Search

Objective: Assessing Alzheimer's disease (AD) using high-dimensional radiology images is clinically important but challenging. Although Artificial Intelligence (AI) has advanced AD diagnosis, it remains unclear how to design AI models embracing predictability and explainability. Here, we propose VisTA, a multimodal language-vision model assisted by contrastive learning, to optimize disease prediction and evidence-based, interpretable explanations for clinical decision-making. Methods: We developed VisTA (Vision-Text Alignment Model) for AD diagnosis. Architecturally, we built VisTA from BiomedCLIP and fine-tuned it using contrastive learning to align images with verified abnormalities and their descriptions. To train VisTA, we used a constructed reference dataset containing images, abnormality types, and descriptions verified by medical experts. VisTA produces four outputs: predicted abnormality type, similarity to reference cases, evidence-driven explanation, and final AD diagnoses. To illustrate VisTA's efficacy, we reported accuracy metrics for abnormality retrieval and dementia prediction. To demonstrate VisTA's explainability, we compared its explanations with human experts' explanations. Results: Compared to 15 million images used for baseline pretraining, VisTA only used 170 samples for fine-tuning and obtained significant improvement in abnormality retrieval and dementia prediction. For abnormality retrieval, VisTA reached 74% accuracy and an AUC of 0.87 (26% and 0.74, respectively, from baseline models). For dementia prediction, VisTA achieved 88% accuracy and an AUC of 0.82 (30% and 0.57, respectively, from baseline models). The generated explanations agreed strongly with human experts' and provided insights into the diagnostic process. Taken together, VisTA optimize prediction, clinical reasoning, and explanation.

The automatic reconstruction of 3D computer-aided design (CAD) models from CAD sketches has recently gained significant attention in the computer vision community. Most existing methods, however, rely on vector CAD sketches and 3D ground truth for supervision, which are often difficult to be obtained in industrial applications and are sensitive to noise inputs. We propose viewing CAD reconstruction as a specific instance of sparse-view 3D reconstruction to overcome these limitations. While this reformulation offers a promising perspective, existing 3D reconstruction methods typically require natural images and corresponding camera poses as inputs, which introduces two major significant challenges: (1) modality discrepancy between CAD sketches and natural images, and (2) difficulty of accurate camera pose estimation for CAD sketches. To solve these issues, we first transform the CAD sketches into representations resembling natural images and extract corresponding masks. Next, we manually calculate the camera poses for the orthographic views to ensure accurate alignment within the 3D coordinate system. Finally, we employ a customized sparse-view 3D reconstruction method to achieve high-quality reconstructions from aligned orthographic views. By leveraging raster CAD sketches for self-supervision, our approach eliminates the reliance on vector CAD sketches and 3D ground truth. Experiments on the Sub-Fusion360 dataset demonstrate that our proposed method significantly outperforms previous approaches in CAD reconstruction performance and exhibits strong robustness to noisy inputs.

The goal of this note is twofold: first, we explain the relation between the isomorphism theorems in the context of vertex reinforced jump process discovered in [BHS19, BHS21] and the standard Markovian isomorphism theorems for Markovian jump processes; second, we introduce the vertex reinforced counterpart of the standard Poissonian loop soup developed by Le Jan [LJ10]. To this end, we propose an algorithm that can be viewed as a variant of Wilson's algorithm with reinforcement. We establish the isomorphism theorems for the erased loops and the random walk from this algorithm, and in particular provide a concrete construction of the reinforced loop soup via a random process with a reinforcement mechanism., Comment: This article supersedes arXiv:1911.09036

Electromagnetic penetration through openings in an otherwise closed electromagnetic scatterer is an important topic in computational electromagnetics. To efficiently model this phenomenon, aperture or slot models are often used in conjunction with surface integral equations to solve Maxwell's equations. To establish the credibility of these models, code verification is necessary to assess the correctness of the implementation of the underlying numerical methods. However, many characteristics of surface integral equations and slot models render traditional code-verification approaches ineffective. In this paper, we present approaches to separately measure the different sources of numerical error arising from the method-of-moments implementation of the electric-field integral equation with an arbitrary-depth slot model. We demonstrate the effectiveness of these approaches for several cases., Comment: arXiv admin note: substantial text overlap with arXiv:2406.14573

