex natural language problem descriptions, generation of code instructions for complex algorithms and data structures, and the successful execution of comprehensive unit tests, presents a significant challenge. While large language models (LLMs) demonstrate impressive proficiency in natural language processing, their performance in code generation tasks remains limited. In this paper, we introduce a new approach to code generation tasks leveraging multi-agent prompting that uniquely replicates the full cycle of program synthesis as observed in human developers. Our framework, MapCoder, consists of four LLM agents specifically designed to emulate the stages of this cycle: recalling relevant examples, planning, code generation, and debugging. After conducting thorough experiments, with multiple LLM ablations and analyses across eight challenging competitive problem-solving and program synthesis benchmarks, MapCoder showcases remarkable code generation capabilities, achieving new state-of-the-art results (pass@1) on HumanEval (93.9%), MBPP (83.1%), APPS (22.0%), CodeContests (28.5%), and xCodeEval (45.3%). Moreover, our method consistently delivers superior performance across various programming languages and varying problem difficulties. We open-source our framework at https://github.com/Md-Ashraful-Pramanik/MapCoder. △ Less

Submitted 18 May, 2024; originally announced May 2024.

GenFighter: A Generative and Evolutive Textual Attack Removal

Authors: Md Athikul Islam, Edoardo Serra, Sushil Jajodia

Abstract: Adversarial attacks pose significant challenges to deep neural networks (DNNs) such as Transformer models in natural language processing (NLP). This paper introduces a novel defense strategy, called GenFighter, which enhances adversarial robustness by learning and reasoning on the training classification distribution. GenFighter identifies potentially malicious instances deviating from the distrib… ▽ More Adversarial attacks pose significant challenges to deep neural networks (DNNs) such as Transformer models in natural language processing (NLP). This paper introduces a novel defense strategy, called GenFighter, which enhances adversarial robustness by learning and reasoning on the training classification distribution. GenFighter identifies potentially malicious instances deviating from the distribution, transforms them into semantically equivalent instances aligned with the training data, and employs ensemble techniques for a unified and robust response. By conducting extensive experiments, we show that GenFighter outperforms state-of-the-art defenses in accuracy under attack and attack success rate metrics. Additionally, it requires a high number of queries per attack, making the attack more challenging in real scenarios. The ablation study shows that our approach integrates transfer learning, a generative/evolutive procedure, and an ensemble method, providing an effective defense against NLP adversarial attacks. △ Less

Submitted 17 April, 2024; originally announced April 2024.

Towards Realistic Few-Shot Relation Extraction: A New Meta Dataset and Evaluation

Authors: Fahmida Alam, Md Asiful Islam, Robert Vacareanu, Mihai Surdeanu

Abstract: We introduce a meta dataset for few-shot relation extraction, which includes two datasets derived from existing supervised relation extraction datasets NYT29 (Takanobu et al., 2019; Nayak and Ng, 2020) and WIKIDATA (Sorokin and Gurevych, 2017) as well as a few-shot form of the TACRED dataset (Sabo et al., 2021). Importantly, all these few-shot datasets were generated under realistic assumptions su… ▽ More We introduce a meta dataset for few-shot relation extraction, which includes two datasets derived from existing supervised relation extraction datasets NYT29 (Takanobu et al., 2019; Nayak and Ng, 2020) and WIKIDATA (Sorokin and Gurevych, 2017) as well as a few-shot form of the TACRED dataset (Sabo et al., 2021). Importantly, all these few-shot datasets were generated under realistic assumptions such as: the test relations are different from any relations a model might have seen before, limited training data, and a preponderance of candidate relation mentions that do not correspond to any of the relations of interest. Using this large resource, we conduct a comprehensive evaluation of six recent few-shot relation extraction methods, and observe that no method comes out as a clear winner. Further, the overall performance on this task is low, indicating substantial need for future research. We release all versions of the data, i.e., both supervised and few-shot, for future research. △ Less

Submitted 5 April, 2024; originally announced April 2024.

Best of Both Worlds: A Pliable and Generalizable Neuro-Symbolic Approach for Relation Classification

Authors: Robert Vacareanu, Fahmida Alam, Md Asiful Islam, Haris Riaz, Mihai Surdeanu

Abstract: This paper introduces a novel neuro-symbolic architecture for relation classification (RC) that combines rule-based methods with contemporary deep learning techniques. This approach capitalizes on the strengths of both paradigms: the adaptability of rule-based systems and the generalization power of neural networks. Our architecture consists of two components: a declarative rule-based model for tr… ▽ More This paper introduces a novel neuro-symbolic architecture for relation classification (RC) that combines rule-based methods with contemporary deep learning techniques. This approach capitalizes on the strengths of both paradigms: the adaptability of rule-based systems and the generalization power of neural networks. Our architecture consists of two components: a declarative rule-based model for transparent classification and a neural component to enhance rule generalizability through semantic text matching. Notably, our semantic matcher is trained in an unsupervised domain-agnostic way, solely with synthetic data. Further, these components are loosely coupled, allowing for rule modifications without retraining the semantic matcher. In our evaluation, we focused on two few-shot relation classification datasets: Few-Shot TACRED and a Few-Shot version of NYT29. We show that our proposed method outperforms previous state-of-the-art models in three out of four settings, despite not seeing any human-annotated training data. Further, we show that our approach remains modular and pliable, i.e., the corresponding rules can be locally modified to improve the overall model. Human interventions to the rules for the TACRED relation \texttt{org:parents} boost the performance on that relation by as much as 26\% relative improvement, without negatively impacting the other relations, and without retraining the semantic matching component. △ Less

Submitted 5 March, 2024; originally announced March 2024.

Multi-Target Two-way Integrated Sensing and Communications with Full Duplex MIMO Radios

Authors: Muhammad Talha, Besma Smida, Md Atiqul Islam, George C. Alexandropoulos

Abstract: In this paper, we propose a multiple input multiple output (MIMO) Full-Duplex Integrated Sensing and Communication System consisting of multiple targets, a single downlink, and a single uplink user. We employed signal-to-interference plus noise ratio (SINR) as the performance metric for radar, downlink, and uplink communication. We use a communication-centric approach in which communication wavefo… ▽ More In this paper, we propose a multiple input multiple output (MIMO) Full-Duplex Integrated Sensing and Communication System consisting of multiple targets, a single downlink, and a single uplink user. We employed signal-to-interference plus noise ratio (SINR) as the performance metric for radar, downlink, and uplink communication. We use a communication-centric approach in which communication waveform is used for both communication and sensing of the environment. We develop a sensing algorithm capable of estimating the direction of arrival (DoA), range, and velocity of each target. We also propose a joint optimization framework for designing A/D transmit and receive beamformers to improve radar, downlink, and uplink SINRs while minimizing self-interference (SI) leakage. We also propose a null space projection (NSP) based approach to improve the uplink rate. Our simulation results, considering orthogonal frequency division multiplexing (OFDM) waveform, show accurate radar parameter estimation with improved downlink and uplink rate. △ Less

Submitted 16 December, 2023; originally announced December 2023.

Comments: 7 pages, 6 figures

Hy-Tracker: A Novel Framework for Enhancing Efficiency and Accuracy of Object Tracking in Hyperspectral Videos

Authors: Mohammad Aminul Islam, Wangzhi Xing, Jun Zhou, Yongsheng Gao, Kuldip K. Paliwal

Abstract: Hyperspectral object tracking has recently emerged as a topic of great interest in the remote sensing community. The hyperspectral image, with its many bands, provides a rich source of material information of an object that can be effectively used for object tracking. While most hyperspectral trackers are based on detection-based techniques, no one has yet attempted to employ YOLO for detecting an… ▽ More Hyperspectral object tracking has recently emerged as a topic of great interest in the remote sensing community. The hyperspectral image, with its many bands, provides a rich source of material information of an object that can be effectively used for object tracking. While most hyperspectral trackers are based on detection-based techniques, no one has yet attempted to employ YOLO for detecting and tracking the object. This is due to the presence of multiple spectral bands, the scarcity of annotated hyperspectral videos, and YOLO's performance limitation in managing occlusions, and distinguishing object in cluttered backgrounds. Therefore, in this paper, we propose a novel framework called Hy-Tracker, which aims to bridge the gap between hyperspectral data and state-of-the-art object detection methods to leverage the strengths of YOLOv7 for object tracking in hyperspectral videos. Hy-Tracker not only introduces YOLOv7 but also innovatively incorporates a refined tracking module on top of YOLOv7. The tracker refines the initial detections produced by YOLOv7, leading to improved object-tracking performance. Furthermore, we incorporate Kalman-Filter into the tracker, which addresses the challenges posed by scale variation and occlusion. The experimental results on hyperspectral benchmark datasets demonstrate the effectiveness of Hy-Tracker in accurately tracking objects across frames. △ Less

Submitted 29 November, 2023; originally announced November 2023.

Unveiling the Potential of Big Data Analytics for Transforming Higher Education in Bangladesh; Needs, Prospects, and Challenges

Authors: Sabbir Ahmed Chowdhury, Md Aminul Islam, Mostafa Azad Kamal

Abstract: Big Data Analytics has gained tremendous momentum in many sectors worldwide. Big Data has substantial influence in the field of Learning Analytics that may allow academic institutions to better understand the learners needs and proactively address them. Hence, it is essential to understand Big Data and its application. With the capability of Big Data to find a broad understanding of the scientific… ▽ More Big Data Analytics has gained tremendous momentum in many sectors worldwide. Big Data has substantial influence in the field of Learning Analytics that may allow academic institutions to better understand the learners needs and proactively address them. Hence, it is essential to understand Big Data and its application. With the capability of Big Data to find a broad understanding of the scientific decision making process, Big Data Analytics (BDA) can be a piece of the answer to accomplishing Bangladesh Higher Education (BHE) objectives. This paper reviews the capacity of BDA, considers possible applications in BHE, gives an insight into how to improve the quality of education or uncover additional values from the data generated by educational institutions, and lastly, identifies needs and difficulties, opportunities, and some frameworks to probable implications about the BDA in BHE sector. Keywords; Big Data Analytics, Learning Analytics, Quality of Education, Challenges, Higher Education, Bangladesh △ Less

Submitted 24 November, 2023; v1 submitted 10 October, 2023; originally announced November 2023.

