Climate change presents significant challenges to ecosystems, economies, and societies globally. In Malaysia, a tropical country highly dependent on its natural resources, the impacts are evident in altered rainfall patterns, rising temperatures, and extreme weather events. Despite these challenges, many studies still predominantly rely on traditional statistical methods, which limit their capacity for making accurate climate predictions and developing effective policy solutions.This study effectively addresses the existing gap in research by analyzing extensive historical climate data using advanced machine learning (ML) techniques. The primary focus is on accurately forecasting trends in both precipitation patterns and surface air temperature fluctuations. Performance measures like Mean Absolute Error (MAE), Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) are used to assess three ML models: Support Vector Regression (SVR), Random Forest Regression (RFR) and Linear Regression (LR). The findings demonstrate that LR performs better than the other models in forecasting patterns of precipitation and temperature. The results suggest a significant increase in temperature and unpredictable patterns of precipitation, and that poses major implications for agriculture, infrastructure resilience, and water management. Malaysia's climate resilience is improved by this research, which promotes data-driven policymaking by assessing current climate adaptation methods and offering practical ideas.

Browser fingerprinting is the identification of a browser through the network traffic captured during communication between the browser and server. This can be done using the HTTP protocol, browser extensions, and other methods. This paper discusses browser fingerprinting using the HTTPS over TLS 1.3 protocol. The study observed that different browsers use a different number of messages to communicate with the server, and the length of messages also varies. To conduct the study, a network was set up using a UTM hypervisor with one virtual machine as the server and another as a VM with a different browser. The communication was captured, and it was found that there was a $30 \%-35 \%$ dissimilarity in the behavior of different browsers.

A collaborative real-time text editor is an application that allows multiple users to edit a document simultaneously and merge their contributions automatically.It can be made collaborative by implementing a conflict resolution algorithm either on the client side (in peerto-peer collaboration) or on the server side (when using web sockets and a central server to monitor state changes).Although web sockets are ideal for real-time text editors, using multiple collaborative editors on one connection can create problems.This is because a single web connection cannot monitor which user is collaborating on which application state, leading to unnecessary network queries and data being delivered to the wrong state.To address this issue, the current solution is to open multiple web socket connections, with one web socket per collaboration application.However, this can add significant overhead proportional to the number of apps utilized.In this study, we demonstrate an algorithm that enables using a single web socket for multiple collaborative applications in a collaborative editor.Our method involves modifying the socket's code to track which application's shared state is being worked on and by whom.This allows for the simultaneous collaboration of multiple states in real-time, with infinite users, without opening a different socket for each application.Our optimized editor showed an efficiency improvement of over 96% in access time duration.This approach can be implemented in other collaborative editors and web applications with similar architecture to improve performance and eliminate issues arising from network overload.

WiFi and private 5G networks, commonly referred to as P5G, provide Internet of Things (IoT) devices the ability to communicate at fast speeds, with low latency and with a high capacity. Will they coexist and share the burden of delivering a connection to devices at home, on the road, in the workplace, and at a park or a stadium? Or will one replace the other to manage the increase in endpoints and traffic in the enterprise, campus, and manufacturing environments? In this research, we describe IoT device testbeds to collect network traffic in a local area network and cyberspace including beyond 5G/6G network traffic traces at different layers. We also describe research problems and challenges, such as traffic classification and traffic prediction by the traffic traces of devices. An AI-enabled hierarchical learning architecture for the problems above using sources like network packets, frames, and signals from the traffic traces with machine learning models is also presented.

Over the years, space systems have evolved considerably to provide high-quality services for demanding applications such as navigation, communication, and weather forecast. Modern space systems rely on extremely fast commercially available off-the-shelf (COTS) processing units, with built-in GPU, DSP, and FPGA in light-weight, energy-efficient hardware. Since such devices are not necessarily designed with security features as a priority, there must be an adaptive controller to protect this mission-critical space system from potential malicious attacks, such as memory leaks, packet drops, algorithmic trojans, and so on. These attacks can lead the system to substantial inefficiency or complete failure. Considering the hardware diversity in current space systems, we propose a framework to explore both diversity and redundancy not only of hardware but also of software to make the overall system fault-tolerant. Our approach deploys mechanisms for monitoring and orchestrating actions of redundancy, diversity, and randomization to render the system resilient unpredictably dynamic, and optimize efficiency as much as possible during abnormalities. Yet, we use rule-based and adaptive engines to keep track of the various computing units to learn the best strategies to take when the system is under attack. The robustness of our approach lies in the fact that it makes the system highly unpredictable to potential attackers and tolerates attacks to some extent, which is crucial for any mission-critical application.

