Search

Early childhood education (ECE) conflict zones face profound challenges that undermine children's cognitive, emotional, and social development. Armed conflicts disrupt access to education through displacement, infrastructure destruction, and insecurity, leaving millions of children without safe learning environments (UNICEF, 2023). This report focuses on Syria, where years of war have deprived over two million children of education (UNESCO, 2018). These children face severe psychological trauma, including post-traumatic stress disorder (PTSD), anxiety, and depression, impairing their ability to learn and develop (Save the Children, 2020). This report emphasizes the dual importance of immediate and sustainable solutions to address these challenges. Immediate measures include providing psychological support, temporary learning spaces, and access to basic educational resources, while sustainable approaches, such as the "Hope Initiative," focus on creating resilient educational systems capable of withstanding future crises. This initiative, inspired by global best practices, proposes an integrated framework of proactive strategies, including teacher training, resource mobilization, and technology-driven learning solutions (Moving Minds Alliance, 2023). Using a mixed-methods approach, the study incorporates interviews with affected families, surveys with victims and their families, and case studies, such as that of Ahmed, a young Syrian child navigating educational challenges amid conflict. Findings reveal that displacement, resource shortages, and psychological stress significantly hinder educational progress, while community-driven initiatives provide hope and resilience. The report underscores the urgent need for regional collaboration and innovative policies to ensure that education remains a priority, even in the most challenging environments (Global Education Monitoring Report, 2019). This research reaffirms the transformative power of education as a tool for resilience, community empowerment, and long-term peacebuilding. By addressing the unique needs of children in conflict zones, we can pave the way for a future where every child, regardless of circumstance, has the opportunity to learn, grow, and contribute to a more peaceful world.

Compute Express Link (CXL) switch allows memory extension via PCIe physical layer to address increasing demand for larger memory capacities in data centers. However, CXL attached memory introduces 170ns to 400ns memory latency. This becomes a significant performance bottleneck for applications that host data in persistent memory as all updates, after traversing the CXL switch, must reach persistent domain to ensure crash consistent updates.We make a case for persistent CXL switch to persist updates as soon as they reach the switch and hence significantly reduce latency of persisting data. To enable this, we presented a system independent persistent buffer (PB) design that ensures data persistency at CXL switch. Our PB design provides 12\% speedup, on average, over volatile CXL switch. Our \textit{read forwarding} optimization improves speedup to 15\%., Comment: 7 pages, accepted work for Work-in-Progress (WIP) poster sessions at the 62nd DAC

Vision-Language-Action (VLA) models aim to predict robotic actions based on visual observations and language instructions. Existing approaches require fine-tuning pre-trained visionlanguage models (VLMs) as visual and language features are independently fed into downstream policies, degrading the pre-trained semantic alignments. We propose OTTER, a novel VLA architecture that leverages these existing alignments through explicit, text-aware visual feature extraction. Instead of processing all visual features, OTTER selectively extracts and passes only task-relevant visual features that are semantically aligned with the language instruction to the policy transformer. This allows OTTER to keep the pre-trained vision-language encoders frozen. Thereby, OTTER preserves and utilizes the rich semantic understanding learned from large-scale pre-training, enabling strong zero-shot generalization capabilities. In simulation and real-world experiments, OTTER significantly outperforms existing VLA models, demonstrating strong zeroshot generalization to novel objects and environments. Video, code, checkpoints, and dataset: https://ottervla.github.io/.

This study investigates the classification of aerial images depicting transmission towers, forests, farmland, and mountains. To complete the classification job, features are extracted from input photos using a Convolutional Neural Network (CNN) architecture. Then, the images are classified using Softmax. To test the model, we ran it for ten epochs using a batch size of 90, the Adam optimizer, and a learning rate of 0.001. Both training and assessment are conducted using a dataset that blends self-collected pictures from Google satellite imagery with the MLRNet dataset. The comprehensive dataset comprises 10,400 images. Our study shows that transfer learning models and MobileNetV2 in particular, work well for landscape categorization. These models are good options for practical use because they strike a good mix between precision and efficiency; our approach achieves results with an overall accuracy of 87% on the built CNN model. Furthermore, we reach even higher accuracies by utilizing the pretrained VGG16 and MobileNetV2 models as a starting point for transfer learning. Specifically, VGG16 achieves an accuracy of 90% and a test loss of 0.298, while MobileNetV2 outperforms both models with an accuracy of 96% and a test loss of 0.119; the results demonstrate the effectiveness of employing transfer learning with MobileNetV2 for classifying transmission towers, forests, farmland, and mountains., Comment: This paper is published in Journal of Computer Science, Volume 21 No. 3, 2025. It contains 635-645 pages

