Network traffic classification is crucial for both network security and management. Despite advances in deep learning-based multi-task traffic classification, existing models often struggle to jointly handle multiple tasks while providing interpretable insights. In multi-task scenarios, different tasks rely on distinct regions of the traffic sequence, motivating the use of dynamic and interpretable attention mechanisms. To this end, we propose Dynamic Quantized Self-Attention (DQSA), a unified framework specifically designed for multi-task network traffic classification. At its core, the Task Gated Attention Router (TGAR) dynamically associates attention heads with different tasks, enabling adaptive focus on task-specific patterns. This mechanism provides interpretable attention scores, which help analyze misclassifications and guide further model refinement. To improve efficiency and handle diverse network traffic features, we introduce the Soft Quantized Self-Attention Head (SQ-SAH) to reduce computational complexity and extend the Rotary Position Embedding (RoPE) to accommodate these features. Extensive experiments on ISCX VPN-NonVPN and DCI-LTE datasets demonstrate that DQSA consistently outperforms state-of-the-art baselines, achieving 92.85% accuracy on the encapsulation-level task of ISCX VPN-NonVPN and 93.17% accuracy on the application-level task of DCI-LTE, surpassing the strongest existing methods by up to 2.65%, while providing interpretable task-specific attention for efficient multi-task network traffic classification.

Academic conferences serve as a useful platform for researchers and educators to share their work and ideas. Upon the paper acceptance, authors need to register their papers, pay registration fees, and present their work at conferences. International conferences often involve multiple parties of users and cross-border payments with multiple channels that incur additional service charges. Conference organisers and reviewers typically contribute voluntarily without any monetary rewards. The emergence of Web 3.0 technology, leveraging the decentralized and secure nature of blockchain, presents an opportunity for this domain. This research creates a decentralized academic conference management system using Web 3.0 (ConfMan Web 3.0) and Rust programming language that simplifies cross-border payments on different channels by using cryptocurrencies instead of fiat currency, thereby reducing service charges. It aims to provide rewards, recognition of contributions, and consolidated historical records to all conference contributors, including authors, reviewers, programme chairs, and organising committees. The Solana blockchain is used to store conference- related data, and a web application is developed for ConfMan Web 3.0. Various testings are conducted to evaluate its performance. The findings highlight the potential of Web 3.0 technology in transforming the academic conference management landscape.

The goal of the feature reconstruction network based on an autoencoder in the training phase is to force the network to reconstruct the input features well. The network tends to learn shortcuts of “identity mapping,” which leads to the network outputting abnormal features as they are in the inference phase. As such, the abnormal features based on reconstruction error cannot be distinguished from normal features, significantly limiting the detection performance of such methods. To address this issue, we propose a feature transformation reconstruction (FTR) network, which can avoid the identity mapping problem. Specifically, we use a normalizing flow model as a feature transformation (FT) network to transform input features into other forms. The training goal of the feature reconstruction (FR) network is no longer to reconstruct the input features but to reconstruct the transformed features, effectively avoiding the shortcut of learning the “identity map.” Furthermore, this paper proposes a masked convolutional attention (MCA) module, which randomly masks the input features in the training phase and reconstructs the input features in a self‐supervised manner. In the testing phase, the MCA can effectively suppress the excessive reconstruction of abnormal features and further improve anomaly detection performance. FTR achieves the scores of the area under the receiver operating characteristic curve (AUROC) at 99.5% and 97.8% on the MVTec AD and BTAD datasets, respectively, outperforming other state‐of‐the‐art methods. Moreover, FTR is faster than the existing methods, with a high speed of 137 frames per second (FPS) on a 3080ti GPU.

This Innovative Practice category full paper describes the strategy of peer evaluation approach. Effective team management is pivotal in the success of any collaborative project, particularly in the domain of learning Software Engineering or Professional Software Development courses. Peer evaluations are commonly used methods to motivate group-based learning. This paper addresses the significance of team management and peer evaluations in the context of two Computing Science (CS) courses offered: Professional Software Development (PSD) and Team Project (TP). The main contributions of this paper are to provide nuanced insights into the challenges and opportunities inherent in collaborative endeavors, emphasizing effective motivation, encouragement, and resolution strategies for tackling teamwork issues within group projects. This study scrutinizes the intricate dimensions of teamwork in Software Engineering education. A central aspect of the paper involves a comprehensive analysis of peer evaluation methodologies, intending to shed light on practical implications for educators and learners. The findings underscore the importance of leveraging these insights to cultivate a conducive learning environment, motivating students to actively participate in collaborative endeavors. The proposed peer evaluations approaches are presented as a combination of the existing quantitative assessment approach and newly qualitative feedback mechanisms. It incorporates these elements: weekly peer review, emotion rating & thoughts, release of feedback, and a centralized dashboard. In the study, 31 CS Sophomore students taking the CSC2101 - PSD and TP1 course participate to evaluation experiments, where positive results are obtained. The collective findings and evaluations contribute to the ongoing refinement of collaborative learning experiences within the dynamic landscape of Software Engineering education.