Unleashing Modified Deep Learning Models in Efficient COVID19 Detection

Authors: Md Aminul Islam, Shabbir Ahmed Shuvo, Mohammad Abu Tareq Rony, M Raihan, Md Abu Sufian

Abstract: The COVID19 pandemic, a unique and devastating respiratory disease outbreak, has affected global populations as the disease spreads rapidly. Recent Deep Learning breakthroughs may improve COVID19 prediction and forecasting as a tool of precise and fast detection, however, current methods are still being examined to achieve higher accuracy and precision. This study analyzed the collection contained… ▽ More The COVID19 pandemic, a unique and devastating respiratory disease outbreak, has affected global populations as the disease spreads rapidly. Recent Deep Learning breakthroughs may improve COVID19 prediction and forecasting as a tool of precise and fast detection, however, current methods are still being examined to achieve higher accuracy and precision. This study analyzed the collection contained 8055 CT image samples, 5427 of which were COVID cases and 2628 non COVID. The 9544 Xray samples included 4044 COVID patients and 5500 non COVID cases. The most accurate models are MobileNet V3 (97.872 percent), DenseNet201 (97.567 percent), and GoogleNet Inception V1 (97.643 percent). High accuracy indicates that these models can make many accurate predictions, as well as others, are also high for MobileNetV3 and DenseNet201. An extensive evaluation using accuracy, precision, and recall allows a comprehensive comparison to improve predictive models by combining loss optimization with scalable batch normalization in this study. Our analysis shows that these tactics improve model performance and resilience for advancing COVID19 prediction and detection and shows how Deep Learning can improve disease handling. The methods we suggest would strengthen healthcare systems, policymakers, and researchers to make educated decisions to reduce COVID19 and other contagious diseases. CCS CONCEPTS Covid,Deep Learning, Image Processing KEYWORDS Covid, Deep Learning, DenseNet201, MobileNet, ResNet, DenseNet, GoogleNet, Image Processing, Disease Detection. △ Less

Submitted 21 October, 2023; originally announced October 2023.

Effect of Triangular Pre-Cracks on the Mechanical Behavior of 2D MoTe$_2$: A Molecular Dynamics Study

Authors: Md. Jobayer Aziz, Md Akibul Islam, Md. Rezwanul Karim, Arafat Ahmed Bhuiyan

Abstract: Among two-dimensional (2D) materials, transition metal dichalcogenides (TMDs) stand out for their remarkable electronic, optical, and chemical properties. In addition to being variable bandgap semiconductor materials, the atomic thinness provides flexibility to TMDs. Therefore, understanding the physical properties of TMDs for applications in flexible and wearable devices is crucial. Despite the g… ▽ More Among two-dimensional (2D) materials, transition metal dichalcogenides (TMDs) stand out for their remarkable electronic, optical, and chemical properties. In addition to being variable bandgap semiconductor materials, the atomic thinness provides flexibility to TMDs. Therefore, understanding the physical properties of TMDs for applications in flexible and wearable devices is crucial. Despite the growing enthusiasm surrounding two-dimensional transition metal dichalcogenides (TMDs), our understanding of the mechanical characteristics of molybdenum ditelluride (MoTe$_2$) remains limited. The mechanical properties of MoTe$_2$ deteriorate in the presence of pre-existing cracks or vacancy defects, which are very common in grown TMDs. In this study, the fracture properties and crack propagation of monolayer molybdenum ditelluride (MoTe$_2$) sheets containing pre-existing triangular cracks with various vertex angles are investigated by performing molecular dynamics (MD) simulations of uniaxial and biaxial tensile loading. Due to pre-crack length, angle, and perimeter variations, monolayer MoTe$_2$ with pre-existing cracks underwent considerable changes in Young's modulus, tensile strength, fracture toughness, and fracture strain values. We have found that the pre-cracked MoTe$_2$ is more brittle than its pristine counterpart. Regulated alteration of pre-crack angle under constant simulation conditions improved the uniaxial mechanical properties. Similarly, regulated alteration of the perimeter of the pre-crack resulted in improved biaxial mechanical properties. This study contributes to the foundational knowledge for advanced design strategies involving strain engineering in MoTe$_2$ and other similar transition metal dichalcogenides. △ Less

Submitted 2 October, 2023; originally announced October 2023.

Comments: Submitted to a peer-reviewed journal, 55 pages, 21 figures

Autonomous Vehicles an overview on system, cyber security, risks, issues, and a way forward

Authors: Md Aminul Islam, Sarah Alqahtani

Abstract: This chapter explores the complex realm of autonomous cars, analyzing their fundamental components and operational characteristics. The initial phase of the discussion is elucidating the internal mechanics of these automobiles, encompassing the crucial involvement of sensors, artificial intelligence (AI) identification systems, control mechanisms, and their integration with cloud-based servers wit… ▽ More This chapter explores the complex realm of autonomous cars, analyzing their fundamental components and operational characteristics. The initial phase of the discussion is elucidating the internal mechanics of these automobiles, encompassing the crucial involvement of sensors, artificial intelligence (AI) identification systems, control mechanisms, and their integration with cloud-based servers within the framework of the Internet of Things (IoT). It delves into practical implementations of autonomous cars, emphasizing their utilization in forecasting traffic patterns and transforming the dynamics of transportation. The text also explores the topic of Robotic Process Automation (RPA), illustrating the impact of autonomous cars on different businesses through the automation of tasks. The primary focus of this investigation lies in the realm of cybersecurity, specifically in the context of autonomous vehicles. A comprehensive analysis will be conducted to explore various risk management solutions aimed at protecting these vehicles from potential threats including ethical, environmental, legal, professional, and social dimensions, offering a comprehensive perspective on their societal implications. A strategic plan for addressing the challenges and proposing strategies for effectively traversing the complex terrain of autonomous car systems, cybersecurity, hazards, and other concerns are some resources for acquiring an understanding of the intricate realm of autonomous cars and their ramifications in contemporary society, supported by a comprehensive compilation of resources for additional investigation. Keywords: RPA, Cyber Security, AV, Risk, Smart Cars △ Less

Submitted 25 September, 2023; originally announced September 2023.

Comparative study of Deep Learning Models for Binary Classification on Combined Pulmonary Chest X-ray Dataset

Authors: Shabbir Ahmed Shuvo, Md Aminul Islam, Md. Mozammel Hoque, Rejwan Bin Sulaiman

Abstract: CNN-based deep learning models for disease detection have become popular recently. We compared the binary classification performance of eight prominent deep learning models: DenseNet 121, DenseNet 169, DenseNet 201, EffecientNet b0, EffecientNet lite4, GoogleNet, MobileNet, and ResNet18 for their binary classification performance on combined Pulmonary Chest Xrays dataset. Despite the widespread ap… ▽ More CNN-based deep learning models for disease detection have become popular recently. We compared the binary classification performance of eight prominent deep learning models: DenseNet 121, DenseNet 169, DenseNet 201, EffecientNet b0, EffecientNet lite4, GoogleNet, MobileNet, and ResNet18 for their binary classification performance on combined Pulmonary Chest Xrays dataset. Despite the widespread application in different fields in medical images, there remains a knowledge gap in determining their relative performance when applied to the same dataset, a gap this study aimed to address. The dataset combined Shenzhen, China (CH) and Montgomery, USA (MC) data. We trained our model for binary classification, calculated different parameters of the mentioned models, and compared them. The models were trained to keep in mind all following the same training parameters to maintain a controlled comparison environment. End of the study, we found a distinct difference in performance among the other models when applied to the pulmonary chest Xray image dataset, where DenseNet169 performed with 89.38 percent and MobileNet with 92.2 percent precision. Keywords: Pulmonary, Deep Learning, Tuberculosis, Disease detection, Xray △ Less

Submitted 3 October, 2023; v1 submitted 16 September, 2023; originally announced September 2023.

A mathematical model for understanding and controlling monkeypox transmission dynamics in the United States and its implications for future epidemic management

Authors: Md. Azmir Ibne Islam, M H M Mubassir, Arindam Kumar Paul, Sharmin Sultana Shanta

Abstract: Although the outbreak of human monkeypox (Mpox) caused by the monkeypox virus (MPXV) has slowed down around the world, little is known about this epidemic-like disease. To identify and re-examine the underlying pattern of the disease through a modified logistic growth model, Mpox data set of the United States from 10 May 2022 to 31 December 2022 was used in this study. The main focus is on the two… ▽ More Although the outbreak of human monkeypox (Mpox) caused by the monkeypox virus (MPXV) has slowed down around the world, little is known about this epidemic-like disease. To identify and re-examine the underlying pattern of the disease through a modified logistic growth model, Mpox data set of the United States from 10 May 2022 to 31 December 2022 was used in this study. The main focus is on the two non-pharmaceutical interventions (policies for reducing human-to-human, and animal-to-human transmissions) which were applied to understand their significance on the epidemic. The interventions are used as control parameters in the model with a view to analyzing the strengths of such controls in minimizing the infected cases. The model reveals a complying acceptance to the United States data. The findings disclose that preventive measures could play important roles in controlling the deadly spread of the transmission in the year of 2022. During the transmission period, better outcomes could have been possible to achieve in the US if both controls were brought to action simultaneously. Our model reflects that to prevent the outbreak of Mpox and/or any similar diseases from a community in future, the continuous application of the preventive strategies displayed through the model might be an effective tool. Moreover, such strategies could play supporting roles during pre-and/or post-vaccination periods. △ Less

Submitted 2 September, 2023; originally announced September 2023.

Comments: 11 pages, 6 figures

Parametric Investigation on Different Bone Densities to avoid Thermal Necrosis during Bone Drilling Process

Authors: Md Ashequl Islam, Nur Saifullah Kamarrudin, M. F. F. Suhaimi, Ruslizam Daud, Ishak Ibrahim, Fauziah Mat

Abstract: Bone drilling is a universal surgical procedure commonly used for internal fracture fixation, implant placement, or reconstructive surgery in orthopedics and dentistry. The increased temperature during such treatment increases the risk of thermal penetration of the bone, which may delay healing or compromise the fixation's integrity. Thus, avoiding penetration during bone drilling is critical to e… ▽ More Bone drilling is a universal surgical procedure commonly used for internal fracture fixation, implant placement, or reconstructive surgery in orthopedics and dentistry. The increased temperature during such treatment increases the risk of thermal penetration of the bone, which may delay healing or compromise the fixation's integrity. Thus, avoiding penetration during bone drilling is critical to ensuring the implant's stability, which needs surgical drills with an optimized design. Bovine femur and mandible bones are chosen as the work material since human bones are not available and are the closest animal bone to human bone in terms of properties. In the present study, the Taguchi fractional factorial approach was used to determine the best design of surgical drills by comparing the drilling properties (i.e., signal-to-noise ratio and temperature rise). The control factors (spindle speed, drill bit diameter, drill site depth, and their levels) were arranged in an L9 orthogonal array. Drilling experiments were done using nine experimental drills with three repetitions. The findings of this study indicate that the ideal values of the surgical drill's three parameters combination (S1D1Di2) and their percentage contribution depend on the parameters' drilling levels. However, the result shows that the spindle speed has the highest temperature effect among other parameters in both (femur and mandible) bones. △ Less

Submitted 24 August, 2023; originally announced August 2023.