Emergency search and rescue (SAR) operations often require rapid and precise target identification in complex environments where traditional manual drone control is inefficient. In order to address these scenarios, a rapid SAR system, UAV-VLRR (Vision-Language-Rapid-Response), is developed in this research. This system consists of two aspects: 1) A multimodal system which harnesses the power of Visual Language Model (VLM) and the natural language processing capabilities of ChatGPT-4o (LLM) for scene interpretation. 2) A non-linearmodel predictive control (NMPC) with built-in obstacle avoidance for rapid response by a drone to fly according to the output of the multimodal system. This work aims at improving response times in emergency SAR operations by providing a more intuitive and natural approach to the operator to plan the SAR mission while allowing the drone to carry out that mission in a rapid and safe manner. When tested, our approach was faster on an average by 33.75% when compared with an off-the-shelf autopilot and 54.6% when compared with a human pilot. Video of UAV-VLRR: https://youtu.be/KJqQGKKt1xY

The UAV-VLPA* (Visual-Language-Planning-and-Action) system represents a cutting-edge advancement in aerial robotics, designed to enhance communication and operational efficiency for unmanned aerial vehicles (UAVs). By integrating advanced planning capabilities, the system addresses the Traveling Salesman Problem (TSP) to optimize flight paths, reducing the total trajectory length by 18.5\% compared to traditional methods. Additionally, the incorporation of the A* algorithm enables robust obstacle avoidance, ensuring safe and efficient navigation in complex environments. The system leverages satellite imagery processing combined with the Visual Language Model (VLM) and GPT's natural language processing capabilities, allowing users to generate detailed flight plans through simple text commands. This seamless fusion of visual and linguistic analysis empowers precise decision-making and mission planning, making UAV-VLPA* a transformative tool for modern aerial operations. With its unmatched operational efficiency, navigational safety, and user-friendly functionality, UAV-VLPA* sets a new standard in autonomous aerial robotics, paving the way for future innovations in the field., Comment: arXiv admin note: text overlap with arXiv:2501.05014

Characterizing a preclinical stage of Alzheimer's Disease (AD) via single imaging is difficult as its early symptoms are quite subtle. Therefore, many neuroimaging studies are curated with various imaging modalities, e.g., MRI and PET, however, it is often challenging to acquire all of them from all subjects and missing data become inevitable. In this regards, in this paper, we propose a framework that generates unobserved imaging measures for specific subjects using their existing measures, thereby reducing the need for additional examinations. Our framework transfers modality-specific style while preserving AD-specific content. This is done by domain adversarial training that preserves modality-agnostic but AD-specific information, while a generative adversarial network adds an indistinguishable modality-specific style. Our proposed framework is evaluated on the Alzheimer's Disease Neuroimaging Initiative (ADNI) study and compared with other imputation methods in terms of generated data quality. Small average Cohen's $d$ $< 0.19$ between our generated measures and real ones suggests that the synthetic data are practically usable regardless of their modality type., Comment: ISBI 2024 (oral)

Hypertensive retinopathy (HR) is a severe eye disease that may cause permanent vision loss if not diagnosed early. Traditional diagnostic methods are time-consuming and subjective, highlighting the need for an automated, reliable system. Existing studies often use a single Deep Learning (DL) model, struggling to distinguish HR stages. This study introduces a three-stage approach to enhance HR diagnosis accuracy. Initially, 14 CNN models were tested, identifying DenseNet169, MobileNet, and ResNet152 as the most effective. DenseNet169 achieved 87.73% accuracy, 87.75% precision, 87.73% recall, 87.67% F1-score, and 0.8359 Cohen's Kappa. MobileNet followed with 86.40% accuracy, 86.60% precision, 86.40% recall, 86.31% F1-score, and 0.8180 Cohen's Kappa. ResNet152 ranked third with 85.87% accuracy, 86.01% precision, 85.87% recall, 85.83% F1-score, and 0.8188 Cohen's Kappa. In the second stage, deep features from these models were fused and classified using Machine Learning (ML) algorithms (SVM, RF, XGBoost). SVM (sigmoid kernel) performed best with 92.00% accuracy, 91.93% precision, 92.00% recall, 91.91% F1-score, and 0.8930 Cohen's Kappa. The third stage applied meta-heuristic optimization (GA, ABC, PSO, HHO) for feature selection. HHO yielded 94.66% accuracy, precision, and recall, 94.64% F1-score, and 0.9286 Cohen's Kappa. The proposed approach surpassed single CNN models and previous studies in HR diagnosis accuracy and generalization.