This Innovative Practice category full paper proposes an innovative approach to enhance peer evaluations in group-based learning. It aims at enhancing the overall peer assessment experience for students in software engineering education. In the software industry, it is all about teamwork for software development projects. There is a need to nurture students to work collaboratively in group-based learning. The collaborative milieu inherent in group projects has long been acknowledged within the sphere of software engineering education. In our university, students are cultivated the effective teamwork within the purview of CSC2101 - Professional Software Development and Team Projects (PSD & TP1), an undergraduate course in the domain of software engineering. The terrain of software group projects still presents its own array of obstacles, such as unequal contributions, internal conflicts, peer evaluation mechanism, etc. This paper first discusses the limitations of the current peer evaluation methods in this course. Next, the proposed methodology is presented that embodies four key elements: transparency, anonymity, structured assessment, and qualitative feedback. The implementation details of the proposed approaches are then described consisting of the following features and mechanisms: online platform, anonymous profiles, public results, and structured evaluation forms. The rationales for the proposed peer evaluation techniques are discussed. The experiment studies are conducted involving 40 students in our university. It includes numerous characteristics of the peer evaluation experience in order to gain a full picture of existing peer evaluation approaches within software engineering education. The survey results show a generally positive mentality toward the proposed features.

Research outcomes have been reported to various industrial applications of Artificial Intelligence generated content (AIGC). This paper explores AI assisted creative production of storyboard and game characters in the game industry, by leveraging Stable Diffusion 1.5 coupled with Control Nets (SD-1.5+CN) for image generations and Generative Pretrained Transformer (GPT) of OpenAI for narrative generations. Experiment evaluations are conducted for visual appeal and narrative flows of the generated contents using SketchBoard. The results illustrate SketchBoard with weighted prompts achieving a higher Contrastive Language-Image Pre-Training (CLIP) score than the baseline models. In addition to reducing the time and effort required for manual creation, it obtains strong alignments between the generated images and provided text prompts. User study is conducted where the user feedback indicates a high level of satisfaction, with 92 % of participants expressing enj oyment and contentment with SketchBoard. Its ability to generate visually appealing storyboard frames with coherent narratives offers opportunities for efficient storyboard creation, paving the way for future advancements in automated storyboard generation.

The widespread use of encrypted communication in mobile networks poses significant challenges in accurately classifying traffic. Detecting out-of-distribution (OOD) samples, which significantly deviate from known classes, adds complexity to the task. This paper proposes a feature analysis-based OOD detection scheme for traffic classification in Long-Term Evolution (LTE) systems. Our method utilizes Long Short-Term Memory (LSTM) networks for feature extraction, capturing the feature vectors of the traffic series. Principal Component Analysis (PCA) is then applied to obtain principal and residual principal components. Leveraging the residual feature vector, we construct an OOD score to quantify deviation from the ID dataset. Extensive experiments on a large-scale encrypted mobile traffic dataset demonstrate the superiority of our approach, achieving high accuracy in OOD detection compared to existing techniques. Our method contributes to enhanced security and reliable traffic classification in LTE systems, addressing challenges posed by OOD samples.

Deep learning enables effective predictions. But deep structures face some challenges on human interpretability compared to conventional techniques, e.g., fuzzy inference systems. It motivates more research works to alleviate the black box nature of deep structures with performance maintained. This paper proposes a fuzzy-embedded recurrent neural network (FE-RNN) to improve interpretability of the underlying neural networks. It is a parallel deep structure comprising an RNN and a Pseudo Outer-Product based Fuzzy Neural Network (POPFNN) that share a common set of input and output linguistic concepts. The inference processes undertaken are associated by RNN using fuzzy rules in the embedded POPFNN. Fuzzy IF-THEN rules provide better interpretability of the inference process of the hybrid networks. It allows an effective realisation of a data driven implication using RNN in the modelling of fuzzy entailment within a fuzzy neural networks (FNN) structure. FE-RNN obtains more consistent results than other FNN in the experiment using the Mackey-Glass dataset. FE-RNN achieves about 99% correlation for forecasting prices of market indexes. Its interpretability is also discussed. FE-RNN then acts as a prediction tool in a financial trading system using forecast-assisted technical indicators optimised with Genetic Algorithms. It outperforms the benchmark trading strategies in the trading experiments.