Temperature measurement methods in an experimental setup during bone drilling: A brief review on the comparison of thermocouple and infrared thermography

Authors: Md Ashequl Islam, Nur Saifullah Kamarrudin, Ruslizam Daud, Ishak Ibrahim, Anas Rahman, Fauziah Mat

Abstract: Predicting thermal response in orthopaedic surgery or dental implantation remains a significant challenge. This study aims to find a practical approach for measuring temperature elevation during a bone drilling experiment by analyzing the existing methods. Traditionally thermocouple has frequently been used to predict the bone temperature in the drilling process. However, several experimental stud… ▽ More Predicting thermal response in orthopaedic surgery or dental implantation remains a significant challenge. This study aims to find a practical approach for measuring temperature elevation during a bone drilling experiment by analyzing the existing methods. Traditionally thermocouple has frequently been used to predict the bone temperature in the drilling process. However, several experimental studies demonstrate that the invasive method using thermocouple is impractical in medical conditions and prefers the thermal infrared (IR) camera as a non-invasive method. This work proposes a simplified experimental model that uses the thermocouple to determine temperature rise coupled with the thermal image source approach. Furthermore, our new method provides a significant opportunity to calibrate the thermal IR camera by discovering the undetected heat elevation in a workpiece depth. △ Less

Submitted 23 August, 2023; originally announced August 2023.

JutePestDetect: An Intelligent Approach for Jute Pest Identification Using Fine-Tuned Transfer Learning

Authors: Md. Simul Hasan Talukder, Mohammad Raziuddin Chowdhury, Md Sakib Ullah Sourav, Abdullah Al Rakin, Shabbir Ahmed Shuvo, Rejwan Bin Sulaiman, Musarrat Saberin Nipun, Muntarin Islam, Mst Rumpa Islam, Md Aminul Islam, Zubaer Haque

Abstract: In certain Asian countries, Jute is one of the primary sources of income and Gross Domestic Product (GDP) for the agricultural sector. Like many other crops, Jute is prone to pest infestations, and its identification is typically made visually in countries like Bangladesh, India, Myanmar, and China. In addition, this method is time-consuming, challenging, and somewhat imprecise, which poses a subs… ▽ More In certain Asian countries, Jute is one of the primary sources of income and Gross Domestic Product (GDP) for the agricultural sector. Like many other crops, Jute is prone to pest infestations, and its identification is typically made visually in countries like Bangladesh, India, Myanmar, and China. In addition, this method is time-consuming, challenging, and somewhat imprecise, which poses a substantial financial risk. To address this issue, the study proposes a high-performing and resilient transfer learning (TL) based JutePestDetect model to identify jute pests at the early stage. Firstly, we prepared jute pest dataset containing 17 classes and around 380 photos per pest class, which were evaluated after manual and automatic pre-processing and cleaning, such as background removal and resizing. Subsequently, five prominent pre-trained models -DenseNet201, InceptionV3, MobileNetV2, VGG19, and ResNet50 were selected from a previous study to design the JutePestDetect model. Each model was revised by replacing the classification layer with a global average pooling layer and incorporating a dropout layer for regularization. To evaluate the models performance, various metrics such as precision, recall, F1 score, ROC curve, and confusion matrix were employed. These analyses provided additional insights for determining the efficacy of the models. Among them, the customized regularized DenseNet201-based proposed JutePestDetect model outperformed the others, achieving an impressive accuracy of 99%. As a result, our proposed method and strategy offer an enhanced approach to pest identification in the case of Jute, which can significantly benefit farmers worldwide. △ Less

Submitted 28 May, 2023; originally announced August 2023.

Comments: 29 Pages, 7 Tables, 7 Figures, 5 Appendix

Uncovering local aggregated air quality index with smartphone captured images leveraging efficient deep convolutional neural network

Authors: Joyanta Jyoti Mondal, Md. Farhadul Islam, Raima Islam, Nowsin Kabir Rhidi, Sarfaraz Newaz, Meem Arafat Manab, A. B. M. Alim Al Islam, Jannatun Noor

Abstract: The prevalence and mobility of smartphones make these a widely used tool for environmental health research. However, their potential for determining aggregated air quality index (AQI) based on PM2.5 concentration in specific locations remains largely unexplored in the existing literature. In this paper, we thoroughly examine the challenges associated with predicting location-specific PM2.5 concent… ▽ More The prevalence and mobility of smartphones make these a widely used tool for environmental health research. However, their potential for determining aggregated air quality index (AQI) based on PM2.5 concentration in specific locations remains largely unexplored in the existing literature. In this paper, we thoroughly examine the challenges associated with predicting location-specific PM2.5 concentration using images taken with smartphone cameras. The focus of our study is on Dhaka, the capital of Bangladesh, due to its significant air pollution levels and the large population exposed to it. Our research involves the development of a Deep Convolutional Neural Network (DCNN), which we train using over a thousand outdoor images taken and annotated. These photos are captured at various locations in Dhaka, and their labels are based on PM2.5 concentration data obtained from the local US consulate, calculated using the NowCast algorithm. Through supervised learning, our model establishes a correlation index during training, enhancing its ability to function as a Picture-based Predictor of PM2.5 Concentration (PPPC). This enables the algorithm to calculate an equivalent daily averaged AQI index from a smartphone image. Unlike, popular overly parameterized models, our model shows resource efficiency since it uses fewer parameters. Furthermore, test results indicate that our model outperforms popular models like ViT and INN, as well as popular CNN-based models such as VGG19, ResNet50, and MobileNetV2, in predicting location-specific PM2.5 concentration. Our dataset is the first publicly available collection that includes atmospheric images and corresponding PM2.5 measurements from Dhaka. Our codes and dataset are available at https://github.com/lepotatoguy/aqi. △ Less

Submitted 18 January, 2024; v1 submitted 6 August, 2023; originally announced August 2023.

Comments: 18 pages, 7 figures, published to Nature Scientific Reports

Journal ref: Sci Rep 14, 1627 (2024)

Temperature Dependent Failure of Atomically Thin MoTe2

Authors: A S M Redwan Haider, Ahmad Fatehi Ali Mohammed Hezam, Md Akibul Islam, Yeasir Arafat, Mohammad Tanvirul Ferdaous, Sayedus Salehin, Md. Rezwanul Karim

Abstract: In this study, we systematically investigated the mechanical responses of monolayer molybdenum ditelluride (MoTe2) using molecular dynamics (MD) simulations. The tensile behavior of trigonal prismatic phase (2H phase) MoTe2 under uniaxial strain was simulated in the armchair and zigzag directions. We also investigated the crack formation and propagation in both armchair and zigzag directions at 10… ▽ More In this study, we systematically investigated the mechanical responses of monolayer molybdenum ditelluride (MoTe2) using molecular dynamics (MD) simulations. The tensile behavior of trigonal prismatic phase (2H phase) MoTe2 under uniaxial strain was simulated in the armchair and zigzag directions. We also investigated the crack formation and propagation in both armchair and zigzag directions at 10K and 300K to understand the fracture behavior of monolayer MoTe2 at different temperatures. The MD simulations show clean cleavage for the armchair direction, and the cracks were numerous and scattered in the case of the zigzag direction. Finally, we investigated the effect of temperature on Young's modulus and fracture stress of monolayer MoTe2. The results show that at a strain rate of 10^-4 ps^-1, the fracture strength of monolayer MoTe2 in the armchair and zigzag directions at 10K is 16.33 GPa (11.43 N/m) and 13.71 GPa (9.46 N/m) under a 24% and 18% fracture strain, respectively. The fracture strength of monolayer MoTe2 in the armchair and zigzag direction at 600K is 10.81 GPa (7.56 N/m) and 10.13 GPa (7.09 N/m) under a 12.5% and 12.47% fracture strain, respectively. △ Less

Submitted 20 January, 2024; v1 submitted 26 June, 2023; originally announced June 2023.

Comments: 22 Pages, 7 Figures

Structural, elastic, electronic, bonding, thermo-mechanical and optical properties of predicted NbAlB MAB phase in comparison to MoAlB: DFT based ab-initio insights

Authors: Mst. Bina Aktar, F. Parvin, A. K. M. Azharul Islam, S. H. Naqib

Abstract: In this study, we have used density functional theory (DFT) based first-principles investigation of the physical properties of prospective NbAlB compound for the first time. From the analysis of the cohesive energy and enthalpy of formation, it was found that NbAlB is chemically stable. The physical properties of NbAlB have been compared and contrasted with those obtained for MoAlB. Both these MAB… ▽ More In this study, we have used density functional theory (DFT) based first-principles investigation of the physical properties of prospective NbAlB compound for the first time. From the analysis of the cohesive energy and enthalpy of formation, it was found that NbAlB is chemically stable. The physical properties of NbAlB have been compared and contrasted with those obtained for MoAlB. Both these MAB phases are elastically anisotropic, mechanically stable, machinable and brittle materials. Structural and elastic features reflect the layered features. The estimated hardness of NbAlB is 19.0 GPa comparable to that of MoAlB (20.8 GPa) suggesting that predicted NbAlB is a hard compound and is suitable for heavy duty industrial applications. NbAlB is more machinable than MoAlB. Electronic band structure calculations reveal conventional metallic behavior with the electronic density of states at the Fermi level arising mainly due to the Nb 4d orbitals in NbAlB. The electronic density of states at the Fermi level is significantly higher in NbAlB in comparison to MoAlB, indicating that NbAlB is expected to exhibit higher level of electrical conductivity. Electronic dispersion is highly anisotropic for both MoAlB and NbAlB with substantially large electronic effective masses in the out-of-plane directions. The bonding features have been elucidated via the analysis of the band structure and charge density distribution. Both the compounds have mixed covalent, ionic and metallic bonding characteristics. The Fermi surfaces of MoAlB and NbAlB consists of electron and hole like sheets. The Debye temperatures of MoAlB and NbAlB are comparable. The estimated melting temperature of NbAlB is somewhat lower than that of MoAlB. NbAlB shows excellent reflection characteristics suitable to be used as an efficient solar reflector. NbAlB is also expected to absorb ultraviolet radiation very effectively. △ Less

Submitted 21 May, 2023; originally announced May 2023.

AI & Blockchain as sustainable teaching and learning tools to cope with the 4IR

Authors: Md Aminul Islam

Abstract: The Fourth Industrial Revolution (4IR) is transforming the way we live and work, and education is no exception. To cope with the challenges of 4IR, there is a need for innovative and sustainable teaching and learning tools. AI and block chain technologies hold great promise in this regard, with potential benefits such as personalized learning, secure credentialing, and decentralized learning netwo… ▽ More The Fourth Industrial Revolution (4IR) is transforming the way we live and work, and education is no exception. To cope with the challenges of 4IR, there is a need for innovative and sustainable teaching and learning tools. AI and block chain technologies hold great promise in this regard, with potential benefits such as personalized learning, secure credentialing, and decentralized learning networks. This paper presents a review of existing research on AI and block chain in education, analyzing case studies and exploring the potential benefits and challenges of these technologies. The paper also suggests a unique model for integrating AI and block chain into sustainable teaching and learning practices. Future research directions are discussed, including the need for more empirical studies and the exploration of ethical and social implications. The key summary of this discussion is that, by enhancing accessibility, efficacy, and security in education, AI and blockchain have the potential to revolutionise the field. In order to ensure that students can benefit from these potentially game-changing technologies as technology develops, it will be crucial to find ways to harness its power while minimising hazards. Overall, this paper highlights the potential of AI and block chain as sustainable tools for teaching and learning in the 4IR era and their respective advantages, issues and future prospects have been discussed in this writing. △ Less

Submitted 17 September, 2023; v1 submitted 1 May, 2023; originally announced May 2023.