Stacking excessive layers in DNN results in highly underdetermined system when training samples are limited, which is very common in medical applications. In this regard, we present a framework capable of deriving an efficient high-dimensional space with reasonable increase in model size. This is done by utilizing a transform (i.e., convolution) that leverages scale-space theory with covariance structure. The overall model trains on this transform together with a downstream classifier (i.e., Fully Connected layer) to capture the optimal multi-scale representation of the original data which corresponds to task-specific components in a dual space. Experiments on neuroimaging measures from Alzheimer's Disease Neuroimaging Initiative (ADNI) study show that our model performs better and converges faster than conventional models even when the model size is significantly reduced. The trained model is made interpretable using gradient information over the multi-scale transform to delineate personalized AD-specific regions in the brain., Comment: ISBI 2023

In this paper, we present a framework to analyze the impact of user velocity on the distribution of the peak age-of-information (PAoI) for both ground and aerial users by using the dominant interferer-based approximation. We first approximate the SINR meta distribution for the uplink transmission using the distances between the serving base station (BS) and each of the user of interest and the dominant interfering user, which is the interferer that provides the strongest average received power at the tagged BS. We then analyze the spatio-temporal correlation coefficient of the conditional success probability by studying the correlation between the aforementioned two distances. Finally, we choose PAoI as a performance metric to showcase how spatio-temporal correlation or user velocity affect system performance. Our results reveal that ground users exhibit higher spatio-temporal correlations compared to aerial users, resulting in a more pronounced impact of velocity on system performance, such as joint probability of the conditional success probability and distribution of PAoI. Furthermore, our work demonstrates that the dominant interferer-based approximation for the SINR meta distribution delivers good matching performance in complex scenarios, such as Nakagami-m fading model, and it can also be effectively utilized in computing spatio-temporal correlation, as this approximation is derived from the distances to the serving BS and the dominant interferer.

As large language models (LLMs) become increasingly integrated into daily life, ensuring their cultural sensitivity and inclusivity is paramount. We introduce our dataset, a year-long community-driven project covering all 22 Arab countries. The dataset includes instructions (input, response pairs) in both Modern Standard Arabic (MSA) and dialectal Arabic (DA), spanning 20 diverse topics. Built by a team of 44 researchers across the Arab world, all of whom are authors of this paper, our dataset offers a broad, inclusive perspective. We use our dataset to evaluate the cultural and dialectal capabilities of several frontier LLMs, revealing notable limitations. For instance, while closed-source LLMs generally exhibit strong performance, they are not without flaws, and smaller open-source models face greater challenges. Moreover, certain countries (e.g., Egypt, the UAE) appear better represented than others (e.g., Iraq, Mauritania, Yemen). Our annotation guidelines, code, and data for reproducibility are publicly available., Comment: More information about our dataset is available at our project page: https://github.com/UBC-NLP/palm

This review has explored the fundamental principles of thermal field theory in the context of a background magnetic field, highlighting its theoretical framework and some of its applications to the thermo-magnetic QCD plasma generated in heavy-ion collisions. Our discussion has been limited to equilibrium systems for clarity and conciseness. We analyzed bulk thermodynamic characteristics including the phase diagram as well as real-time observables, shedding light on the behaviour and dynamics of the thermo-magnetic QCD medium relevant to heavy-ion physics., Comment: An invited review in Progress in Particle and Nuclear Physics

Composite light-trapping structures offer a promising approach to achieving broadband absorption and high efficiency in thin-film solar cells (TFSCs) in order to accelerate sustainable energy solutions. As the leading material in thin-film solar technology, cadmium telluride (CdTe) faces challenges from surface reflective losses across the solar spectrum and weak absorption in the near-infrared (NIR) range. This computational study addresses these limitations by employing a dual light trapping technique: the top surfaces of both the CdS and CdTe layers are tapered as nanocones (NCs), while germanium (Ge) spherical nanoparticles (NPs) are embedded within the CdTe absorber layer to enhance broadband absorption. Numerical simulations using Finite-Difference Time Domain (FDTD) and other methods are used to optimize the parameters and configurations of both nanostructures, aiming to achieve peak optoelectronic performance. The results show that a short-circuit current density ($J_{sc}$) of 35.38 mA/$cm^2$ and a power conversion efficiency (PCE) of 27.76% can be achieved with optimal nanocone (NC) texturing and spherical Ge nanoparticle (NP) configurations, a 45.45% and 80.72% increase compared to baseline structure in $J_{sc}$ and PCE respectively. To understand the enhancement mechanisms, the study includes analyses using diffraction grating theory and Mie theory. Fabricability of these structures is also evaluated. Furthermore, an additional study on the effects of incident angle variation and polarization change demonstrates that the optimal structure is robust under practical conditions, maintaining consistent performance., Comment: 18 pages of main paper and 5 pages of supplementary material. 16 figures and 5 tables in total