Teamwork in software development life cycle (SDLC) and Software Engineering (SE) is a cooperative process that all Computing Science (CS) undergraduates need to undergo. It is a critical skill for the industry and is usually trained through group projects in Higher Education. Due to the nature of software development, most software projects involve collaborative efforts of a group of developers. Although teamwork has been studied in many prior works, it is still considered as a dynamic element in SDLC. As the level of complexity, type of deliverables and range of stakeholders in software projects can vary widely, prior experience cannot be applied directly to new projects. The current implementation of SE education in the Professional Software Development (PSD) and Team Project (TP) subjects contains elements to promote teamwork. Students are required to work in groups on real-world problems. This paper examines the current teamwork simulating real-world software projects through an evaluation with the existing and previous cohorts of students, who have experienced the PSD and TP subjects. Several improvements are then proposed by this study. Based on the results, the majority of the respondents agree that our proposed methods such as self-selection of groups, pair programming, and prototyping model will bring about improved teamwork in their group projects.

In computing science (CS) education, software programming and software engineering subjects are parts of core CS subjects where students learn various programming languages and software development life cycle (SDLC). With the aim of equipping students with the necessary project skills and experiences essential for workplaces, Institutes of Higher Learning (IHL) have increasingly adopted the educational practice of bringing real-world team-based projects in the curriculum. Such software projects commonly support the use of Agile process practices, which place more emphasis on communications and teamwork among students in a team, stakeholders and industry customers. This could be manageable for small scale projects where student teams are in their learning journey with active and continuous communications with customers, while developing the necessary cohesion within the team to facilitate incremental deliverables of their software products. However, teamwork remains a bottleneck that is difficult to be consistently realized among members in group projects. It is not easy to motivate and measure all team members working towards the common goals or complete project tasks on time. As such, this paper proposes the Weekly Team Assessment (WTA) framework to continuously encourage teamwork throughout the software project development duration. We also propose the settings of dedicated leadership and constructive feedback iteratively in the WTA framework. The proposed methods have been applied and evaluated through the experiments in two software engineering modules, Professional Software Development and Team Project, involving 50 sophomore CS students who are divided into control group and experimental group equally. Experiment results are analyzed and discussed in this paper. Positive outcomes are observed from the experiment where about 72% of participants appreciate the proposed methods in improving their teamwork in group projects. The degree of effectiveness of the proposed methods is further evaluated from three aspects in the experiment.

Classification of traffic service types is a valuable function for wireless communication networks. Even though some progress has been made, the recognition of the type of the traffic services cannot be done in real time. In this paper, we propose a novel method for classifying traffic series in real time based on transfer learning techniques. We pre-train a deep learning model with long traffic series and fine-tune the model with short traffic series. In this way, the developed model achieves the capability of recognising traffic services in real time. In other words, the model can recognize traffic services by using short traffic series. We collect Downlink Control Information (DCI) from commercial LTE networks when using five common types of traffic services. Then we use the dataset to validate our method. Our experimental results show that, by using proposed method, LSTM accuracy rates will increase to 80% and 88.5% when the length of the traffic series is 5 seconds and 10 seconds respectively, which is higher than the baseline. The strategy is also suitable for one dimension convolution neural network (1D-CNN).