Documentation Practices in Agile Software Development: A Systematic Literature Review

Authors: Md Athikul Islam, Rizbanul Hasan, Nasir U. Eisty

Abstract: Context: Agile development methodologies in the software industry have increased significantly over the past decade. Although one of the main aspects of agile software development (ASD) is less documentation, there have always been conflicting opinions about what to document in ASD. Objective: This study aims to systematically identify what to document in ASD, which documentation tools and methods… ▽ More Context: Agile development methodologies in the software industry have increased significantly over the past decade. Although one of the main aspects of agile software development (ASD) is less documentation, there have always been conflicting opinions about what to document in ASD. Objective: This study aims to systematically identify what to document in ASD, which documentation tools and methods are in use, and how those tools can overcome documentation challenges. Method: We performed a systematic literature review of the studies published between 2010 and June 2021 that discusses agile documentation. Then, we systematically selected a pool of 74 studies using particular inclusion and exclusion criteria. After that, we conducted a quantitative and qualitative analysis using the data extracted from these studies. Results: We found nine primary vital factors to add to agile documentation from our pool of studies. Our analysis shows that agile practitioners have primarily developed their documentation tools and methods focusing on these factors. The results suggest that the tools and techniques in agile documentation are not in sync, and they separately solve different challenges. Conclusions: Based on our results and discussion, researchers and practitioners will better understand how current agile documentation tools and practices perform. In addition, investigation of the synchronization of these tools will be helpful in future research and development. △ Less

Submitted 15 April, 2023; originally announced April 2023.

Comments: Accepted to 21st IEEE/ACIS International Conference on Software Engineering, Management and Applications (SERA 2023). May 23-25, 2023, Orlando, USA

Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models

Authors: Pengfei Li, Jianyi Yang, Mohammad A. Islam, Shaolei Ren

Abstract: The growing carbon footprint of artificial intelligence (AI) models, especially large ones such as GPT-3, has been undergoing public scrutiny. Unfortunately, however, the equally important and enormous water (withdrawal and consumption) footprint of AI models has remained under the radar. For example, training GPT-3 in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 l… ▽ More The growing carbon footprint of artificial intelligence (AI) models, especially large ones such as GPT-3, has been undergoing public scrutiny. Unfortunately, however, the equally important and enormous water (withdrawal and consumption) footprint of AI models has remained under the radar. For example, training GPT-3 in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand may be accountable for 4.2 -- 6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4 -- 6 Denmark or half of the United Kingdom. This is very concerning, as freshwater scarcity has become one of the most pressing challenges shared by all of us in the wake of the rapidly growing population, depleting water resources, and aging water infrastructures. To respond to the global water challenges, AI models can, and also must, take social responsibility and lead by example by addressing their own water footprint. In this paper, we provide a principled methodology to estimate the water footprint of AI models, and also discuss the unique spatial-temporal diversities of AI models' runtime water efficiency. Finally, we highlight the necessity of holistically addressing water footprint along with carbon footprint to enable truly sustainable AI. △ Less

Submitted 29 October, 2023; v1 submitted 6 April, 2023; originally announced April 2023.

Comments: New updates include discussion on water withdrawal and water consumption, scope definition for water, and new estimates of GPT-3's water footprint based on Microsoft's new WUE and PUE data. Source codes available at: https://github.com/Ren-Research/Making-AI-Less-Thirsty

Full Duplex Holographic MIMO for Near-Field Integrated Sensing and Communications

Authors: Ioannis Gavras, Md Atiqul Islam, Besma Smida, George C. Alexandropoulos

Abstract: This paper presents an in-band Full Duplex (FD) integrated sensing and communications system comprising a holographic Multiple-Input Multiple-Output (MIMO) base station, which is capable to simultaneously communicate with multiple users in the downlink direction, while sensing targets being randomly distributed within its coverage area. Considering near-field wireless operation at THz frequencies,… ▽ More This paper presents an in-band Full Duplex (FD) integrated sensing and communications system comprising a holographic Multiple-Input Multiple-Output (MIMO) base station, which is capable to simultaneously communicate with multiple users in the downlink direction, while sensing targets being randomly distributed within its coverage area. Considering near-field wireless operation at THz frequencies, the FD node adopts dynamic metasurface antenna panels for both transmission and reception, which consist of massive numbers of sub-wavelength-spaced metamaterials, enabling reduced cost and power consumption analog precoding and combining. We devise an optimization framework for the FD node's reconfigurable parameters with the dual objective of maximizing the targets' parameters estimation accuracy and the downlink communication performance. Our simulation results verify the integrated sensing and communications capability of the proposed FD holographic MIMO system, showcasing the interplays among its various design parameters. △ Less

Submitted 7 August, 2023; v1 submitted 14 March, 2023; originally announced March 2023.

Comments: 5 pages, 3 figures, EUSIPCO 2023

Sherlock in OSS: A Novel Approach of Content-Based Searching in Object Storage System

Authors: Jannatun Noor, Rizwanul Haque Ratul, Mir Rownak Ali Uday, Joyanta Jyoti Mondal, Md. Sadiqul Islam Sakif, A. B. M. Alim Al Islam

Abstract: Object Storage Systems (OSS) inside a cloud promise scalability, durability, availability, and concurrency. However, open-source OSS does not have a specific approach to letting users and administrators search based on the data, which is contained inside the object storage, without involving the entire cloud infrastructure. Therefore, in this paper, we propose Sherlock, a novel Content-Based Searc… ▽ More Object Storage Systems (OSS) inside a cloud promise scalability, durability, availability, and concurrency. However, open-source OSS does not have a specific approach to letting users and administrators search based on the data, which is contained inside the object storage, without involving the entire cloud infrastructure. Therefore, in this paper, we propose Sherlock, a novel Content-Based Searching (CoBS) architecture to extract additional information from images and documents. Here, we store the additional information in an Elasticsearch-enabled database, which helps us to search for our desired data based on its contents. This approach works in two sequential stages. First, the data will be uploaded to a classifier that will determine the data type and send it to the specific model for the data. Here, the images that are being uploaded are sent to our trained model for object detection, and the documents are sent for keyword extraction. Next, the extracted information is sent to Elasticsearch, which enables searching based on the contents. Because the precision of the models is so fundamental to the search's correctness, we train our models with comprehensive datasets (Microsoft COCO Dataset for multimedia data and SemEval2017 Dataset for document data). Furthermore, we put our designed architecture to the test with a real-world implementation of an open-source OSS called OpenStack Swift. We upload images into the dataset of our implementation in various segments to find out the efficacy of our proposed model in real-life Swift object storage. △ Less

Submitted 6 May, 2023; v1 submitted 24 January, 2023; originally announced March 2023.

Comments: 13 Pages, 9 Figures, Submitted to IEEE Transactions on Parallel and Distributed Systems for possible publication. arXiv admin note: substantial text overlap with arXiv:1306.3075 by other authors; substantial text overlap with arXiv:1910.05786 by other authors without attribution

Data analytics on key indicators for the city's urban services and dashboards for leadership and decision-making

Authors: Md Aminul Islam, Md Abu Sufian

Abstract: Cities are continuously evolving human settlements. Our cities are under strain in an increasingly urbanized world, and planners, decision-makers, and communities must be ready to adapt. Data is an important resource for municipal administration. Some technologies aid in the collection, processing, and visualization of urban data, assisting in the interpretation and comprehension of how urban syst… ▽ More Cities are continuously evolving human settlements. Our cities are under strain in an increasingly urbanized world, and planners, decision-makers, and communities must be ready to adapt. Data is an important resource for municipal administration. Some technologies aid in the collection, processing, and visualization of urban data, assisting in the interpretation and comprehension of how urban systems operate. The relationship between data analytics and smart cities has come to light in recent years as interest in both has grown. A sophisticated network of interconnected systems, including planners and inhabitants, is what is known as a smart city. Data analysis has the potential to support data-driven decision-making in the context of smart cities. Both urban managers and residents are becoming more interested in city dashboards. Dashboards may collect, display, analyze, and provide information on regional performance to help smart cities development have sustainability. In order to assist decision-making processes and enhance the performance of cities, we examine how dashboards might be used to acquire accurate and representative information regarding urban challenges. This chapter culminates Data Analytics on key indicators for the city's urban services and dashboards for leadership and decision-making. A single web page with consolidated information, real-time data streams pertinent to planners and decision-makers as well as residents' everyday lives, and site analytics as a method to assess user interactions and preferences are among the proposals for urban dashboards. Keywords: -Dashboard, data analytics, smart city, sustainability, Smart cities, City dashboards, Urban services, Decision-making, Interconnected systems, Real-time data streams, Key indicators, and Urban challenges. △ Less

Submitted 12 September, 2023; v1 submitted 1 December, 2022; originally announced December 2022.

Blockchain Technology: A tool to solve the challenges of education sector in developing countries

Authors: Md Aminul Islam

Abstract: The education system is getting diversified, challenged, and blended for the overwhelming advancement of disruptive technology. The core purpose of this chapter is to visualize the probable solutions of the modern education system using blockchain technology. The entire chapter has been discussed on the basis of present solution and projection of future inventions to smoothen the education system.… ▽ More The education system is getting diversified, challenged, and blended for the overwhelming advancement of disruptive technology. The core purpose of this chapter is to visualize the probable solutions of the modern education system using blockchain technology. The entire chapter has been discussed on the basis of present solution and projection of future inventions to smoothen the education system. The fourth industrial revolution (4IR) is changing our experiences in terms of education and other lifestyle. Delivering lectures, interacting between learners and educations, evaluating learning outcomes, and verifying educational credentials might be smoother, easier, faster, cheaper, and jollier than before. Blockchain technology can contribute to the education provider to tackle all those existing problems to create a comfortable learning environment to all irrespective to their economic backgrounds and geographic location. How this technology can contribute to improve Reviewing recent inventions in this technology, the chapter explains some of the strategies to go beyond the ongoing projects around the world. A set of models are arranged to enable the readers mind for future inventions in the realm of educationists. Keywords: -Blockchain, 4IR, educators, learning outcome. △ Less

Submitted 18 November, 2022; originally announced November 2022.