Recent advancements have positioned AI, and particularly Large Language Models (LLMs), as transformative tools for scientific research, capable of addressing complex tasks that require reasoning, problem-solving, and decision-making. Their exceptional capabilities suggest their potential as scientific research assistants but also highlight the need for holistic, rigorous, and domain-specific evaluation to assess effectiveness in real-world scientific applications. This paper describes a multifaceted methodology for Evaluating AI models as scientific Research Assistants (EAIRA) developed at Argonne National Laboratory. This methodology incorporates four primary classes of evaluations. 1) Multiple Choice Questions to assess factual recall; 2) Open Response to evaluate advanced reasoning and problem-solving skills; 3) Lab-Style Experiments involving detailed analysis of capabilities as research assistants in controlled environments; and 4) Field-Style Experiments to capture researcher-LLM interactions at scale in a wide range of scientific domains and applications. These complementary methods enable a comprehensive analysis of LLM strengths and weaknesses with respect to their scientific knowledge, reasoning abilities, and adaptability. Recognizing the rapid pace of LLM advancements, we designed the methodology to evolve and adapt so as to ensure its continued relevance and applicability. This paper describes the methodology state at the end of February 2025. Although developed within a subset of scientific domains, the methodology is designed to be generalizable to a wide range of scientific domains., Comment: 33 pages, 18 figures

We present an analysis of the cluster X-ray morphology and active galactic nucleus (AGN) activity in nine $z\sim1$ galaxy clusters from the Massive and Distant Clusters of $WISE$ Survey (MaDCoWS) observed with $Chandra$. Using photon asymmetry ($A_{\text{phot}}$) to quantify X-ray morphologies, we find evidence that the four most dynamically disturbed clusters are likely to be mergers. Employing a luminosity cut of $7.6\times10^{42}$ erg/s to identify AGN in the 0.7-7.0 keV, we show that the majority of these clusters host excess AGN compared to the local field. We use the cumulative number-count ($\log N-\log S$) model to predict AGN incidence in cluster isophotes under this luminosity cut. Our analysis finds evidence (at $> 2\sigma$) of a positive correlation between AGN surface densities and photon asymmetry, suggesting that a disturbed cluster environment plays a pivotal role in regulating AGN triggering. Studying AGN incidence in cluster X-ray isophotes equivalent in area to $1.0r_{500}$, we find that the AGN space density inversely scales with cluster mass as $\sim M^{-0.5^{+0.18}_{-0.18}}$ at the 3.18$\sigma$ level. Finally, when we separately explore the cluster mass dependence of excess AGN surface density in disturbed and relaxed clusters, we see tentative evidence that the two morphologically distinct sub-populations exhibit diverging trends, especially near the outskirts, likely due to cluster merger-driven AGN triggering/suppression., Comment: 20 pages, 9 figures, 3 tables. Accepted for publication in ApJ

Recent advances in human shape learning have focused on achieving accurate human reconstruction from single-view images. However, in the real world, humans share space with other objects. Reconstructing images with humans and objects is challenging due to the occlusions and lack of 3D spatial awareness, which leads to depth ambiguity in the reconstruction. Existing methods in monocular human-object reconstruction fail to capture intricate details of clothed human bodies and object surfaces due to their template-based nature. In this paper, we jointly reconstruct clothed humans and objects in a spatially coherent manner from single-view images, while addressing human-object occlusions. A novel attention-based neural implicit model is proposed that leverages image pixel alignment to retrieve high-quality details, and incorporates semantic features extracted from the human-object pose to enable 3D spatial awareness. A generative diffusion model is used to handle human-object occlusions. For training and evaluation, we introduce a synthetic dataset with rendered scenes of inter-occluded 3D human scans and diverse objects. Extensive evaluation on both synthetic and real datasets demonstrates the superior quality of proposed human-object reconstructions over competitive methods.

Accurate multi-sensor calibration is essential for deploying robust perception systems in applications such as autonomous driving, robotics, and intelligent transportation. Existing LiDAR-camera calibration methods often rely on manually placed targets, preliminary parameter estimates, or intensive data preprocessing, limiting their scalability and adaptability in real-world settings. In this work, we propose a fully automatic, targetless, and online calibration framework, CalibRefine, which directly processes raw LiDAR point clouds and camera images. Our approach is divided into four stages: (1) a Common Feature Discriminator that trains on automatically detected objects--using relative positions, appearance embeddings, and semantic classes--to generate reliable LiDAR-camera correspondences, (2) a coarse homography-based calibration, (3) an iterative refinement to incrementally improve alignment as additional data frames become available, and (4) an attention-based refinement that addresses non-planar distortions by leveraging a Vision Transformer and cross-attention mechanisms. Through extensive experiments on two urban traffic datasets, we show that CalibRefine delivers high-precision calibration results with minimal human involvement, outperforming state-of-the-art targetless methods and remaining competitive with, or surpassing, manually tuned baselines. Our findings highlight how robust object-level feature matching, together with iterative and self-supervised attention-based adjustments, enables consistent sensor fusion in complex, real-world conditions without requiring ground-truth calibration matrices or elaborate data preprocessing., Comment: Submitted to Transportation Research Part C: Emerging Technologies

