As mental health (MH) disorders become increasingly prevalent, their multifaceted symptoms and comorbidities with other conditions introduce complexity to diagnosis, posing a risk of underdiagnosis. While machine learning (ML) has been explored to mitigate these challenges, we hypothesized that multiple data modalities support more comprehensive detection and that non-intrusive collection approaches better capture natural behaviors. To understand the current trends, we systematically reviewed 184 studies to assess feature extraction, feature fusion, and ML methodologies applied to detect MH disorders from passively sensed multimodal data, including audio and video recordings, social media, smartphones, and wearable devices. Our findings revealed varying correlations of modality-specific features in individualized contexts, potentially influenced by demographics and personalities. We also observed the growing adoption of neural network architectures for model-level fusion and as ML algorithms, which have demonstrated promising efficacy in handling high-dimensional features while modeling within and cross-modality relationships. This work provides future researchers with a clear taxonomy of methodological approaches to multimodal detection of MH disorders to inspire future methodological advancements. The comprehensive analysis also guides and supports future researchers in making informed decisions to select an optimal data source that aligns with specific use cases based on the MH disorder of interest.

Educators are increasingly concerned about the usage of Large Language Models (LLMs) such as ChatGPT in programming education, particularly regarding the potential exploitation of imperfections in Artificial Intelligence Generated Content (AIGC) Detectors for academic misconduct. In this paper, we present an empirical study where the LLM is examined for its attempts to bypass detection by AIGC Detectors. This is achieved by generating code in response to a given question using different variants. We collected a dataset comprising 5,069 samples, with each sample consisting of a textual description of a coding problem and its corresponding human-written Python solution codes. These samples were obtained from various sources, including 80 from Quescol, 3,264 from Kaggle, and 1,725 from Leet-Code. From the dataset, we created 13 sets of code problem variant prompts, which were used to instruct ChatGPT to generate the outputs. Subsequently, we assessed the performance of five AIGC detectors. Our results demonstrate that existing AIGC Detectors perform poorly in distinguishing between human-written code and AI-generated code.

Fairness of deepfake detectors in the presence of anomalies are not well investigated, especially if those anomalies are more prominent in either male or female subjects. The primary motivation for this work is to evaluate how deepfake detection model behaves under such anomalies. However, due to the black-box nature of deep learning (DL) and artificial intelligence (AI) systems, it is hard to predict the performance of a model when the input data is modified. Crucially, if this defect is not addressed properly, it will adversely affect the fairness of the model and result in discrimination of certain sub-population unintentionally. Therefore, the objective of this work is to adopt metamorphic testing to examine the reliability of the selected deepfake detection model, and how the transformation of input variation places influence on the output. We have chosen MesoInception-4, a state-of-the-art deepfake detection model, as the target model and makeup as the anomalies. Makeups are applied through utilizing the Dlib library to obtain the 68 facial landmarks prior to filling in the RGB values. Metamorphic relations are derived based on the notion that realistic perturbations of the input images, such as makeup, involving eyeliners, eyeshadows, blushes, and lipsticks (which are common cosmetic appearance) applied to male and female images, should not alter the output of the model by a huge margin. Furthermore, we narrow down the scope to focus on revealing potential gender biases in DL and AI systems. Specifically, we are interested to examine whether MesoInception-4 model produces unfair decisions, which should be considered as a consequence of robustness issues. The findings from our work have the potential to pave the way for new research directions in the quality assurance and fairness in DL and AI systems.

Sign language recognition (SLR) refers to interpreting sign language glosses from given videos automatically. This research area presents a complex challenge in computer vision because of the rapid and intricate movements inherent in sign languages, which encompass hand gestures, body postures, and even facial expressions. Recently, skeleton-based action recognition has attracted increasing attention due to its ability to handle variations in subjects and backgrounds independently. However, current skeleton-based SLR methods exhibit three limitations: 1) they often neglect the importance of realistic hand poses, where most studies train SLR models on non-realistic skeletal representations; 2) they tend to assume complete data availability in both training or inference phases, and capture intricate relationships among different body parts collectively; 3) these methods treat all sign glosses uniformly, failing to account for differences in complexity levels regarding skeletal representations. To enhance the realism of hand skeletal representations, we present a kinematic hand pose rectification method for enforcing constraints. Mitigating the impact of missing data, we propose a feature-isolated mechanism to focus on capturing local spatial-temporal context. This method captures the context concurrently and independently from individual features, thus enhancing the robustness of the SLR model. Additionally, to adapt to varying complexity levels of sign glosses, we develop an input-adaptive inference approach to optimise computational efficiency and accuracy. Experimental results demonstrate the effectiveness of our approach, as evidenced by achieving a new state-of-the-art (SOTA) performance on WLASL100 and LSA64. For WLASL100, we achieve a top-1 accuracy of 86.50%, marking a relative improvement of 2.39% over the previous SOTA. For LSA64, we achieve a top-1 accuracy of 99.84%. The artefacts and code related to this study are made publicly online (https://github.com/mpuu00001/Siformer.git).

Urban heat islands (UHIs) pose significant environmental challenges in rapidly urbanising regions, highlighting the need for high-quality spatial datasets to better characterise urban thermal patterns. The Local Climate Zones (LCZ) approach provides a standardised land-use framework for UHI research and is known to strongly correlate with Land Surface Temperature (LST). In this study, 27 LCZ maps for five Southeast Asian cities were generated using cloud-free Landsat 8/9 scenes. Corresponding LST maps were retrieved to construct a paired dataset. Spearman’s rank correlation analysis confirmed a similar relationship between LCZ classes and LST existed in all pairs, shown by Spearman’s r ranging from -0.675 to -0.874, indicating a strong negative direction. When evaluated using the same testing polygons, our LCZ maps consistently outperformed the existing global LCZ dataset. A fully convolutional network was trained using the dataset to perform pixel-wise LST regression using LCZ maps, achieving an RMSE of 1.371℃ and an R ² of 0.891 under random-split setting and RMSE values ranging from 1.212 °C to 1.623 °C across individual city validations. The results highlight the substantial predictive performance of LCZ information for spatial LST estimation and demonstrate the suitability of the dataset for UHI machine-learning applications. The dataset is publicly available at https://doi.org/10.5281/zenodo.18223368 . Graphical Abstract Based on the graphical abstract snapshot, this study utilises Landsat 8 and 9 to develop a paired Local Climate Zones (LCZ) and Land Surface Temperature (LST) for five major Southeast Asian Cities. For each city, all Landsat 8/9 scenes from 2013 to 2025 were retrieved and clipped to the city extent and checked for cloud cover, if the cloud cover is less than 10%, the scene is selected for LCZ mapping through a Convolutional Neural Network (CNN) using our carefully-selected LCZ polygons, if the validation overall accuracy of the mapped LCZ scene is greater than or equal to 70%, the corresponding LST scene is calculated and clipped to form an LCZ and LST pair. Spearman’s rank correlation analysis resulted in an average Spearman’s rank of -0.783 across all generated pairs. The formed SEA-LCZ-LST dataset was tested for pixel-wise LST regression from LCZ using a Fully Convolutional Network (FCN) and achieved a RMSE of 1.371℃ and R2 of 0.891 under random-split setting. These findings demonstrate the dataset’s suitability for machine-learning-based UHI applications in the Southeast Asian region.