Agriculture contributes a significant amount to the economy of India due to the dependence on human beings for their survival. The main obstacle to food security is population expansion leading to rising demand for food. Farmers must produce more on the same land to boost the supply. Through crop yield prediction, technology can assist farmers in producing more. This paper's primary goal is to predict crop yield utilizing the variables of rainfall, crop, meteorological conditions, area, production, and yield that have posed a serious threat to the long-term viability of agriculture. Crop yield prediction is a decision-support tool that uses machine learning and deep learning that can be used to make decisions about which crops to produce and what to do in the crop's growing season. It can decide which crops to produce and what to do in the crop's growing season. Regardless of the distracting environment, machine learning and deep learning algorithms are utilized in crop selection to reduce agricultural yield output losses. To estimate the agricultural yield, machine learning techniques: decision tree, random forest, and XGBoost regression; deep learning techniques - convolutional neural network and long-short term memory network have been used. Accuracy, root mean square error, mean square error, mean absolute error, standard deviation, and losses are compared. Other machine learning and deep learning methods fall short compared to the random forest and convolutional neural network. The random forest has a maximum accuracy of 98.96%, mean absolute error of 1.97, root mean square error of 2.45, and standard deviation of 1.23. The convolutional neural network has been evaluated with a minimum loss of 0.00060. Consequently, a model is developed that, compared to other algorithms, predicts the yield quite well. The findings are then analyzed using the root mean square error metric to understand better how the model's errors compare to those of the other methods.

Conversational artificial intelligence (AI) is a type of artificial intelligence that uses machine learning techniques to understand and respond to user inputs. This paper presents a conversational chat system that uses an attention mechanism to respond to COVID-19 inquiries. The model is based on the Luong Attention Mechanism’s three scoring methodologies the Dot Attention Mechanism, the General Attention Mechanism, and the Concat Attention Mechanism. The results show that the accuracy of the dot attention mechanism is highest and is 87% when the test questions were obtained directly from the database, as determined by an examination of the results, compared to 38% when the attention mechanism is not used. Furthermore, when the questions are asked with natural variations, human verification accuracy is 63% compared to 16% when the attention mechanism is not used. The research suggests that chatbots can be used everywhere due to their accuracy and accessibility around the clock.

Artificial Intelligence (AI) development has encouraged many new research areas, including AI-enabled Internet of Things (IoT) network. AI analytics and intelligent paradigms greatly improve learning efficiency and accuracy. Applying these learning paradigms to network scenarios provide technical advantages of new networking solutions. In this paper, we propose an improved approach for IoT security from data perspective. The network traffic of IoT devices can be analyzed using AI techniques. The Adversary Learning (AdLIoTLog) model is proposed using Recurrent Neural Network (RNN) with attention mechanism on sequences of network events in the network traffic. We define network events as a sequence of the time series packets of protocols captured in the log. We have considered different packets TCP packets, UDP packets, and HTTP packets in the network log to make the algorithm robust. The distributed IoT devices can collaborate to cripple our world which is extending to Internet of Intelligence. The time series packets are converted into structured data by removing noise and adding timestamps. The resulting data set is trained by RNN and can detect the node pairs collaborating with each other. We used the BLEU score to evaluate the model performance. Our results show that the predicting performance of the AdLIoTLog model trained by our method degrades by 3-4% in the presence of attack in comparison to the scenario when the network is not under attack. AdLIoTLog can detect adversaries because when adversaries are present the model gets duped by the collaborative events and therefore predicts the next event with a biased event rather than a benign event. We conclude that AI can provision ubiquitous learning for the new generation of Internet of Things.

Received Oct 2, 2021 Revised Apr 4, 2022 Accepted Apr 28, 2022 Half of the entire social media users in Indonesia has experienced cyberbullying. Cyberbullying is one of the treatments received as an attack with abusive words. An abusive word is a word or phrase that contained harassment and is expressed be it spoken or in the form of text. This is a serious problem that must be controlled because the act has an impact on the victim's psychology and causes trauma resulting in depression. This study proposed to identify abusive comments from social media in Indonesian language using a deep learning approach. The architecture used is a hybrid model, a combination between recurrent neural network (RNN) and long short-term memory (LSTM). RNN can map the input sequences to fixed-size vectors on hidden vector components and LSTM implemented to overcome gradient vector growth components that have the potential to exist in RNN. The steps carried out include preprocessing, modelling, implementation, and evaluation. The dataset used is indonesian abusive and hate speech from Twitter data. The evaluation result showed that the model proposed produced an f-measure value of 94% with an increase in accuracy of 23%.