Glaucoma is a prevalent eye disease that progresses silently without symptoms. If not detected and treated early, it can cause permanent vision loss. Computer-assisted diagnosis systems play a crucial role in timely and efficient identification. This study introduces MaxGlaViT, a lightweight model based on the restructured Multi-Axis Vision Transformer (MaxViT) for early glaucoma detection. First, MaxViT was scaled to optimize block and channel numbers, resulting in a lighter architecture. Second, the stem was enhanced by adding attention mechanisms (CBAM, ECA, SE) after convolution layers to improve feature learning. Third, MBConv structures in MaxViT blocks were replaced by advanced DL blocks (ConvNeXt, ConvNeXtV2, InceptionNeXt). The model was evaluated using the HDV1 dataset, containing fundus images of different glaucoma stages. Additionally, 40 CNN and 40 ViT models were tested on HDV1 to validate MaxGlaViT's efficiency. Among CNN models, EfficientB6 achieved the highest accuracy (84.91%), while among ViT models, MaxViT-Tiny performed best (86.42%). The scaled MaxViT reached 87.93% accuracy. Adding ECA to the stem block increased accuracy to 89.01%. Replacing MBConv with ConvNeXtV2 further improved it to 89.87%. Finally, integrating ECA in the stem and ConvNeXtV2 in MaxViT blocks resulted in 92.03% accuracy. Testing 80 DL models for glaucoma stage classification, this study presents a comprehensive and comparative analysis. MaxGlaViT outperforms experimental and state-of-the-art models, achieving 92.03% accuracy, 92.33% precision, 92.03% recall, 92.13% f1-score, and 87.12% Cohen's kappa score.

Hybrid paper interfaces leverage augmented reality to combine the desired tangibility of paper documents with the affordances of interactive digital media. Typically, virtual content can be embedded through direct links (e.g., QR codes); however, this impacts the aesthetics of the paper print and limits the available visual content space. To address this problem, we present Imprinto, an infrared inkjet watermarking technique that allows for invisible content embeddings only by using off-the-shelf IR inks and a camera. Imprinto was established through a psychophysical experiment, studying how much IR ink can be used while remaining invisible to users regardless of background color. We demonstrate that we can detect invisible IR content through our machine learning pipeline, and we developed an authoring tool that optimizes the amount of IR ink on the color regions of an input document for machine and human detectability. Finally, we demonstrate several applications, including augmenting paper documents and objects., Comment: 18 pages, 13 figures. To appear in the Proceedings of the 2025 ACM CHI Conference on Human Factors in Computing Systems. https://imprinto.github.io

Breast cancer cell lines are indispensable tools for unraveling disease mechanisms, developing new drugs and personalized medicine, yet their heterogeneity and inconsistent classification pose significant challenges in model selection and data reproducibility. This article aims at providing a comprehensive and user-friendly framework for broadly mapping the essential features of publicly available human breast cancer cell lines. The cells are classified using (1) absolute criteria, i.e. objective features such as origin (e.g., MDA-MB, MCF), histological subtype (ductal, lobular), hormone receptor status (ER/PR/HER2), and genetic mutations (BRCA1, TP53), and (2) relative criteria, which contextualize functional behaviors like metastatic potential, drug sensitivity and genomic instability. We systematically catalog over 70 cell lines, detailing their molecular profiles, research applications and clinical relevance. Critical gaps are addressed, including the underrepresentation of cell lines from young patients, male breast cancer, and diverse ethnicities, as well as genetic drift during long-term culture. This article bridges in vitro studies with meaningful applications, offering a tool for easily selecting lines that mirror specific research objectives in a clinical setting. The goal is to empower researchers to optimize experimental design, enhance translational relevance and accelerate therapeutic development to advance precision oncology in breast cancer research., Comment: 2 figures, 3 tables

In this article, we examine two double-phase variable exponent problems, each formulated within a distinct framework. The first problem is non-variational, as the nonlinear term may depend on the gradient of the solution. The first main result establishes an existence property from the nonlinear monotone operator theory given by Browder and Minty. The second problem is set up within a variational framework, where we employ a well-known critical point result by Bonanno and Chinn\`{\i}. In both cases, we demonstrate the existence of at least one nontrivial solution. To illustrate the practical application of the main results, we provide examples for each problem.

This study investigates the application of Super-Resolution techniques in holographic microscopy to enhance quantitative phase imaging. An off-axis Mach-Zehnder interferometric setup was employed to capture interferograms. The study evaluates two Super-Resolution models, RCAN and Real-ESRGAN, for their effectiveness in reconstructing high-resolution interferograms from a microparticle-based dataset. The models were assessed using two primary approaches: image-based analysis for structural detail enhancement and morphological evaluation for maintaining sample integrity and phase map accuracy. The results demonstrate that RCAN achieves superior numerical precision, making it ideal for applications requiring highly accurate phase map reconstruction, while Real-ESRGAN enhances visual quality and structural coherence, making it suitable for visualization-focused applications. This study highlights the potential of Super-Resolution models in overcoming diffraction-imposed resolution limitations in holographic microscopy, opening the way for improved imaging techniques in biomedical diagnostics, materials science, and other high-precision fields.

