Mohamed bin Zayed University of Artificial Intelligence

Abstract Virtual Reality (VR) applications aim to create an immersive virtual world, which demands a high level of visual realism. The analytical material models commonly used in VR often fall short of reproducing complex real‐world appearances. Recently, neural materials have emerged as a promising alternative, offering a compact yet effective representation of real‐world materials. Deploying neural materials on low‐power mobile VR devices poses significant challenges due to the computational complexity of neural networks and the high display resolution and frame rate requirements of VR devices (commonly 72+ frames per second). We address these challenges by leveraging texture‐space shading with spatiotemporal computation amortization, driven by a compact, coarse‐to‐fine neural material model of extremely low capacity. Thanks to our distillation training scheme, our compact neural materials achieve visual quality comparable to NeuMIP [KMX*21] at a much lower cost. Our method reaches over 90 FPS on a mobile VR device (Meta Quest 3) even under multiple light sources.

The Dense Audio-Visual Event Localization (DAVEL) task aims to temporally localize events in untrimmed videos that occur simultaneously in both the audio and visual modalities. This paper explores DAVEL under a new and more challenging weakly-supervised setting (W-DAVEL task), where only video-level event labels are provided and the temporal boundaries of each event are unknown. We address W-DAVEL by exploiting cross-modal salient anchors, which are defined as reliable timestamps that are well predicted under weak supervision and exhibit highly consistent event semantics across audio and visual modalities. Specifically, we propose a Mutual Event Agreement Evaluation module, which generates an agreement score by measuring the discrepancy between the predicted audio and visual event classes. Then, the agreement score is utilized in a Cross-modal Salient Anchor Identification module, which identifies the audio and visual anchor features through global-video and local temporal window identification mechanisms. The anchor features after multimodal integration are fed into an Anchor-based Temporal Propagation module to enhance event semantic encoding in the original temporal audio and visual features, facilitating better temporal localization under weak supervision. We establish benchmarks for W-DAVEL on both the UnAV-100 and ActivityNet1.3 datasets. Extensive experiments demonstrate that our method achieves state-of-the-art performance.

Understanding the causes of the exceptional rate of evolution of the mammalian placenta is likely to aid the understanding of placental development and the etiology of the human-specific pregnancy disorder pre-eclampsia (PE). As retroelements are often lineage-specific and known to be co-opted for placental function, here we consider the binding of the transcription factors GATA3 and DLX5 to retroelements. These factors are dysregulated in pre-eclampsia, as are their downstream consequences. We identify retrovirus-derived LTR8B as a placentally-relevant cis-regulatory element (CRE), not least within the PSG array, a primate-specific genomic region that exhibits high intraspecies variability. LTR8B at PSG9 is particularly influential affecting other PSG family members. Moreover, unique among PSGs, PSG9 produces both secreted and membrane-anchored isoforms. The retroelement MER65-int provides alternative polyA signals that enable the evolution of secreted PSG variants by truncating the ancestral CEACAM protein's transmembrane domain. Functional characterization finds that LTR8B/PSG9 regulates the differentiation of multinucleated trophoblasts (syncytialization) and, like chorionic gonadotropin and syncytin1, determines the identity of syncytiotrophoblasts. Notably, PSG9 is the most upregulated PSG in PE, with levels correlated with GATA3 and DLX5 levels. Retroelements contribute to the structural and expression evolution of PSG genes, facilitating lineage-specific placental evolution. The LTR8B/PSG9 regulatory network plays a central role in syncytiotrophoblast differentiation. Given the association between DLX5/GATA3 dysregulation and elevated PSG9 levels, along with PSG9's expression in the first trimester, PSG9 shows potential as a predictive biomarker for preeclampsia.

Finding good branching orders is key to solving SAT problems efficiently, but finding such branching orders is a difficult problem. Using a learning based approach to predict a good branching order before solving, therefore, has potential. In this paper, we investigate predicting branching orders using graph neural networks as a preprocessing step to conflict-driven clause learning (CDCL) SAT solvers. We show that there are significant gains to be made in existing CDCL SAT solvers by providing a good initial branching. Further, we provide three labeling methods to find such initial branching orders in a tractable way. Finally, we train a graph neural network to predict these branching orders and show through our evaluations that a GNN-initialized ordering yields significant speedups on random 3-CNF and pseudo-industrial benchmarks, with generalization capabilities to instances much larger than the training set. However, we also find that the predictions fail at speeding up more difficult and industrial instances. We attribute this to the solver's dynamic heuristics, which rapidly overwrite the provided initialization, and to the complexity of these instances, making GNN prediction hard.