In this paper, we systematically investigate the structural and electronic properties of twisted h-BN bilayers to understand the role of the twisting angle. Using first-principles methods with relaxation taken into account, we simulate h-BN bilayers with commensurate supercells with the smallest angle being $2.88^{\circ}$ until $60^{\circ}$. We find that the interlayer separation is not constant throughout each bilayer because of the various stacking patterns of AA, AA', AB, AB', and A'B throughout the layers, which also play a significant role in their unique charge redistribution. The calculations for the 110 generated structures show the existence of flat bands in several twisted h-BN bilayers, as well as the emergence of different trends in their properties as a function of the twist angle. These results are useful for establishing a systematic base line of registry-dependent relations for the development of more advanced computational methods to access incommensurate h-BN bilayers., Comment: 9 pages, 6 figures, submitted

We report the observations of the radio shell of the supernova remnant (SNR) G310.6-1.6 at 943 MHz from the Evolutionary Map of the Universe (EMU) and the Polarization Sky Survey of the Universe's Magnetism (POSSUM) surveys by using the Australian Square Kilometre Array Pathfinder (ASKAP). We detect polarized emission from the central pulsar wind nebula (PWN) with rotation measures varying from -696 rad m$^{-2}$ to -601 rad m$^{-2}$. We measure the integrated flux density of the shell to be 36.4 +/- 2.2 mJy at 943 MHz and derive a spectral index of -0.4 +/- 0.1 for the PWN and -0.7 +/- 0.3 for the SNR shell. From the combined radio and X-ray observations, the object can be identified as a supernova explosion of about 2500 yr ago with energy of about 1.3 x 10$^{50}$ erg, suggesting an ejected mass of about 10 M_sun. The circular radio shell outside the circular hard X-ray shell is unique among Galactic SNRs. We discuss several possible scenarios, including blast wave, reverse shock, and pulsar-fed emission, but find that none of them can fully explain the observed characteristics of the shell. This poses a challenge for understanding the evolution of SNRs. The results of this paper demonstrate the potential of the ASKAP EMU and POSSUM surveys in discovering more objects of small angular size and low surface brightness., Comment: 16 pages, 10 figures, published in ApJ 980, 162

The rapid advancement of information and communication technology has facilitated easier access to information. However, this progress has also necessitated more stringent verification measures to ensure the accuracy of information, particularly within the context of Vietnam. This paper introduces an approach to address the challenges of Fact Verification using the Vietnamese dataset by integrating both sentence selection and classification modules into a unified network architecture. The proposed approach leverages the power of large language models by utilizing pre-trained PhoBERT and XLM-RoBERTa as the backbone of the network. The proposed model was trained on a Vietnamese dataset, named ISE-DSC01, and demonstrated superior performance compared to the baseline model across all three metrics. Notably, we achieved a Strict Accuracy level of 75.11\%, indicating a remarkable 28.83\% improvement over the baseline model., Comment: accepted for Oral Presentation in CITA 2024 (The 13th Conference on Information Technology and Its Applications) and will be published in VOLUME 1 OF CITA 2024 (Volume of the Lecture Notes in Network and Systems, Springer)

Linguistic richness is essential for advancing natural language processing (NLP), as dataset characteristics often directly influence model performance. However, traditional metrics such as Type-Token Ratio (TTR), Vocabulary Diversity (VOCD), and Measure of Lexical Text Diversity (MTLD) do not adequately capture contextual relationships, semantic richness, and structural complexity. In this paper, we introduce an autoencoder-based framework that uses neural network capacity as a proxy for vocabulary richness, diversity, and complexity, enabling a dynamic assessment of the interplay between vocabulary size, sentence structure, and contextual depth. We validate our approach on two distinct datasets: the DIFrauD dataset, which spans multiple domains of deceptive and fraudulent text, and the Project Gutenberg dataset, representing diverse languages, genres, and historical periods. Experimental results highlight the robustness and adaptability of our method, offering practical guidance for dataset curation and NLP model design. By enhancing traditional vocabulary evaluation, our work fosters the development of more context-aware, linguistically adaptive NLP systems., Comment: Extended version of "Vocabulary Quality in NLP Datasets: An Autoencoder-Based Framework Across Domains and Languages" in IDA 2025