We propose a machine learning-based local search approach for finding feasible solutions of large-scale maritime inventory routing optimization problems. Given the combinatorial complexity of the problems, we integrate a graph neural network-based neighborhood selection method to enhance local search efficiency. Our approach enables a structured exploration of vessel neighborhoods, improving solution quality while maintaining computational efficiency. Through extensive computational experiments on realistic instances, we demonstrate that our method outperforms direct mixed-integer programming in solution time.

In this paper, we derive multiple anisotropic analogs from the established isotropic model by means of the gravitational decoupling approach in a fluid-geometry interaction based theory. To accomplish this, we initially consider a static spherical perfect-fluid configuration and then introduce a new matter source to induce anisotropic behavior in the system. The resulting field equations encapsulate the entire matter distribution and thus become much complicated. We then split these equations into two sets through implementing a particular transformation, each set delineating characteristics attributed to their original fluid sources. We adopt the Heintzmann's ansatz and some constraints on extra gravitating source to deal with the first and second systems of equations, respectively. Furthermore, the two fundamental forms of the matching criteria are used to make the constant in the considered solution known. By utilizing the preliminary information of a star candidate LMC X-4, we assess the physical validity of the developed models. Our analysis indicates that both our models exhibit characteristics which are well-agreed with the acceptability criteria for certain parametric values., Comment: 18 pages, 10 figures

We introduce SigLIP 2, a family of new multilingual vision-language encoders that build on the success of the original SigLIP. In this second iteration, we extend the original image-text training objective with several prior, independently developed techniques into a unified recipe -- this includes captioning-based pretraining, self-supervised losses (self-distillation, masked prediction) and online data curation. With these changes, SigLIP 2 models outperform their SigLIP counterparts at all model scales in core capabilities, including zero-shot classification, image-text retrieval, and transfer performance when extracting visual representations for Vision-Language Models (VLMs). Furthermore, the new training recipe leads to significant improvements on localization and dense prediction tasks. We also train variants which support multiple resolutions and preserve the input's native aspect ratio. Finally, we train on a more diverse data-mixture that includes de-biasing techniques, leading to much better multilingual understanding and improved fairness. To allow users to trade off inference cost with performance, we release model checkpoints at four sizes: ViT-B (86M), L (303M), So400m (400M), and g (1B)., Comment: Model checkpoints are available at https://github.com/google-research/big_vision/tree/main/big_vision/configs/proj/image_text/README_siglip2.md

We present an end-to-end framework for generating synthetic users for evaluating interactive agents designed to encourage positive behavior changes, such as in health and lifestyle coaching. The synthetic users are grounded in health and lifestyle conditions, specifically sleep and diabetes management in this study, to ensure realistic interactions with the health coaching agent. Synthetic users are created in two stages: first, structured data are generated grounded in real-world health and lifestyle factors in addition to basic demographics and behavioral attributes; second, full profiles of the synthetic users are developed conditioned on the structured data. Interactions between synthetic users and the coaching agent are simulated using generative agent-based models such as Concordia, or directly by prompting a language model. Using two independently-developed agents for sleep and diabetes coaching as case studies, the validity of this framework is demonstrated by analyzing the coaching agent's understanding of the synthetic users' needs and challenges. Finally, through multiple blinded evaluations of user-coach interactions by human experts, we demonstrate that our synthetic users with health and behavioral attributes more accurately portray real human users with the same attributes, compared to generic synthetic users not grounded in such attributes. The proposed framework lays the foundation for efficient development of conversational agents through extensive, realistic, and grounded simulated interactions.

The prediction of the Remaining Useful Life of aircraft engines is a critical area in high-reliability sectors such as aerospace and defense. Early failure predictions help ensure operational continuity, reduce maintenance costs, and prevent unexpected failures. Traditional regression models struggle with censored data, which can lead to biased predictions. Survival models, on the other hand, effectively handle censored data, improving predictive accuracy in maintenance processes. This paper introduces a novel approach based on the Rashomon perspective, which considers multiple models that achieve similar performance rather than relying on a single best model. This enables uncertainty quantification in survival probability predictions and enhances decision-making in predictive maintenance. The Rashomon survival curve was introduced to represent the range of survival probability estimates, providing insights into model agreement and uncertainty over time. The results on the CMAPSS dataset demonstrate that relying solely on a single model for RUL estimation may increase risk in some scenarios. The censoring levels significantly impact prediction uncertainty, with longer censoring times leading to greater variability in survival probabilities. These findings underscore the importance of incorporating model multiplicity in predictive maintenance frameworks to achieve more reliable and robust failure predictions. This paper contributes to uncertainty quantification in RUL prediction and highlights the Rashomon perspective as a powerful tool for predictive modeling., Comment: 4 pages, 1 figures