Device fingerprinting is a problem of identifying a network device using network traffic data to secure against cyber-attacks. Automated device classification from a large set of network traffic features space is challenging for the devices connected in the cyberspace. In this work, the idea is to define a device-specific unique fingerprint by analysing solely inter-arrival time of packets as a feature to identify a device. Neural networks are the universal function approximation which learn abstract, high-level, nonlinear representation of training data. Deep convolution neural network is used on images of inter-arrival time signature for device fingerprinting of 58 non-IoT devices of 5-11 types. To evaluate the performance, we compared ResNet-50 layer and basic CNN-5 layer architectures. We observed that device type identification models perform better than device identification. We also found that when deep learning models are attacked over device signature, the models identify the change in signature, and classify the device in the wrong class thereby the classification performance of the models degrades. The performance of the models to detect the attacks are significantly different from each other though both models indicate the system under attack.

Received Aug 22, 2021 Revised Jun 30, 2022 Accepted Jul 11, 2022 Media news are making a large part of public opinion and, therefore, must not be fake. News on web sites, blogs, and social media must be analyzed before being published. In this paper, we present linguistic characteristics of media news items to differentiate between fake news and real news using machine learning algorithms. Neural fake news generation, headlines created by machines, semantic incongruities in text and image captions generated by machine are other types of fake news problems. These problems use neural networks which mainly control distributional features rather than evidence. We propose applying correlation between features set and class, and correlation among the features to compute correlation attribute evaluation metric and covariance metric to compute variance of attributes over the news items. Features unique, negative, positive, and cardinal numbers with high values on the metrics are observed to provide a high area under the curve (AUC) and F1-score.

Media news are making a large part of public opinion and, therefore, must not be fake. News on web sites, blogs, and social media must be analyzed before being published. In this paper, we present linguistic characteristics of media news items to differentiate between fake news and real news using machine learning algorithms. Neural fake news generation, headlines created by machines, semantic incongruities in text and image captions generated by machine are other types of fake news problems. These problems use neural networks which mainly control distributional features rather than evidence. We propose applying correlation between features set and class, and correlation among the features to compute correlation attribute evaluation metric and covariance metric to compute variance of attributes over the news items. Features unique, negative, positive, and cardinal numbers with high values on the metrics are observed to provide a high area under the curve (AUC) and F1-score.

Despite substantial advances in network architecture performance, the susceptibility of adversarial attacks makes deep learning challenging to implement in safety-critical applications. This paper proposes a data-centric approach to addressing this problem. A nonlocal denoising method with different luminance values has been used to generate adversarial examples from the Modified National Institute of Standards and Technology database (MNIST) and Canadian Institute for Advanced Research (CIFAR-10) data sets. Under perturbation, the method provided absolute accuracy improvements of up to 9.3% in the MNIST data set and 13% in the CIFAR-10 data set. Training using transformed images with higher luminance values increases the robustness of the classifier. We have shown that transfer learning is disadvantageous for adversarial machine learning. The results indicate that simple adversarial examples can improve resilience and make deep learning easier to apply in various applications.

Delay tolerant networks (DTNs) have been standardised as a solution for wireless networking scenarios with intermittent connectivity. Limited buffer space and limited battery power in DTNs can give rise to malicious nodes. These malicious nodes misuse network resources and exhibit malicious behaviour which can cripple the network. This paper discusses various forms of malicious behaviour: flooding attack, black hole attack, and selfish attack. Several reputation based defending mechanisms against these attacks have been proposed by the DTN research community. In this paper, a taxonomy of these reputation-based defending mechanisms is defined, in terms of how reputations are calculated and disseminated. The mechanisms are categorised as source-based, peer-based, trusted-authority based or destination-based, depending on which node takes the decision to assign reputation. Under each category, the mechanisms are further classified as either node-centric (the node itself), or node-in-contact-centric (contacting node), based on which node keeps evidence of the job performed. This paper presents a review of all these reputation mechanisms and compare them in terms of computation and other parameters. Destination-based reputation mechanism seems to be a better approach.

With the growth of wireless network technology-based devices, identifying the communication behaviour of wireless connectivity enabled devices, e.g. Internet of Things (IoT) devices, is one of the vital aspects, in managing and securing IoT networks. Initially, devices use frames to connect to the access point on the local area network and then, use packets of typical communication protocols through the access point to communicate over the Internet. Toward this goal, network packet and IEEE 802.11 media access control (MAC) frame analysis may assist in managing IoT networks efficiently, and allow investigation of inclusive behaviour of IoT devices. This paper presents network traffic traces data of D-Link IoT devices from packet and frame levels. Data collection experiment has been conducted in the Network Systems and Signal Processing (NSSP) laboratory at Universiti Brunei Darussalam (UBD). All the required devices, such as IoT devices, workstation, smartphone, laptop, USB Ethernet adapter, and USB WiFi adapter, have been configured accordingly, to capture and store network traffic traces of the 14 IoT devices in the laboratory. These IoT devices were from the same manufacture (D-Link) with different types, such as camera, home-hub, door-window sensor, and smart-plug.