Quantifying and Learning Static vs. Dynamic Information in Deep Spatiotemporal Networks

Authors: Matthew Kowal, Mennatullah Siam, Md Amirul Islam, Neil D. B. Bruce, Richard P. Wildes, Konstantinos G. Derpanis

Abstract: There is limited understanding of the information captured by deep spatiotemporal models in their intermediate representations. For example, while evidence suggests that action recognition algorithms are heavily influenced by visual appearance in single frames, no quantitative methodology exists for evaluating such static bias in the latent representation compared to bias toward dynamics. We tackl… ▽ More There is limited understanding of the information captured by deep spatiotemporal models in their intermediate representations. For example, while evidence suggests that action recognition algorithms are heavily influenced by visual appearance in single frames, no quantitative methodology exists for evaluating such static bias in the latent representation compared to bias toward dynamics. We tackle this challenge by proposing an approach for quantifying the static and dynamic biases of any spatiotemporal model, and apply our approach to three tasks, action recognition, automatic video object segmentation (AVOS) and video instance segmentation (VIS). Our key findings are: (i) Most examined models are biased toward static information. (ii) Some datasets that are assumed to be biased toward dynamics are actually biased toward static information. (iii) Individual channels in an architecture can be biased toward static, dynamic or a combination of the two. (iv) Most models converge to their culminating biases in the first half of training. We then explore how these biases affect performance on dynamically biased datasets. For action recognition, we propose StaticDropout, a semantically guided dropout that debiases a model from static information toward dynamics. For AVOS, we design a better combination of fusion and cross connection layers compared with previous architectures. △ Less

Submitted 3 November, 2022; originally announced November 2022.

Comments: arXiv admin note: substantial text overlap with arXiv:2206.02846

Analysis and prediction of heart stroke from ejection fraction and serum creatinine using LSTM deep learning approach

Authors: Md Ershadul Haque, Salah Uddin, Md Ariful Islam, Amira Khanom, Abdulla Suman, Manoranjan Paul

Abstract: The combination of big data and deep learning is a world-shattering technology that can greatly impact any objective if used properly. With the availability of a large volume of health care datasets and progressions in deep learning techniques, systems are now well equipped to predict the future trend of any health problems. From the literature survey, we found the SVM was used to predict the hear… ▽ More The combination of big data and deep learning is a world-shattering technology that can greatly impact any objective if used properly. With the availability of a large volume of health care datasets and progressions in deep learning techniques, systems are now well equipped to predict the future trend of any health problems. From the literature survey, we found the SVM was used to predict the heart failure rate without relating objective factors. Utilizing the intensity of important historical information in electronic health records (EHR), we have built a smart and predictive model utilizing long short-term memory (LSTM) and predict the future trend of heart failure based on that health record. Hence the fundamental commitment of this work is to predict the failure of the heart using an LSTM based on the patient's electronic medicinal information. We have analyzed a dataset containing the medical records of 299 heart failure patients collected at the Faisalabad Institute of Cardiology and the Allied Hospital in Faisalabad (Punjab, Pakistan). The patients consisted of 105 women and 194 men and their ages ranged from 40 and 95 years old. The dataset contains 13 features, which report clinical, body, and lifestyle information responsible for heart failure. We have found an increasing trend in our analysis which will contribute to advancing the knowledge in the field of heart stroke prediction. △ Less

Submitted 27 September, 2022; originally announced September 2022.

The rise of 212 MAX phase borides, Ti$_2$PB$_2$, Zr$_2$PbB$_2$, and Nb$_2$AB$_2$ [A = P, S]: DFT insights into the physical properties for thermo-mechanical applications

Authors: M. A. Ali, M. M. Hossain, M. M. Uddin, A. K. M. A. Islam, S. H. Naqib

Abstract: An interesting class of ternary metallic borides, known as the 212 MAX phase borides, is the recent advancement of the MAX phase family. In this article, results from ab-initio calculations on unexplored Ti$_2$PB$_2$, Zr$_2$PbB$_2$, and Nb$_2$AB$_2$ [A = P, S] are reported wherein Ti$_2$PB$_2$ along with its 211 boride phase Ti$_2$PB are predicted for the first time. The stability was confirmed by… ▽ More An interesting class of ternary metallic borides, known as the 212 MAX phase borides, is the recent advancement of the MAX phase family. In this article, results from ab-initio calculations on unexplored Ti$_2$PB$_2$, Zr$_2$PbB$_2$, and Nb$_2$AB$_2$ [A = P, S] are reported wherein Ti$_2$PB$_2$ along with its 211 boride phase Ti$_2$PB are predicted for the first time. The stability was confirmed by calculating the formation energy, phonon dispersion curve, and elastic stiffness constants. The obtained elastic constants, elastic moduli, and Vickers hardness values of Ti$_2$PB$_2$, Zr$_2$PbB$_2$, and Nb$_2$AB$_2$ [A = P, S] were found to be significantly larger than those of their counterparts 211 borides and carbides, in a trend similar to other 212 borides. The studied compounds are brittle like most of the MAX and MAB phases. The electronic band structure and density of states revealed the metallic nature of the titled borides. Several thermal parameters were explored, certifying the suitability of Ti2PB2, Zr2PbB2, and Nb2AB2 [A = P, S] compared to their counterparts, and a similar trend was found for the other 212 borides. The obtained results predict that Ti2PB2, Zr2PbB2, and Nb2AB2 [A = P, S] have significant potential for use as efficient thermal barrier coating materials. The response of Ti$_2$PB$_2$, Zr$_2$PbB$_2$, and Nb$_2$AB$_2$ [A = P, S] to the incident photon was studied by computing the dielectric constant (real and imaginary part), refractive index, absorption coefficient, photoconductivity, reflectivity, and energy loss function. The ability to protect from solar heating was revealed from the studied reflectivity spectra. In this work, we have explored the physical basis of the improved thermo-mechanical properties of 212 MAX phase borides compared to their carbide and boride counterparts. △ Less

Submitted 2 September, 2022; v1 submitted 17 July, 2022; originally announced July 2022.

Comments: 36 pages

A Deeper Dive Into What Deep Spatiotemporal Networks Encode: Quantifying Static vs. Dynamic Information

Authors: Matthew Kowal, Mennatullah Siam, Md Amirul Islam, Neil D. B. Bruce, Richard P. Wildes, Konstantinos G. Derpanis

Abstract: Deep spatiotemporal models are used in a variety of computer vision tasks, such as action recognition and video object segmentation. Currently, there is a limited understanding of what information is captured by these models in their intermediate representations. For example, while it has been observed that action recognition algorithms are heavily influenced by visual appearance in single static… ▽ More Deep spatiotemporal models are used in a variety of computer vision tasks, such as action recognition and video object segmentation. Currently, there is a limited understanding of what information is captured by these models in their intermediate representations. For example, while it has been observed that action recognition algorithms are heavily influenced by visual appearance in single static frames, there is no quantitative methodology for evaluating such static bias in the latent representation compared to bias toward dynamic information (e.g. motion). We tackle this challenge by proposing a novel approach for quantifying the static and dynamic biases of any spatiotemporal model. To show the efficacy of our approach, we analyse two widely studied tasks, action recognition and video object segmentation. Our key findings are threefold: (i) Most examined spatiotemporal models are biased toward static information; although, certain two-stream architectures with cross-connections show a better balance between the static and dynamic information captured. (ii) Some datasets that are commonly assumed to be biased toward dynamics are actually biased toward static information. (iii) Individual units (channels) in an architecture can be biased toward static, dynamic or a combination of the two. △ Less

Submitted 6 June, 2022; originally announced June 2022.

Comments: CVPR 2022

Distributed Ledger Technology based Integrated Healthcare Solution for Bangladesh

Authors: Md. Ariful Islam, Md. Antonin Islam, Md. Amzad Hossain Jacky, Md. Al-Amin, M. Saef Ullah Miah, Md Muhidul Islam Khan, Md. Iqbal Hossain

Abstract: Healthcare data is sensitive and requires great protection. Encrypted electronic health records (EHRs) contain personal and sensitive data such as names and addresses. Having access to patient data benefits all of them. This paper proposes a blockchain-based distributed healthcare application platform for Bangladeshi public and private healthcare providers. Using data immutability and smart contra… ▽ More Healthcare data is sensitive and requires great protection. Encrypted electronic health records (EHRs) contain personal and sensitive data such as names and addresses. Having access to patient data benefits all of them. This paper proposes a blockchain-based distributed healthcare application platform for Bangladeshi public and private healthcare providers. Using data immutability and smart contracts, the suggested application framework allows users to create safe digital agreements for commerce or collaboration. Thus, all enterprises may securely collaborate using the same blockchain network, gaining data openness and read/write capacity. The proposed application consists of various application interfaces for various system users. For data integrity, privacy, permission and service availability, the proposed solution leverages Hyperledger fabric and Blockchain as a Service. Everyone will also have their own profile in the portal. A unique identity for each person and the installation of digital information centres across the country have greatly eased the process. It will collect systematic health data from each person which will be beneficial for research institutes and health-related organisations. A national data warehouse in Bangladesh is feasible for this application and It is also possible to keep a clean health sector by analysing data stored in this warehouse and conducting various purification algorithms using technologies like Data Science. Given that Bangladesh has both public and private health care, a straightforward digital strategy for all organisations is essential. △ Less

Submitted 30 May, 2022; originally announced May 2022.

Comments: 21 pages, 16 figures, 4 tables

Simultaneous Multi-User MIMO Communications and Multi-Target Tracking with Full Duplex Radios

Authors: Md Atiqul Islam, George C. Alexandropoulos, Besma Smida

Abstract: In this paper, we present an Integrated Sensing and Communications (ISAC) system enabled by in-band Full Duplex (FD) radios, where a massive Multiple-Input Multiple-Output (MIMO) base station equipped with hybrid Analog and Digital (A/D) beamformers is communicating with multiple DownLink (DL) users, and simultaneously estimates via the same signaling waveforms the Direction of Arrival (DoA) as we… ▽ More In this paper, we present an Integrated Sensing and Communications (ISAC) system enabled by in-band Full Duplex (FD) radios, where a massive Multiple-Input Multiple-Output (MIMO) base station equipped with hybrid Analog and Digital (A/D) beamformers is communicating with multiple DownLink (DL) users, and simultaneously estimates via the same signaling waveforms the Direction of Arrival (DoA) as well as the range of radar targets randomly distributed within its coverage area. Capitalizing on a recent reduced-complexity FD hybrid A/D beamforming architecture, we devise a joint radar target tracking and DL data transmission protocol. An optimization framework for the joint design of the massive A/D beamformers and the Self-Interference (SI) cancellation unit, with the dual objective of maximizing the radar tracking accuracy and DL communication performance, is presented. Our simulation results at millimeter wave frequencies using 5G NR wideband waveforms, showcase the accuracy of the radar target tracking performance of the proposed system, which simultaneously offers increased sum rate compared with benchmark schemes. △ Less

Submitted 17 May, 2022; originally announced May 2022.