Some common technical indicators, such as moving average convergence divergence (MACD), relative strength index (RSI), and MACD histogram (MACDH) are used in technical analyses and stock trading. However, some of them are lagging indicators, affecting the effectiveness in the stock trading and portfolio management. A forecasted MACDH (fMACDH) indicator for predicting next day price by a neuro-fuzzy network, Self-reorganizing Fuzzy Associative Machine (SeroFAM) which has been reported in the prior research work. In order to further reduce the lagging effect, two trading indicators are proposed in this paper: the optimised fMACDH indicator and the fMACDH-fRSI indicator. The optimised fMACDH indicator is derived to extend price forecasting to 1–5 days ahead as the prediction depth, using 1–5 days of historical price data as the input depth. The fMACDH-fRSI indicator is derived by combining the optimized fMACDH indicator and the forecasted RSI (fRSI) indicator. A genetic algorithm (GA) and the fitness functions are designed with the SeroFAM in this paper, which are utilised for optimising parameters of these two proposed indicators. Experiments have been conducted to evaluate and benchmark of the proposed trading indicators optimised by the GA. Two rule-based portfolio rebalancing algorithms are then proposed using the optimised fMACDH trading indicator tuned by the GA: the Tactical Buy and Hold (TBH) and the Rule-Based Business Cycle (RBBC) portfolio rebalancing algorithms. The TBH algorithm takes advantage of relative differences in risk levels to perform rebalancing during trend reversals. The RBBC portfolio rebalancing algorithm takes advantage of the offsets between the business cycles of different market sectors. Experiments have been conducted to evaluate the performance of both algorithms using two sets of portfolios consisting of different assets. The TBH portfolio rebalancing algorithm outperforms the equally weighted portfolio strategy by about 26 % − 27 %; as well outperforms the Buy and Hold strategy by 5 % − 40 %. The RBBC portfolio rebalancing algorithm outperforms the equally weighted portfolio strategy by 54 % − 55 %; it also outperforms 12 out of the 13 assets with the Buy and Hold strategy, by an average performance of about 166 %. The results are highly encouraging with consistent performances achieved in dynamic portfolio rebalancing.

In this paper, we develop the weighted error entropy-regularized tensor learning method for multi-view subspace clustering (WETMSC), which integrates the noise disturbance removal and subspace structure discovery into one unified framework. Unlike most existing methods which focus only on the affinity matrix learning for the subspace discovery by different optimization models and simply assume that the noise is independent and identically distributed (i.i.d.), our WETMSC method adopts the weighted error entropy to characterize the underlying noise by assuming that noise is independent and piecewise identically distributed (i.p.i.d.). Meanwhile, WETMSC constructs the self-representation tensor by storing all self-representation matrices from the view dimension, preserving high-order correlation of views based on the tensor nuclear norm. To solve the proposed nonconvex optimization method, we design a half-quadratic (HQ) additive optimization technology and iteratively solve all subproblems under the alternating direction method of multipliers framework. Extensive comparison studies with state-of-the-art clustering methods on real-world datasets and synthetic noisy datasets demonstrate the ascendancy of the proposed WETMSC method.

Multi-view subspace clustering aims to integrate the complementary information contained in different views to facilitate data representation. Currently, low-rank representation (LRR) serves as a benchmark method. However, we observe that these LRR-based methods would suffer from two issues: limited clustering performance and high computational cost since (1) they usually adopt the nuclear norm with biased estimation to explore the low-rank structures; (2) the singular value decomposition of large-scale matrices is inevitably involved. Moreover, LRR may not achieve low-rank properties in both intra-views and inter-views simultaneously. To address the above issues, this paper proposes the Bi-nuclear tensor Schatten- p norm minimization for multi-view subspace clustering (BTMSC). Specifically, BTMSC constructs a third-order tensor from the view dimension to explore the high-order correlation and the subspace structures of multi-view features. The Bi-Nuclear Quasi-Norm (BiN) factorization form of the Schatten- p norm is utilized to factorize the third-order tensor as the product of two small-scale third-order tensors, which not only captures the low-rank property of the third-order tensor but also improves the computational efficiency. Finally, an efficient alternating optimization algorithm is designed to solve the BTMSC model. Extensive experiments with ten datasets of texts and images illustrate the performance superiority of the proposed BTMSC method over state-of-the-art methods.

Abstract Portfolio managements in financial markets involve risk management strategies and opportunistic responses to individual trading behaviours. Optimal portfolios constructed aim to have a minimal risk with highest accompanying investment returns, regardless of market conditions. This paper focuses on providing an alternative view in maximising portfolio returns using Reinforcement Learning (RL) by considering dynamic risks appropriate to market conditions through dynamic portfolio rebalancing. The proposed algorithm is able to improve portfolio management by introducing the dynamic rebalancing of portfolios with vigorous risk through an RL agent. This is done while accounting for market conditions, asset diversifications, risk and returns in the global financial market. Studies have been performed in this paper to explore four types of methods with variations in fully portfolio rebalancing and gradual portfolio rebalancing, which combine with and without the use of the Long Short-Term Memory (LSTM) model to predict stock prices for adjusting the technical indicator centring. Performances of the four methods have been evaluated and compared using three constructed financial portfolios, including one portfolio with global market index assets with different risk levels, and two portfolios with uncorrelated stock assets from different sectors and risk levels. Observed from the experiment results, the proposed RL agent for gradual portfolio rebalancing with the LSTM model on price prediction outperforms the other three methods, as well as returns of individual assets in these three portfolios. The improvements of the returns using the RL agent for gradual rebalancing with prediction model are achieved at about 27.9–93.4% over those of the full rebalancing without prediction model. It has demonstrated the ability to dynamically adjust portfolio compositions according to the market trends, risks and returns of the global indices and stock assets.