Received Aug 22, 2021 Revised May 20, 2022 Accepted Jun 6, 2022 Despite substantial advances in network architecture performance, the susceptibility of adversarial attacks makes deep learning challenging to implement in safety-critical applications. This paper proposes a data-centric approach to addressing this problem. A nonlocal denoising method with different luminance values has been used to generate adversarial examples from the Modified National Institute of Standards and Technology database (MNIST) and Canadian Institute for Advanced Research (CIFAR-10) data sets. Under perturbation, the method provided absolute accuracy improvements of up to 9.3% in the MNIST data set and 13% in the CIFAR10 data set. Training using transformed images with higher luminance values increases the robustness of the classifier. We have shown that transfer learning is disadvantageous for adversarial machine learning. The results indicate that simple adversarial examples can improve resilience and make deep learning easier to apply in various applications.

Transfer learning allows us to exploit knowledge gained from one task to assist in solving another but relevant task. In modern computer vision research, the question is which architecture performs better for a given dataset. In this paper, we compare the performance of 14 pre-trained ImageNet models on the histopathologic cancer detection dataset, where each model has been configured as a naive model, feature extractor model, or fine-tuned model. Densenet161 has been shown to have high precision whilst Resnet101 has a high recall. A high precision model is suitable to be used when follow-up examination cost is high, whilst low precision but a high recall/sensitivity model can be used when the cost of follow-up examination is low. Results also show that transfer learning helps to converge a model faster.

Received Aug 22, 2021 Revised May 20, 2022 Accepted Jun 6, 2022 Despite substantial advances in network architecture performance, the susceptibility of adversarial attacks makes deep learning challenging to implement in safety-critical applications. This paper proposes a data-centric approach to addressing this problem. A nonlocal denoising method with different luminance values has been used to generate adversarial examples from the Modified National Institute of Standards and Technology database (MNIST) and Canadian Institute for Advanced Research (CIFAR-10) data sets. Under perturbation, the method provided absolute accuracy improvements of up to 9.3% in the MNIST data set and 13% in the CIFAR10 data set. Training using transformed images with higher luminance values increases the robustness of the classifier. We have shown that transfer learning is disadvantageous for adversarial machine learning. The results indicate that simple adversarial examples can improve resilience and make deep learning easier to apply in various applications.

Device identification is the process of identifying a device on Internet without using its assigned network or other credentials. The sharp rise of usage in Internet of Things (IoT) devices has imposed new challenges in device identification due to a wide variety of devices, protocols and control interfaces. In a network, conventional IoT devices identify each other by utilizing IP or MAC addresses, which are prone to spoofing. Moreover, IoT devices are low power devices with minimal embedded security solution. To mitigate the issue in IoT devices, fingerprint (DFP) for device identification can be used. DFP identifies a device by using implicit identifiers, such as network traffic (or packets), radio signal, which a device used for its communication over the network. These identifiers are closely related to the device hardware and software features. In this paper, we exploit TCP/IP packet header features to create a device fingerprint utilizing device originated network packets. We present a set of three metrics which separate some features from a packet which contribute actively for device identification. To evaluate our approach, we used publicly accessible two datasets. We observed the accuracy of device genre classification 99.37% and 83.35% of accuracy in the identification of an individual device from IoT Sentinel dataset. However, using UNSW dataset device type identification accuracy reached up to 97.78%.

Machine translation has many applications such as news translation, email translation, official letter translation etc. Commercial translators, e.g. Google Translation lags in regional vocabulary and are unable to learn the bilingual text in the source and target languages within the input. In this paper, a regional vocabulary-based application-oriented Neural Machine Translation (NMT) model is proposed over the data set of emails used at the University for communication over a period of three years. A state-of-the-art Sequence-to-Sequence Neural Network for ML -> EN and EN -> ML translations is compared with Google Translate using Gated Recurrent Unit Recurrent Neural Network machine translation model with attention decoder. The low BLEU score of Google Translation in comparison to our model indicates that the application based regional models are better. The low BLEU score of EN -> ML of our model and Google Translation indicates that the Malay Language has complex language features corresponding to English.