Comments: 6 pages, 5 figures. Submitted for publication in the Proceedings of IEEE Global Communications Conference (GLOBECOM), 2022

Full Duplex Massive MIMO Architectures: Recent Advances, Applications, and Future Directions

Authors: George C. Alexandropoulos, Md Atiqul Islam, Besma Smida

Abstract: The increasingly demanding objectives for next generation wireless communications have spurred recent research activities on multi-antenna transceiver hardware architectures and relevant intelligent communication schemes. Among them belong the Full Duplex (FD) Multiple-Input Multiple-Output (MIMO) architectures, which offer the potential for simultaneous uplink and downlink operations in the entir… ▽ More The increasingly demanding objectives for next generation wireless communications have spurred recent research activities on multi-antenna transceiver hardware architectures and relevant intelligent communication schemes. Among them belong the Full Duplex (FD) Multiple-Input Multiple-Output (MIMO) architectures, which offer the potential for simultaneous uplink and downlink operations in the entire frequency band. However, as the number of antenna elements increases, the interference signal leaking from the transmitter of the FD radio to its receiver becomes more severe. In this article, we present a unified FD massive MIMO architecture comprising analog and digital transmit and receive BeamForming (BF), as well as analog and digital SI cancellation, which can be jointly optimized for various performance objectives and complexity requirements. Performance evaluation results for applications of the proposed architecture to fully digital and hybrid analog and digital BF operations using recent algorithmic designs, as well as simultaneous communication of data and control signals are presented. It is shown that the proposed architecture, for both small and large numbers of antennas, enables improved spectral efficiency FD communications with fewer analog cancellation elements compared to various benchmark schemes. The article is concluded with a list of open challenges and research directions for future FD massive MIMO communication systems and their promising applications. △ Less

Submitted 17 May, 2022; originally announced May 2022.

Comments: 7 pages, 5 figure. Under review for publication in IEEE Vehicular Technology Magazine, 2022

Shashthosheba: Dissecting Perception of Bangladeshi People towards Telemedicine Apps through the Lens of Features of the Apps

Authors: Waqar Hassan Khan, Md Al Imran, Ahmed Nafis Fuad, Mohammed Latif Siddiq, A. B. M. Alim Al Islam

Abstract: Bangladesh, a developing country with a large and dense population, has recently seen significant economic as well as technological developments. The growth of technology has resulted in a dramatic increase in the number of smartphone users in Bangladesh, and as such, mobile apps have become an increasingly important part of peoples' life, even encompassing healthcare services. However, the apps u… ▽ More Bangladesh, a developing country with a large and dense population, has recently seen significant economic as well as technological developments. The growth of technology has resulted in a dramatic increase in the number of smartphone users in Bangladesh, and as such, mobile apps have become an increasingly important part of peoples' life, even encompassing healthcare services. However, the apps used in healthcare (telemedicine to be specific) in Bangladesh are yet to be studied from the perspective of their features as per the voices of the users as well as service providers. Therefore, in this study, we focus on the features of the telemedicine apps used in Bangladesh. First, we evaluated the present status of existing telemedicine apps in Bangladesh, as well as their benefits and drawbacks in the context of HCI. We analyzed publicly accessible reviews of several Bangladeshi telemedicine apps (N = 14) to evaluate the user impressions. Additionally, to ascertain the public opinion of these apps, we performed a survey in which the patients (N = 87) participated willingly. Our analysis of the collected opinions reveals what users experience, what they appreciate, and what they are concerned about when they use telemedicine apps. Additionally, our study demonstrates what users expect from telemedicine apps, independent of their past experience. Finally, we explore how to address the issues we discovered and how telemedicine may be used to effectively offer healthcare services throughout the country. To the best of our knowledge, this study is the first to analyze the perception of the people of Bangladesh towards telemedicine apps from the perspective of features of the apps. △ Less

Submitted 6 May, 2022; v1 submitted 5 May, 2022; originally announced May 2022.

Comments: 12 pages

Practical Efficient Microservice Autoscaling with QoS Assurance

Authors: Md Rajib Hossen, Mohammad A. Islam, Kishwar Ahmed

Abstract: Cloud applications are increasingly moving away from monolithic services to agile microservices-based deployments. However, efficient resource management for microservices poses a significant hurdle due to the sheer number of loosely coupled and interacting components. The interdependencies between various microservices make existing cloud resource autoscaling techniques ineffective. Meanwhile, ma… ▽ More Cloud applications are increasingly moving away from monolithic services to agile microservices-based deployments. However, efficient resource management for microservices poses a significant hurdle due to the sheer number of loosely coupled and interacting components. The interdependencies between various microservices make existing cloud resource autoscaling techniques ineffective. Meanwhile, machine learning (ML) based approaches that try to capture the complex relationships in microservices require extensive training data and cause intentional SLO violations. Moreover, these ML-heavy approaches are slow in adapting to dynamically changing microservice operating environments. In this paper, we propose PEMA (Practical Efficient Microservice Autoscaling), a lightweight microservice resource manager that finds efficient resource allocation through opportunistic resource reduction. PEMA's lightweight design enables novel workload-aware and adaptive resource management. Using three prototype microservice implementations, we show that PEMA can find efficient resource allocation and save up to 33% resource compared to the commercial rule-based resource allocations. △ Less

Submitted 9 August, 2022; v1 submitted 31 January, 2022; originally announced February 2022.

Comments: Appeared on HPDC'22, 13 pages, 20 figures

Integrated Sensing and Communication with Millimeter Wave Full Duplex Hybrid Beamforming

Authors: Md Atiqul Islam, George C. Alexandropoulos, Besma Smida

Abstract: Integrated Sensing and Communication (ISAC) has attracted substantial attraction in recent years for spectral efficiency improvement, enabling hardware and spectrum sharing for simultaneous sensing and signaling operations. In-band Full Duplex (FD) is being considered as a key enabling technology for ISAC applications due to its simultaneous transmission and reception capability. In this paper, we… ▽ More Integrated Sensing and Communication (ISAC) has attracted substantial attraction in recent years for spectral efficiency improvement, enabling hardware and spectrum sharing for simultaneous sensing and signaling operations. In-band Full Duplex (FD) is being considered as a key enabling technology for ISAC applications due to its simultaneous transmission and reception capability. In this paper, we present an FD-based ISAC system operating at millimeter Wave (mmWave) frequencies, where a massive Multiple-Input Multiple-Output (MIMO) Base Station (BS) node employing hybrid Analog and Digital (A/D) beamforming is communicating with a DownLink (DL) multi-antenna user and the same waveform is utilized at the BS receiver for sensing the radar targets in its coverage environment. We develop a sensing algorithm that is capable of estimating Direction of Arrival (DoA), range, and relative velocity of the radar targets. A joint optimization framework for designing the A/D transmit and receive beamformers as well as the Self-Interference (SI) cancellation is presented with the objective to maximize the achievable DL rate and the accuracy of the radar target sensing performance. Our simulation results, considering fifth Generation (5G) Orthogonal Frequency Division Multiplexing (OFDM) waveforms, verify our approach's high precision in estimating DoA, range, and velocity of multiple radar targets, while maximizing the DL communication rate. △ Less

Submitted 13 January, 2022; originally announced January 2022.

Comments: 6 pages, 4 figures, Submitted for publication in the Proceedings of IEEE ICC 2022, Seoul, South Korea

Joint Analog and Digital Transceiver Design for Wideband Full Duplex MIMO Systems

Authors: Md Atiqul Islam, George C. Alexandropoulos, Besma Smida

Abstract: In this paper, we propose a wideband Full Duplex (FD) Multiple-Input Multiple-Output (MIMO) communication system comprising of an FD MIMO node simultaneously communicating with two multi-antenna UpLink (UL) and DownLink (DL) nodes utilizing the same time and frequency resources. To suppress the strong Self-Interference (SI) signal due to simultaneous transmission and reception in FD MIMO systems,… ▽ More In this paper, we propose a wideband Full Duplex (FD) Multiple-Input Multiple-Output (MIMO) communication system comprising of an FD MIMO node simultaneously communicating with two multi-antenna UpLink (UL) and DownLink (DL) nodes utilizing the same time and frequency resources. To suppress the strong Self-Interference (SI) signal due to simultaneous transmission and reception in FD MIMO systems, we propose a joint design of Analog and Digital (A/D) cancellation as well as transmit and receive beamforming capitalizing on baseband Orthogonal Frequency-Division Multiplexing (OFDM) signal modeling. Considering practical transmitter impairments, we present a multi-tap wideband analog canceller architecture whose number of taps does not scale with the number of transceiver antennas and multipath SI components. We also propose a novel adaptive digital cancellation based on truncated singular value decomposition that reduces the residual SI signal estimation parameters. To maximize the FD sum rate, a joint optimization framework is presented for A/D cancellation and digital beamforming. Finally, our extensive waveform simulation results demonstrate that the proposed wideband FD MIMO design exhibits higher SI cancellation capability with reduced complexity compared to existing cancellation techniques, resulting in improved achievable rate performance. △ Less

Submitted 16 December, 2021; originally announced December 2021.

Comments: 32 pages, 10 figures, Under Review for Publication in the IEEE Transactions on Wireless Communications

A comprehensive first principles calculations on (Ba0.82K0.18)(Bi0.53Pb0.47)O3 single-cubic-perovskite superconductor

Authors: Mirza H. K. Rubel, Sujon Kumar Mitro, Khandaker Monower Hossain, Md. Mijanur Rahaman, M. Khalid Hossain, Jaker Hossain, B. K. Mondal, Istiak Ahmed, A. K. M. A. Islam, A. El-Denglawey

Abstract: In this present study, the pseudopotential plane-wave (PP-PW) pathway in the scheme of density functional theory (DFT) is utilized to investigate the various physical properties on (Ba0.82K0.18)(Bi0.53Pb0.47)O3 (BKBPO) single perovskite superconductor. We have analyzed elastic constants and moduli at zero and elevated pressures (up to 25 GPa) as well. We also have investigated the anisotropic natu… ▽ More In this present study, the pseudopotential plane-wave (PP-PW) pathway in the scheme of density functional theory (DFT) is utilized to investigate the various physical properties on (Ba0.82K0.18)(Bi0.53Pb0.47)O3 (BKBPO) single perovskite superconductor. We have analyzed elastic constants and moduli at zero and elevated pressures (up to 25 GPa) as well. We also have investigated the anisotropic nature incorporating both the theoretical indices and graphical representations in 2D and 3D dimensions, which reveals a high level of anisotropy. The flatness of the energy bands near EF is a sign of Van-Hf singularity that might increase the electron pairing and origination of high-TC superconductivity. The computed band structure exhibits its metallic characteristics is confirmed by band overlapping. A band of DOS is formed for the strong hybridization of the constituent elements. The orbital electrons of O-2p contribute most dominantly at EF in contrast to all orbital electrons. The orbital electrons at the EF are higher from both the partial density of states and charge density mapping investigation. The coexistence of the electron and hole-like Fermi sheets exhibits the multi-band nature of BKBPO. On the other hand, Fermi surfaces with flat faces promote transport features and Fermi surface nesting as well. The calculated value of the electron-phonon coupling constant (λ = 1.46) is slightly lower than the isostructural superconductor, which indicates that the studied BKBPO can be treated as a strongly coupled superconductor similar to the reported isostructural perovskite superconductors. Furthermore, the thermodynamic properties have been evaluated and analyzed at elevated temperature and pressure by using harmonic Debye approximation (QHDA). △ Less

Submitted 7 December, 2021; originally announced December 2021.

