Researchers use replication to confirm, strengthen, and advance Computing Education Research (CER). However, prior research shows that replication is infrequently used in CER, even though the community encourages its use. Previous research suggests that the CER community uses different terms to describe replication, which we aim to confirm in this Working Group (WG). We will conduct a Systematic Literature Review (SLR) across influential international CER venues to understand how researchers present and conduct replication studies, possibly identifying venues receptive to papers applying this research design. In addition, we will interview Computing Education researchers, conference leaders, and journal editors to understand their experiences and perceptions with replication. We expect to collect suggestions and recommendations on how CER can encourage more replication in future studies. Our work will highlight and confirm the terms the CER community uses to present replication studies, enabling researchers and educators to better identify these studies.

Collaboration and Teamwork are some of the most important non-technical skills required by computer science graduates. However, the research on collaboration in CS education is diffuse and inconsistent. In this work, we set out to systematically catalog and map the research on collaboration as it pertains to computer science education. We produce a literature mapping of research on the axes of population being studied, intervention being researched, and method of evaluation. We find a large number of papers studying group projects or assessments, but very few with directly assigned roles or clear hierarchies, and much of the research is on the student experience, with relatively limited insight into the efficacy or learning outcomes. The full literature map, heat maps, and systematic details are made available for the community.

Long-Tailed (LT) recognition has been widely studied to tackle the challenge of imbalanced data distributions in real-world applications. However, the design of neural architectures for LT settings has received limited attention, despite evidence showing that architecture choices can substantially affect performance. This paper aims to bridge the gap between LT challenges and neural network design by providing an in-depth analysis of how various architectures influence LT performance. Specifically, we systematically examine the effects of key network components on LT handling, such as topology, convolutions, and activation functions. Based on these observations, we propose two convolutional operations optimized for improved performance. Recognizing that operation interactions are also crucial to network effectiveness, we apply Neural Architecture Search (NAS) to facilitate efficient exploration. We propose LT-DARTS, a NAS method with a novel search space and search strategy specifically designed for LT data. Experimental results demonstrate that our approach consistently outperforms existing architectures across multiple LT datasets, achieving parameter-efficient, state-of-the-art results when integrated with current LT methods.

Database education is a cornerstone under many of the more popular topics in computer science such as machine learning and visualization. Although, in recent years, more fundamental research into database education has come out, there are many more ways in which it can be extended. Research on the practice of teaching databases, namely on the educational materials and explanations of teachers, can help us create new building blocks for fundamental research. This working group aims to collect and present notional machines of different types, for a wide range of database subtopics. These materials offer and updated context for database educators to design their courses from, as well as open up pathways of further research into database education.

Involving undergraduate students in research has a wide array of benefits, for the students themselves, for the research team, and for the community.

AI-aided Financial Regulation (AIFR) is a practical and significant task, but current solutions have yet to be optimized with customized model designs. Given the privacy concerns surrounding financial data, we aim to employ Neural Architecture Search (NAS) to help non-expert end-users automatically design architectures. The genetic algorithm-based NAS stands out due to its relatively low hardware requirements and robust theoretical foundation. However, constrained by limited data, the model would undergo architecture search on a general regulatory dataset while being deployed on private one owned by each organization. The data distribution of the private dataset may vary from that of public datasets, giving rise to the challenge of data domain shift. To alleviate this problem, we propose a novel fitness evaluation method. When scoring the fitness, we take into account both the architecture’s validation accuracy and its potential for generalization by the metric of loss landscape. In addition, we improve the training paradigm for evaluation, utilizing a prototype-based training paradigm based on embedding distances for classification, allowing for rapid domain adaptation and improve performance on the distribution-shift data. We further introduce GA-TextCNN, a GA-based NAS framework specifically designed for text recognition, enhancing its suitability for text data within AIFR tasks. To demonstrate the effectiveness of our approach, we collect two related datasets and evaluate our method on it. The extensive experiments demonstrate that our method significantly improves baseline models and is effective in solving the AIFR problem.

Recently, the continuous evolution of deep learning has opened up promising avenues to groundbreaking advancements in wireless sensing systems, which significantly enhance the practical applications of WiFi-based Human Activity Recognition (HAR) systems. However, despite these strides, such systems remain susceptible to adversarial attacks. This article unveils the vulnerability of existing WiFi-based HAR systems to common adversaries, revealing their insufficient robustness. While the intuitive approach is to employ adversarial training to fortify the models, our investigation exposes inherent deficiencies in the current approach. Specifically, we confirm that the strength of perturbations directly influences training outcomes. Moreover, even when confined within a specified perturbation radius, the perturbation strength exhibits variability within a prescribed range, potentially giving rise to “extreme” samples that could compromise training results. To address this challenge, we propose a two-stage Selective Adversarial Training (SAT) approach that integrates model confidence calibration and sample selection. Specifically, we start with calibrating the model and then selectively choose samples from all adversarial examples based on the calibrated confidence outputs that align with the desired criteria for adversarial training. This sample-wise perturbation intensity control effectively prevents the inclusion of inappropriate samples in training, a capability lacking in previous domain-wise perturbation control. Our experiments demonstrate that the proposed fine-grained training method, SAT, is both straightforward and effective in augmenting adversarial training results.

Two-step linear-exponential architectures can be applied to incremental ADCs (IADC) to achieve high resolution. In the first step, the ADC works as a normal first-order IADC while, in the second step, the exponential integrator is used to implement extended counting. There exist two architectures for the implementation of the exponential step, where the only difference depends on whether the input signal is connected or disconnected. By conducting a comparative analysis on such two slightly different linear-exponential architectures, we propose to combine the exponential and slope techniques to further boost the resolution without degrading its original thermal-noise suppression ability and DWA effectiveness. Mathematical analysis and simulation results are presented to confirm the principle of the proposed IADC.

This paper presents a two-phase linear-exponential incremental analog-to-digital converter (IADC) with using second-order noise coupling (NC). In the first phase, it works as a first-order IADC. Then the second-order NC path is activated in the second phase to significantly expedite the accumulation speed. Moreover, during the second phase, the integrator is disabled to achieve a large maximum stable amplitude (MSA). Simulations demonstrated that the proposed architecture could achieve a higher signal-to-quantization-noise ratio (SQNR) while avoiding the noise penalty and keeping the high effectiveness of data weighting averaging (DWA) compared with the prior art with using first-order NC. Mathematical analysis and further simulation results are presented to confirm the theory of the proposed structure.