Creating new fonts requires a lot of human effort and professional typographic knowledge. Despite the rapid advancements of automatic font generation models, existing methods require users to prepare pre-designed characters with target styles using font-editing software, which poses a problem for non-expert users. To address this limitation, we propose FontCraft, a system that enables font generation without relying on pre-designed characters. Our approach integrates the exploration of a font-style latent space with human-in-the-loop preferential Bayesian optimization and multimodal references, facilitating efficient exploration and enhancing user control. Moreover, FontCraft allows users to revisit previous designs, retracting their earlier choices in the preferential Bayesian optimization process. Once users finish editing the style of a selected character, they can propagate it to the remaining characters and further refine them as needed. The system then generates a complete outline font in OpenType format. We evaluated the effectiveness of FontCraft through a user study comparing it to a baseline interface. Results from both quantitative and qualitative evaluations demonstrate that FontCraft enables non-expert users to design fonts efficiently., Comment: 14 pages

BabyLM aims to dissolve the boundaries between cognitive modeling and language modeling. We call for both workshop papers and for researchers to join the 3rd BabyLM competition. As in previous years, we call for participants in the data-efficient pretraining challenge in the general track. This year, we also offer a new track: INTERACTION. This new track encourages interactive behavior, learning from a teacher, and adapting the teaching material to the student. We also call for papers outside the competition in any relevant areas. These include training efficiency, cognitively plausible research, weak model evaluation, and more., Comment: EMNLP 2025 BabyLM Workshop. arXiv admin note: text overlap with arXiv:2404.06214

This paper explores the implications of employing non-volatile memory (NVM) as primary storage for a data base management system (DBMS). We investigate the modifications necessary to be applied on top of a traditional relational DBMS to take advantage of NVM features. As a case study, we modify the storage engine (SE) of PostgreSQL enabling efficient use of NVM hardware. We detail the necessary changes and challenges such modifications entail and evaluate them using a comprehensive emulation platform. Results indicate that our modified SE reduces query execution time by up to 45% and 13% when compared to disk and NVM storage, with average reductions of 19% and 4%, respectively. Detailed analysis of these results shows that while our modified SE is able to access data more efficiently, data is not close to the processing units when needed for processing, incurring long latency misses that hinder the performance. To solve this, we develop a general purpose library that employs helper threads to prefetch data from NVM hardware via a simple API. Our library further improves query execution time for our modified SE when compared to disk and NVM storage by up to 54% and 17%, with average reductions of 23% and 8%, respectively., Comment: 32 pages (including last blank page), 15 Figures, 2 Tables

We introduce MuJoCo Playground, a fully open-source framework for robot learning built with MJX, with the express goal of streamlining simulation, training, and sim-to-real transfer onto robots. With a simple "pip install playground", researchers can train policies in minutes on a single GPU. Playground supports diverse robotic platforms, including quadrupeds, humanoids, dexterous hands, and robotic arms, enabling zero-shot sim-to-real transfer from both state and pixel inputs. This is achieved through an integrated stack comprising a physics engine, batch renderer, and training environments. Along with video results, the entire framework is freely available at playground.mujoco.org

In the context of 5G, infrastructure sharing has been identified as a potential solution to reduce the investment costs of cellular networks. In particular, it can help low-income regions build 5G networks more affordably and further bridge the digital divide. There are two main kinds of infrastructure sharing: passive sharing (i.e. site sharing) and active sharing (i.e. access sharing), which require mobile network operators (MNOs) to share their non-electronic elements or electronic elements, respectively. Because co-construction and sharing can achieve broader coverage with lower investment, through percolation theory, we investigate how different sharing strategies can deliver large-scale continuous services. First, we examine the percolation characteristics in signal-to-interference-plus-noise ratio (SINR) coverage graphs and the necessary conditions for percolation. Second, we propose an 'average coverage radius' to approximate the SINR graph with a low base station (BS) density based on the Gilbert disk model. Finally, we estimate the critical conditions of BS densities of MNOs for different sharing strategies and compare the percolation probabilities under different infrastructure sharing strategies.

With the advent of the 6G era, global connectivity has become a common goal in the evolution of communications, aiming to bring Internet services to more unconnected regions. Additionally, the rise of applications such as the Internet of Everything and remote education also requires global connectivity. Non-terrestrial networks (NTN), particularly low earth orbit (LEO) satellites, play a crucial role in this future vision. Although some literature already analyze the coverage performance using stochastic geometry, the ability of generating large-scale continuous service area is still expected to analyze. Therefore, in this paper, we mainly investigate the necessary conditions of LEO satellite deployment for large-scale continuous service coverage on the earth. Firstly, we apply percolation theory to a closed spherical surface and define the percolation on a sphere for the first time. We introduce the sub-critical and super-critical cases to prove the existence of the phase transition of percolation probability. Then, through stereographic projection, we introduce the tight bounds and closed-form expression of the critical number of LEO satellites on the same constellation. In addition, we also investigate how the altitude and maximum slant range of LEO satellites affect percolation probability, and derive the critical values of them. Based on our findings, we provide useful recommendations for companies planning to deploy LEO satellite networks to enhance connectivity.