Comments: 20 pages, 7 figures, 6 tables

Journal ref: Materials Today Communications, 33 (2022) 104302

Simpler Does It: Generating Semantic Labels with Objectness Guidance

Authors: Md Amirul Islam, Matthew Kowal, Sen Jia, Konstantinos G. Derpanis, Neil D. B. Bruce

Abstract: Existing weakly or semi-supervised semantic segmentation methods utilize image or box-level supervision to generate pseudo-labels for weakly labeled images. However, due to the lack of strong supervision, the generated pseudo-labels are often noisy near the object boundaries, which severely impacts the network's ability to learn strong representations. To address this problem, we present a novel f… ▽ More Existing weakly or semi-supervised semantic segmentation methods utilize image or box-level supervision to generate pseudo-labels for weakly labeled images. However, due to the lack of strong supervision, the generated pseudo-labels are often noisy near the object boundaries, which severely impacts the network's ability to learn strong representations. To address this problem, we present a novel framework that generates pseudo-labels for training images, which are then used to train a segmentation model. To generate pseudo-labels, we combine information from: (i) a class agnostic objectness network that learns to recognize object-like regions, and (ii) either image-level or bounding box annotations. We show the efficacy of our approach by demonstrating how the objectness network can naturally be leveraged to generate object-like regions for unseen categories. We then propose an end-to-end multi-task learning strategy, that jointly learns to segment semantics and objectness using the generated pseudo-labels. Extensive experiments demonstrate the high quality of our generated pseudo-labels and effectiveness of the proposed framework in a variety of domains. Our approach achieves better or competitive performance compared to existing weakly-supervised and semi-supervised methods. △ Less

Submitted 19 October, 2021; originally announced October 2021.

Comments: BMVC 2021

Predicting Users' Value Changes by the Friends' Influence from Social Media Usage

Authors: Md. Saddam Hossain Mukta, Ahmed Shahriar Sakib, Md. Adnanul Islam, Mohiuddin Ahmed, Mumshad Ahamed Rifat

Abstract: Basic human values represent a set of values such as security, independence, success, kindness, and pleasure, which we deem important to our lives. Each of us holds different values with different degrees of significance. Existing studies show that values of a person can be identified from their social network usage. However, the value priority of a person may change over time due to different fac… ▽ More Basic human values represent a set of values such as security, independence, success, kindness, and pleasure, which we deem important to our lives. Each of us holds different values with different degrees of significance. Existing studies show that values of a person can be identified from their social network usage. However, the value priority of a person may change over time due to different factors such as life experiences, influence, social structure and technology. Existing studies do not conduct any analysis regarding the change of users' value from the social influence, i.e., group persuasion, form the social media usage. In our research, first, we predict users' value score by the influence of friends from their social media usage. We propose a Bounded Confidence Model (BCM) based value dynamics model from 275 different ego networks in Facebook that predicts how social influence may persuade a person to change their value over time. Then, to predict better, we use particle swarm optimization based hyperparameter tuning technique. We observe that these optimized hyperparameters produce accurate future value score. We also run our approach with different machine learning based methods and find support vector regression (SVR) outperforms other regressor models. By using SVR with the best hyperparameters of BCM model, we find the lowest Mean Squared Error (MSE) score 0.00347. △ Less

Submitted 12 September, 2021; originally announced September 2021.

Comments: 11 pages

ACM Class: I.2.8

Journal ref: http://sbp-brims.org/2021/papers/working-papers/2021_SBP-BRiMS_FinalPDF_20.pdf

Direction-Assisted Beam Management in Full Duplex Millimeter Wave Massive MIMO Systems

Authors: Md Atiqul Islam, George C. Alexandropoulos, Besma Smida

Abstract: Recent applications of the Full Duplex (FD) technology focus on enabling simultaneous control communication and data transmission to reduce the control information exchange overhead, which impacts end-to-end latency and spectral efficiency. In this paper, we present a simultaneous direction estimation and data transmission scheme for millimeter Wave (mmWave) massive Multiple-Input Multiple-Output… ▽ More Recent applications of the Full Duplex (FD) technology focus on enabling simultaneous control communication and data transmission to reduce the control information exchange overhead, which impacts end-to-end latency and spectral efficiency. In this paper, we present a simultaneous direction estimation and data transmission scheme for millimeter Wave (mmWave) massive Multiple-Input Multiple-Output (MIMO) systems, enabled by a recent FD MIMO technology with reduced hardware complexity Self-Interference (SI) cancellation. We apply the proposed framework in the mmWave analog beam management problem, considering a base station equipped with a large transmit antenna array realizing downlink analog beamforming and few digitally controlled receive antenna elements used for uplink Direction-of-Arrival (DoA) estimation. A joint optimization framework for designing the DoA-assisted analog beamformer and the analog as well as digital SI cancellation is presented with the objective to maximize the achievable downlink rate. Our simulation results showcase that the proposed scheme outperforms its conventional half-duplex counterpart, yielding reduced DoA estimation error and superior downlink data rate. △ Less

Submitted 15 September, 2021; originally announced September 2021.

Comments: 6 pages, 5 figures. To appear in the Proceedings of the IEEE Global Communications Conference (GLOBECOM), 7-11 Dec. 2021, Madrid, Spain

SegMix: Co-occurrence Driven Mixup for Semantic Segmentation and Adversarial Robustness

Authors: Md Amirul Islam, Matthew Kowal, Konstantinos G. Derpanis, Neil D. B. Bruce

Abstract: In this paper, we present a strategy for training convolutional neural networks to effectively resolve interference arising from competing hypotheses relating to inter-categorical information throughout the network. The premise is based on the notion of feature binding, which is defined as the process by which activations spread across space and layers in the network are successfully integrated to… ▽ More In this paper, we present a strategy for training convolutional neural networks to effectively resolve interference arising from competing hypotheses relating to inter-categorical information throughout the network. The premise is based on the notion of feature binding, which is defined as the process by which activations spread across space and layers in the network are successfully integrated to arrive at a correct inference decision. In our work, this is accomplished for the task of dense image labelling by blending images based on (i) categorical clustering or (ii) the co-occurrence likelihood of categories. We then train a feature binding network which simultaneously segments and separates the blended images. Subsequent feature denoising to suppress noisy activations reveals additional desirable properties and high degrees of successful predictions. Through this process, we reveal a general mechanism, distinct from any prior methods, for boosting the performance of the base segmentation and saliency network while simultaneously increasing robustness to adversarial attacks. △ Less

Submitted 23 August, 2021; originally announced August 2021.

Comments: Under submission at IJCV (BMVC 2020 Extension). arXiv admin note: substantial text overlap with arXiv:2008.05667

Newly synthesized 3D boron-rich chalcogenides B12X (X = S, Se): Theoretical characterization of physical properties for optoelectronic and mechanical applications

Authors: M. M. Hossain, M. A. Ali, M. M. Uddin, S. H. Naqib, A. K. M. A. Islam

Abstract: Boron rich chalcogenides have been predicted to have excellent properties for optical and mechanical applications in recent times. In this regard, we report the electronic, optical and mechanical properties of recently synthesized boron rich chalcogenide compounds, B12X (X = S and Se) using density functional theory for the first time. The effects of exchange and correlation functional on these pr… ▽ More Boron rich chalcogenides have been predicted to have excellent properties for optical and mechanical applications in recent times. In this regard, we report the electronic, optical and mechanical properties of recently synthesized boron rich chalcogenide compounds, B12X (X = S and Se) using density functional theory for the first time. The effects of exchange and correlation functional on these properties are also investigated. The consistency of the obtained crystal structure with the reported experimental results has been checked in terms of lattice parameters. The considered materials are mechanically stable, brittle and elastically anisotropic. Furthermore, the elastic moduli and hardness parameters are calculated, which show that B12S is likely to be a prominent member of hard materials family compared to B12Se. The origin of different in hardness is explained on the basis of density of states near the Fermi level. Reasonably good values of fracture toughness and machinability index for B12X (X= S and Se) are reported. The melting point, Tm for the B12S and B12Se compounds suggests that both solids are stable, at least up to 4208 and 3577 K, respectively. Indirect band gap of B12S (2.27 eV) and B12Se (1.30 eV) are obtained using the HSE06 functional.The electrons of B12Se compound show lighter average effective mass compared to that of B12S compound, which signifies higher mobility of charge carriers in B12Se. The optical properties are characterized using GGA-PBE and HSE06 method and discussed in detail. These compounds possess bulk optical anisotropy and excellent absorption coefficients in visible light region along with very low static value of reflectivity spectra (range: 7.42-14.0% using both functionals) are noted. Such useful features of the compounds under investigation show promise for applications in optoelectronic and mechanical sectors. △ Less

Submitted 19 August, 2021; originally announced August 2021.

Comments: 34 pages, 13 figures, 6 Tables, This study will be submitted for publication internationally very soon

Global Pooling, More than Meets the Eye: Position Information is Encoded Channel-Wise in CNNs

Authors: Md Amirul Islam, Matthew Kowal, Sen Jia, Konstantinos G. Derpanis, Neil D. B. Bruce

Abstract: In this paper, we challenge the common assumption that collapsing the spatial dimensions of a 3D (spatial-channel) tensor in a convolutional neural network (CNN) into a vector via global pooling removes all spatial information. Specifically, we demonstrate that positional information is encoded based on the ordering of the channel dimensions, while semantic information is largely not. Following th… ▽ More In this paper, we challenge the common assumption that collapsing the spatial dimensions of a 3D (spatial-channel) tensor in a convolutional neural network (CNN) into a vector via global pooling removes all spatial information. Specifically, we demonstrate that positional information is encoded based on the ordering of the channel dimensions, while semantic information is largely not. Following this demonstration, we show the real world impact of these findings by applying them to two applications. First, we propose a simple yet effective data augmentation strategy and loss function which improves the translation invariance of a CNN's output. Second, we propose a method to efficiently determine which channels in the latent representation are responsible for (i) encoding overall position information or (ii) region-specific positions. We first show that semantic segmentation has a significant reliance on the overall position channels to make predictions. We then show for the first time that it is possible to perform a `region-specific' attack, and degrade a network's performance in a particular part of the input. We believe our findings and demonstrated applications will benefit research areas concerned with understanding the characteristics of CNNs. △ Less

Submitted 17 August, 2021; originally announced August 2021.