Currently, illegal parking detection tasks are mainly achieved through manually checking by enforcement officers on patrol or using Closed-Circuit Television (CCTV) cameras. However, these methods either need high human labour costs or demand installation costs and procedures. Therefore, illegal parking detection solutions, which can reduce significant labour and equipment installation costs, are highly demanded. This paper proposes a novel voting based detection algorithm using deep learning networks implemented using in-vehicle cameras to achieve illegal parking detection with multiple offences’ types. Adopting in-vehicle cameras better matches real-world mobile scenarios than using traditional CCTV cameras as this helps enforcement authorities to reduce manpower and installation costs. A well-constructed new dataset with more than 10000 high-quality labelled images with seven object categories is built for illegal parking detection tasks. Additionally, one novel labelling method named “minimal illegal units” is proposed for illegal parking detection. It reduces the time and human labelling costs significantly, achieving a better correlation of a vehicle and its parking type. The experiments have been conducted in the urban areas of Singapore. Furthermore, the illumination robustness test has also been performed to illustrate that the proposed detection algorithm exhibits strong resistance to changing illumination conditions in varied operating environments. Our proposed detection algorithm can provide a benchmark for research in illegal parking detection.

Abstract In applications of augmented reality or mixed reality, rendering virtual objects in real scenes with consistent illumination is crucial for realistic visualization experiences. Prior learning-based methods reported in the literature usually attempt to reconstruct complicated high dynamic range environment maps from limited input, and rely on a separate rendering pipeline to light up the virtual object. In this paper, an object-based illumination transferring and rendering algorithm is proposed to tackle this problem within a unified framework. Given a single low dynamic range image, instead of recovering lighting environment of the entire scene, the proposed algorithm directly infers the relit virtual object. It is achieved by transferring implicit illumination features which are extracted from its nearby planar surfaces. A generative adversarial network is adopted in the proposed algorithm for implicit illumination features extraction and transferring. Compared to previous works in the literature, the proposed algorithm is more robust, as it is able to efficiently recover spatially varying illumination in both indoor and outdoor scene environments. Experiments have been conducted. It is observed that notable experiment results and comparison outcomes have been obtained quantitatively and qualitatively by the proposed algorithm in different environments. It shows the effectiveness and robustness for realistic virtual object insertion and improved realism.

In the past few years, augmented reality (AR) has rapidly advanced and has been applied in different fields. One of the successful AR applications is the immersive and interactive serious games, which can be used for education and learning purposes. In this project, a prototype of an AR serious game is developed and demonstrated. Gamers utilize a head-mounted device and a vibrotactile feedback jacket to explore and interact with the AR serious game. Fourteen vibration actuators are embedded in the vibrotactile feedback jacket to generate immersive AR experience. These vibration actuators are triggered in accordance with the designed game scripts. Various vibration patterns and intensity levels are synthesized in different game scenes. This article presents the details of the entire software development of the AR serious game, including game scripts, game scenes with AR effects design, signal processing flow, behavior design, and communication configuration. Graphics computations are processed using the graphics processing unit in the system. The performance of the AR serious game prototype is evaluated and analyzed. The computation loads and resource utilization of normal game scenes and heavy computation scenes are compared. With 14 vibration actuators placed at different body positions, various vibration patterns and intensity levels can be generated by the vibrotactile feedback jacket, providing different real-world feedback. The prototype of this AR serious game can be valuable in building large-scale AR or virtual reality educational and entertainment games. Possible future improvements of the proposed prototype are also discussed in this article.

This paper describes a novel design of a fuzzy inference chip that allows for real-time online context switching. A context refers to a situation or scenario of an application requiring specific domain knowledge. In particular, our focus is on the class of applications involving embedded fuzzy control. The domain knowledge therefore refers to fuzzy rules and memberships. The kind of applications being considered is real-time in nature, which necessitates the implementation of hardware for fuzzy inferencing. The chip architecture is described and details on the design of the chip is presented.