The task of graph-level out-of-distribution (OOD) detection is crucial for deploying graph neural networks in real-world settings. In this paper, we observe a significant difference in the relationship between the largest and second-largest eigenvalues of the Laplacian matrix for in-distribution (ID) and OOD graph samples: OOD samples often exhibit anomalous spectral gaps (the difference between the largest and second-largest eigenvalues). This observation motivates us to propose SpecGap, an effective post-hoc approach for OOD detection on graphs. SpecGap adjusts features by subtracting the component associated with the second-largest eigenvalue, scaled by the spectral gap, from the high-level features (i.e., X - (λn - λn-1) u_n-1 v_n-1^T). SpecGap achieves state-of-the-art performance across multiple benchmark datasets. We present extensive ablation studies and comprehensive theoretical analyses to support our empirical results. As a parameter-free post-hoc method, SpecGap can be easily integrated into existing graph neural network models without requiring any additional training or model modification.

Background/Objectives: Artificial intelligence (AI) is revolutionising healthcare for people with disabilities, including those with autism spectrum disorder (ASD), in the era of advanced technology. This work explicitly addresses the challenges posed by inconsistent data from various sources by developing and evaluating a robust deep ensemble learning system for the accurate and reliable classification of autism spectrum disorder (ASD) based on facial images. Methods: We created a system that learns from two publicly accessible datasets of ASD images (Kaggle and YTUIA), each with unique demographics and image characteristics. Utilising a weighted ensemble strategy (FPPR), our innovative ASD-UANet ensemble combines the Xception and ResNet50V2 models to maximise model contributions. This methodology underwent extensive testing on a range of groups stratified by age and gender, including a critical assessment of an unseen, real-time dataset (UIFID) to determine how well it generalised to new domains. Results: The performance of the ASD-UANet ensemble was consistently better. It significantly outperformed individual transfer learning models (e.g., Xception alone on T1+T2 yielded an accuracy of 83%), achieving an impressive 96.0% accuracy and an AUC of 0.990 on the combined-domain dataset (T1+T2). Notably, the ASD-UANet ensemble demonstrated strong generalisation on the unseen real-time dataset (T3), achieving 90.6% accuracy and an AUC of 0.930. This demonstrates how well it generalises to new data distributions. Conclusions: Our findings demonstrate significant potential for widespread, equitable, and clinically beneficial ASD screening using this promising, reasonably priced, and non-invasive method. This study establishes the foundation for more precise diagnoses and greater inclusion for people with autism spectrum disorder (ASD) by integrating methods for diverse data and combining deep learning models.

With the ever-increasing computing power of supercomputers and the growing scale of scientific applications, the efficiency of MPI collective communication turns out to be a critical bottleneck in large-scale distributed and parallel processing. The large message size in MPI collectives is particularly concerning because it can significantly degrade overall parallel performance. To address this issue, prior research simply applies off-the-shelf fixed-rate lossy compressors in the MPI collectives, leading to suboptimal performance, limited generalizability, and unbounded errors. In this paper, we propose a novel solution, called ZCCL, which leverages error-bounded lossy compression to significantly reduce the message size, resulting in a substantial reduction in communication costs. The key contributions are three-fold. (1) We develop two general, optimized lossy-compression-based frameworks for both types of MPI collectives (collective data movement as well as collective computation), based on their particular characteristics. Our framework not only reduces communication costs but also preserves data accuracy. (2) We customize fZ-light, an ultra-fast error-bounded lossy compressor, to meet the specific needs of collective communication. (3) We integrate ZCCL into multiple collectives, such as Allgather, Allreduce, Scatter, and Broadcast, and perform a comprehensive evaluation based on real-world scientific application datasets. Experiments show that our solution outperforms the original MPI collectives as well as multiple baselines by 1.9--8.9X.