Comments: ICCV 2021

Fibro-CoSANet: Pulmonary Fibrosis Prognosis Prediction using a Convolutional Self Attention Network

Authors: Zabir Al Nazi, Fazla Rabbi Mashrur, Md Amirul Islam, Shumit Saha

Abstract: Idiopathic pulmonary fibrosis (IPF) is a restrictive interstitial lung disease that causes lung function decline by lung tissue scarring. Although lung function decline is assessed by the forced vital capacity (FVC), determining the accurate progression of IPF remains a challenge. To address this challenge, we proposed Fibro-CoSANet, a novel end-to-end multi-modal learning-based approach, to predi… ▽ More Idiopathic pulmonary fibrosis (IPF) is a restrictive interstitial lung disease that causes lung function decline by lung tissue scarring. Although lung function decline is assessed by the forced vital capacity (FVC), determining the accurate progression of IPF remains a challenge. To address this challenge, we proposed Fibro-CoSANet, a novel end-to-end multi-modal learning-based approach, to predict the FVC decline. Fibro-CoSANet utilized CT images and demographic information in convolutional neural network frameworks with a stacked attention layer. Extensive experiments on the OSIC Pulmonary Fibrosis Progression Dataset demonstrated the superiority of our proposed Fibro-CoSANet by achieving the new state-of-the-art modified Laplace Log-Likelihood score of -6.68. This network may benefit research areas concerned with designing networks to improve the prognostic accuracy of IPF. The source-code for Fibro-CoSANet is available at: \url{https://github.com/zabir-nabil/Fibro-CoSANet}. △ Less

Submitted 12 April, 2021; originally announced April 2021.

Comments: 12 Pages

Enhanced thermo-mechanical properties of 212 MAX phase borides Zr2AB2 (A = In, Tl): an ab-initio understanding

Authors: M. A. Ali, M. M. Hossain, M. M. Uddin, A. K. M. A. Islam, S. H. Naqib

Abstract: The discovery of MAX phase borides has added a new dimension for research in the materials science community. In this paper, a first-principles study of the newly known MAX phase borides Zr2AB2 (A = In, Tl) has been carried out. The estimated lattice constants and volumes of the unit cell are found to be consistent with previous study. The dynamical and mechanical stability of the titled compounds… ▽ More The discovery of MAX phase borides has added a new dimension for research in the materials science community. In this paper, a first-principles study of the newly known MAX phase borides Zr2AB2 (A = In, Tl) has been carried out. The estimated lattice constants and volumes of the unit cell are found to be consistent with previous study. The dynamical and mechanical stability of the titled compounds have been checked. Fundamental insights into the stiffness constants, elastic moduli, hardness parameters, brittleness and anisotropy indices are presented. The Variation of these mechanical properties was explained based on the Mulliken population analysis and charge density mapping (CDM). The electronic properties have been dealt with by considering electronic band structure and density of states (DOS) which confirmed the metallic nature of Zr2AB2 (A = In, Tl). The lowly dispersive energy bands along the c-direction confirmed anisotropy in conductivity. The analysis of DOS revealed the dominant contribution from Zr-d orbitals to the conductivity with a small contribution from the In/Tl-p states contributing at the Fermi level. The Debye temperature (Theta-D), minimum thermal conductivity (Kmin), Gruneisen parameter and melting temperature (Tm) have been calculated. The higher values of Theta-D and Tm, and lower value of Kmin for Zr2AB2 (A = In, Tl) compared to those of Zr2AC (A = In, Tl). Besides, the specific heat capacities, thermal expansion coefficient, and different thermodynamic potential functions have been calculated. The optical constants have been studied to reveal their possible relevance for application purposes. The reflectivity spectra revealed the applicability of Zr2AB2 (A = In, Tl) as cover materials to diminish the solar heating. The studied physical properties of Zr2AB2 (A = In, Tl) are compared with those of other relevant 212 and 211 MAX phase nanolaminates. △ Less

Submitted 21 March, 2021; originally announced March 2021.

Comments: 33 pages, 9 figures

Journal ref: Journal of Materials Research and Technology (2021)

Origin of high hardness and optoelectronic and thermo-physical properties of boron-rich compounds B6X (X = S, Se): a comprehensive study via DFT approach

Authors: M. M. Hossain, M. A. Ali, M. M. Uddin, A. K. M. A. Islam, S. H. Naqib

Abstract: In the present study, the structural and hitherto uninvestigated mechanical (elastic stiffness constants, machinability index, Cauchy pressure, anisotropy indices, brittleness/ductility, Poissons ratio), electronic, optical, and thermodynamic properties of novel boron-rich compounds B6X (X = S, Se) have been explored using density functional theory. The estimated structural lattice parameters were… ▽ More In the present study, the structural and hitherto uninvestigated mechanical (elastic stiffness constants, machinability index, Cauchy pressure, anisotropy indices, brittleness/ductility, Poissons ratio), electronic, optical, and thermodynamic properties of novel boron-rich compounds B6X (X = S, Se) have been explored using density functional theory. The estimated structural lattice parameters were consistent with the prior report. The mechanical and dynamical stability of these compounds have been established theoretically. The materials are brittle in nature and elastically anisotropic. The value of fracture toughness, KIC for the B6S and B6Se are ~ 2.07 MPam0.5, evaluating the resistance to limit the crack propagation inside the materials. Both B6S and B6Se compounds possess high hardness values in the range 31-35 GPa, and have the potential to be prominent members of the class of hard compounds. Strong covalent bonding and sharp peak at low energy below the Fermi level confirmed by partial density of states (PDOS) resulted in the high hardness. The profile of band structure, as well as DOS, assesses the indirect semiconducting nature of the titled compounds. The comparatively high value of Debye temperature (ΘD), minimum thermal conductivity (Kmin), lattice thermal conductivity (kph), low thermal expansion coefficient, and low density suggest that both boron-rich chalcogenides might be used as thermal management materials. Large absorption capacities in the mid ultraviolet region (3.2-15 eV) of the studied materials and low reflectivity (~16 %) are significantly noted. Such favorable features give promise to the compounds under investigation to be used in UV surface-disinfection devices as well as medical sterilizer equipment applications. Excellent correlations are found among all the studied physical properties of these compounds. △ Less

Submitted 31 January, 2021; originally announced February 2021.

Comments: 35 pages, 12 figures, 8 Tables (will be submitted for publication in reputed journal)

Position, Padding and Predictions: A Deeper Look at Position Information in CNNs

Authors: Md Amirul Islam, Matthew Kowal, Sen Jia, Konstantinos G. Derpanis, Neil D. B. Bruce

Abstract: In contrast to fully connected networks, Convolutional Neural Networks (CNNs) achieve efficiency by learning weights associated with local filters with a finite spatial extent. An implication of this is that a filter may know what it is looking at, but not where it is positioned in the image. In this paper, we first test this hypothesis and reveal that a surprising degree of absolute position info… ▽ More In contrast to fully connected networks, Convolutional Neural Networks (CNNs) achieve efficiency by learning weights associated with local filters with a finite spatial extent. An implication of this is that a filter may know what it is looking at, but not where it is positioned in the image. In this paper, we first test this hypothesis and reveal that a surprising degree of absolute position information is encoded in commonly used CNNs. We show that zero padding drives CNNs to encode position information in their internal representations, while a lack of padding precludes position encoding. This gives rise to deeper questions about the role of position information in CNNs: (i) What boundary heuristics enable optimal position encoding for downstream tasks?; (ii) Does position encoding affect the learning of semantic representations?; (iii) Does position encoding always improve performance? To provide answers, we perform the largest case study to date on the role that padding and border heuristics play in CNNs. We design novel tasks which allow us to quantify boundary effects as a function of the distance to the border. Numerous semantic objectives reveal the effect of the border on semantic representations. Finally, we demonstrate the implications of these findings on multiple real-world tasks to show that position information can both help or hurt performance. △ Less

Submitted 28 January, 2021; originally announced January 2021.

Shape or Texture: Understanding Discriminative Features in CNNs

Authors: Md Amirul Islam, Matthew Kowal, Patrick Esser, Sen Jia, Bjorn Ommer, Konstantinos G. Derpanis, Neil Bruce

Abstract: Contrasting the previous evidence that neurons in the later layers of a Convolutional Neural Network (CNN) respond to complex object shapes, recent studies have shown that CNNs actually exhibit a `texture bias': given an image with both texture and shape cues (e.g., a stylized image), a CNN is biased towards predicting the category corresponding to the texture. However, these previous studies cond… ▽ More Contrasting the previous evidence that neurons in the later layers of a Convolutional Neural Network (CNN) respond to complex object shapes, recent studies have shown that CNNs actually exhibit a `texture bias': given an image with both texture and shape cues (e.g., a stylized image), a CNN is biased towards predicting the category corresponding to the texture. However, these previous studies conduct experiments on the final classification output of the network, and fail to robustly evaluate the bias contained (i) in the latent representations, and (ii) on a per-pixel level. In this paper, we design a series of experiments that overcome these issues. We do this with the goal of better understanding what type of shape information contained in the network is discriminative, where shape information is encoded, as well as when the network learns about object shape during training. We show that a network learns the majority of overall shape information at the first few epochs of training and that this information is largely encoded in the last few layers of a CNN. Finally, we show that the encoding of shape does not imply the encoding of localized per-pixel semantic information. The experimental results and findings provide a more accurate understanding of the behaviour of current CNNs, thus helping to inform future design choices. △ Less

Submitted 27 January, 2021; originally announced January 2021.

Comments: Accepted to ICLR 2021

A Comparative Study of AHP and Fuzzy AHP Method for Inconsistent Data

Authors: Md. Ashek-Al-Aziz, Sagar Mahmud, Md. Azizul Islam, Jubayer Al Mahmud, Khan Md. Hasib

Abstract: In various cases of decision analysis we use two popular methods: Analytical Hierarchical Process (AHP) and Fuzzy based AHP or Fuzzy AHP. Both the methods deal with stochastic data and can determine decision result through Multi Criteria Decision Making (MCDM) process. Obviously resulting values of the two methods are not same though same set of data is fed into them. In this research work, we hav… ▽ More In various cases of decision analysis we use two popular methods: Analytical Hierarchical Process (AHP) and Fuzzy based AHP or Fuzzy AHP. Both the methods deal with stochastic data and can determine decision result through Multi Criteria Decision Making (MCDM) process. Obviously resulting values of the two methods are not same though same set of data is fed into them. In this research work, we have tried to observe similarities and dissimilarities between two methods outputs. Almost same trend or fluctuations in outputs have been seen for both methods for same set of input data which are not consistent. Both method outputs ups and down fluctuations are same for fifty percent cases. △ Less

Submitted 23 December, 2020; originally announced January 2021.

Comments: 22 Pages, 9 Figures

Journal ref: International Journal of Sciences: Basic and Applied Research (IJSBAR), Volume 54 Issue 4, Year 2020, Page - 16 -37