Asthma is a chronic respiratory disease with systemic and ocular manifestations. We present a multimodal deep-learning framework to predict asthma by fusing colour fundus photographs (CFP) with optical coherence tomography (OCT), and introduce a novel retinal dataset pairing fundus images with OCT measurements. The pipeline combined fundus CNN features with a feed-forward projection of OCT-derived optic nerve head (ONH) measurements for joint learning. Across five CNN backbones, fusion outperformed unimodal baselines; the best achieved AUC of 0.98, accuracy of 0.92, and F1 of 0.91. Findings supported retinal imaging as a noninvasive biomarker for systemic diseases (e.g., Parkinson, Alzheimer and cardiometabolic), constituting the first multimodal ocular-fusion approach to asthma diagnosis.

Abstract Background The genetic architecture of Breast Cancer (BC) in Arab populations remains largely understudied, limiting the precision of current prevention and screening programs. The Emirati Genome Program (EGP), one of the world’s first nation-wide sequencing initiatives, offers an unprecedented opportunity to study inherited BC risk across an entire population. Methods We analyzed 436,780 EGP individuals, including 229,309 women, integrating whole-genome sequencing (WGS) with electronic health records (EHRs). We quantified the prevalence and penetrance of pathogenic and likely pathogenic (P/LP) variants across 13 NCCN-recommended BC genes, evaluated the performance of established polygenic risk scores (PRS), and reconstructed >48,000 pedigrees to measure familial aggregation. Results P/LP variants were identified in 0.84% of women, accounting for 5.2% of BC cases (mean age of 45.9±11.1 years). Highly penetrant BRCA1 c.4065_4068del (p.Asn1355fs) and BRCA2 c.2808_2811del (p.Ala938Profs) variants showed age-specific cumulative risks of 37.6% and 31% by age 60, respectively, and allele frequencies up to tenfold higher in the Emirati population than in global reference datasets. The European-derived PRS model (PGS000004) demonstrated strong performance, advancing 10-year BC risk onset by a decade for women in the top decile. Family-based PRS discriminated affected from unaffected individuals, revealing higher polygenic risk even within sister pairs. Conclusions Nation-scale genome sequencing reveals, for the first time, the comprehensive landscape of inherited BC susceptibility within a Middle Eastern population. The integration of monogenic, polygenic, and familial data establishes a national framework for genomic risk stratification, transforming population genomics into a foundation for precision prevention and early detection in the UAE and beyond.

One of many impediments to applying graph neural networks (GNNs) in processing large-volume real-world graph-structured data is that it disapproves of a centralized training scheme which involves gathering data belonging to different organizations due to privacy concerns. As a distributed data processing scheme, federated graph learning (FGL) enables learning GNN models collaboratively without sharing participants' private data. Though theoretically feasible, a core challenge in FGL systems is the variation of local training data distributions among clients, also known as the data heterogeneity problem. Most existing solutions suffer from two problems: (1) The typical optimizer based on empirical risk minimization tends to cause local models to fall into sharp valleys and weakens their generalization to out-of-distribution graph data. (2) The prevalent dimensional collapse in the learned representations of local graph data has an adverse impact on the classification capacity of the GNN model. To this end, we formulate a novel optimization objective that is aware of the sharpness (i.e., the curvature of the loss surface) of local GNN models. By minimizing the loss function and its sharpness simultaneously, we seek out model parameters in a flat region with uniformly low loss values, thus improving the generalization over heterogeneous data. By introducing a regularizer based on the correlation matrix of local representations, we relax the correlations of representations generated by individual local graph samples, so as to alleviate the dimensional collapse of the learned model. The proposed Sharpness-aware fEderated grAph Learning (SEAL) algorithm can enhance the classification accuracy and generalization ability of local GNN models in federated graph learning. Experimental studies on several graph classification benchmarks show that SEAL consistently outperforms SOTA FGL baselines and provides gains for more participants.

Zero-shot diffusion posterior sampling offers a flexible framework for inverse problems by accommodating arbitrary degradation operators at test time, but incurs high computational cost due to repeated likelihood-guided updates. In contrast, previous amortized diffusion approaches enable fast inference by replacing likelihood-based sampling with implicit inference models, but at the expense of robustness to unseen degradations. We introduce an amortization strategy for diffusion posterior sampling that preserves explicit likelihood guidance by amortizing the inner optimization problems arising in variational diffusion posterior sampling. This accelerates inference for in-distribution degradations while maintaining robustness to previously unseen operators, thereby improving the trade-off between efficiency and flexibility in diffusion-based inverse problems.