Due to the involved massive number of devices, radio frequency (RF) energy harvesting is indispensable to realize the foreseen Internet-of-Everything (IoE) within 6G networks. Analogous to the cellular networks concept, shared energy stations (ESs) are foreseen to supply energy-as-a-service (EaaS) in order to recharge devices that belong to different IoE operators who are offering diverse use cases. Considering the capital expenditure (CAPEX) for ES deployment along with their finite wireless energy transfer (WET) zones, spatial energy gaps are plausible. Furthermore, the ESs deployment cannot cover 100% of the energy-harvesting devices of all coexisting IoE use cases. In this context, we utilize percolation theory to characterize the feasibility of large-scale device-to-device (D2D) connectivity of IoE networks operating under EaaS platforms. Assuming that ESs and IoE devices follow independent Poisson point processes (PPPs), we construct a connectivity graph for the IoE devices that are within the WET zones of ESs. Continuum percolation on the construct graph is utilized to derive necessary and sufficient conditions for large-scale RF-powered D2D connectivity in terms of the required IoE device density and communication range along with the required ESs density and WET zone size. Fixing the IoE network parameters along with the size of WET zones, we obtain the approximate critical value of the ES density that ensures large-scale connectivity using the inner-city and Gilbert disk models. By imitating the bounds and combining the approximations, we construct an approximate expression for the critical ES density function, which is necessary to minimize the EaaS CAPEX under the IoE connectivity constraint.

Segmentation of colorectal cancer (CRC) tumors in 3D medical imaging is both complex and clinically critical, providing vital support for effective radiation therapy planning and survival outcome assessment. Recently, 3D volumetric segmentation architectures incorporating long-range sequence modeling mechanisms, such as Transformers and Mamba, have gained attention for their capacity to achieve high accuracy in 3D medical image segmentation. In this work, we evaluate the effectiveness of these global token modeling techniques by pitting them against our proposed MambaOutUNet within the context of our newly introduced colorectal tumor segmentation dataset (CTS-204). Our findings suggest that robust local token interactions can outperform long-range modeling techniques in cases where the region of interest is small and anatomically complex, proposing a potential shift in 3D tumor segmentation research., Comment: 5 pages, 1 figures

The smart metering infrastructure may become one of the key elements in efficiently managing energy in smart cities. At the same time, traditional measurement record collection is performed by manual methods, which raises cost, safety, and accuracy issues. This paper proposes an innovative SMI architecture based on an unmanned aerial vehicle swarm organizing itself for the autonomous data collection in smart metering infrastructure with scalability and cost-effectiveness while minimizing risks. We design an architecture-based comprehensive system with various phases of operation, communication protocols, and robust failure-handling mechanisms to ensure reliable operations. We further perform extensive simulations in maintenance of precise formations during flight, efficient data collection from smart meters, and adaptation to various failure scenarios. Importantly, we analyze the energy consumption of the proposed system in both drone flight operations and network communication. We now propose a battery sizing strategy and provide an estimate of the operational lifetime of the swarm, underlining the feasibility and practicality of our approach. Our results show that UAV swarms have great potential to revolutionize smart metering and to bring a further brick to greener and more resilient smart cities.

In laparoscopy surgical training and evaluation, real-time detection of surgical actions with interpretable outputs is crucial for automated and real-time instructional feedback and skill development. Such capability would enable development of machine guided training systems. This paper presents a rapid deployment approach utilizing automated machine learning methods, based on surgical action data collected from both experienced and trainee surgeons. The proposed approach effectively tackles the challenge of highly imbalanced class distributions, ensuring robust predictions across varying skill levels of surgeons. Additionally, our method partially incorporates model transparency, addressing the reliability requirements in medical applications. Compared to deep learning approaches, traditional machine learning models not only facilitate efficient rapid deployment but also offer significant advantages in interpretability. Through experiments, this study demonstrates the potential of this approach to provide quick, reliable and effective real-time detection in surgical training environments

In this paper, we study a class of nonlocal multi-phase variable exponent problems within the framework of a newly introduced Musielak-Orlicz Sobolev space. We consider two problems, each distinguished by the type of nonlinearity it includes. To establish the existence of at least one nontrivial solution for each problem, we employ two different monotone operator methods., Comment: arXiv admin note: text overlap with arXiv:2501.17344