Image deblurring aims to restore high-quality images from blurred ones. While existing deblurring methods have made significant progress, most overlook the fact that the degradation degree varies across different regions. In this paper, we propose AIBNet, a network that adaptively identifies the blurred regions, enabling differential restoration of these regions. Specifically, we design a spatial feature differential handling block (SFDHBlock), with the core being the spatial domain feature enhancement module (SFEM). Through the feature difference operation, SFEM not only helps the model focus on the key information in the blurred regions but also eliminates the interference of implicit noise. Additionally, based on the fact that the difference between sharp and blurred images primarily lies in the high-frequency components, we propose a high-frequency feature selection block (HFSBlock). The HFSBlock first uses learnable filters to extract high-frequency features and then selectively retains the most important ones. To fully leverage the decoder's potential, we use a pre-trained model as the encoder and incorporate the above modules only in the decoder. Finally, to alleviate the resource burden during training, we introduce a progressive training strategy. Extensive experiments demonstrate that our AIBNet achieves superior performance in image deblurring.

Recent top-performing temporal 3D detectors based on Lidars have increasingly adopted region-based paradigms. They first generate coarse proposals, followed by encoding and fusing regional features. However, indiscriminate sampling and fusion often overlook the varying contributions of individual points and lead to exponentially increased complexity as the number of input frames grows. Moreover, arbitrary result-level concatenation limits the global information extraction. In this paper, we propose a Focal Token Acquring-and-Scaling Transformer (FASTer), which dynamically selects focal tokens and condenses token sequences in an adaptive and lightweight manner. Emphasizing the contribution of individual tokens, we propose a simple but effective Adaptive Scaling mechanism to capture geometric contexts while sifting out focal points. Adaptively storing and processing only focal points in historical frames dramatically reduces the overall complexity. Furthermore, a novel Grouped Hierarchical Fusion strategy is proposed, progressively performing sequence scaling and Intra-Group Fusion operations to facilitate the exchange of global spatial and temporal information. Experiments on the Waymo Open Dataset demonstrate that our FASTer significantly outperforms other state-of-the-art detectors in both performance and efficiency while also exhibiting improved flexibility and robustness. The code is available at https://github.com/MSunDYY/FASTer.git., Comment: 10pages,6 figures

Dynamic Algorithm Configuration (DAC) has garnered significant attention in recent years, particularly in the prevalence of machine learning and deep learning algorithms. Numerous studies have leveraged the robustness of decision-making in Reinforcement Learning (RL) to address the optimization challenges associated with algorithm configuration. However, making an RL agent work properly is a non-trivial task, especially in reward design, which necessitates a substantial amount of handcrafted knowledge based on domain expertise. In this work, we study the importance of reward design in the context of DAC via a case study on controlling the population size of the $(1+(\lambda,\lambda))$-GA optimizing OneMax. We observed that a poorly designed reward can hinder the RL agent's ability to learn an optimal policy because of a lack of exploration, leading to both scalability and learning divergence issues. To address those challenges, we propose the application of a reward shaping mechanism to facilitate enhanced exploration of the environment by the RL agent. Our work not only demonstrates the ability of RL in dynamically configuring the $(1+(\lambda,\lambda))$-GA, but also confirms the advantages of reward shaping in the scalability of RL agents across various sizes of OneMax problems.

Image deblurring aims to restore high-quality images by removing undesired degradation. Although existing methods have yielded promising results, they either overlook the varying degrees of degradation across different regions of the blurred image, or they approximate nonlinear function properties by stacking numerous nonlinear activation functions. In this paper, we propose a differential handling network (DHNet) to perform differential processing for different blur regions. Specifically, we design a Volterra block (VBlock) to integrate the nonlinear characteristics into the deblurring network, avoiding the previous operation of stacking the number of nonlinear activation functions to map complex input-output relationships. To enable the model to adaptively address varying degradation degrees in blurred regions, we devise the degradation degree recognition expert module (DDRE). This module initially incorporates prior knowledge from a well-trained model to estimate spatially variable blur information. Consequently, the router can map the learned degradation representation and allocate weights to experts according to both the degree of degradation and the size of the regions. Comprehensive experimental results show that DHNet effectively surpasses state-of-the-art (SOTA) methods on both synthetic and real-world datasets.

In wireless communication networks, it is difficult to solve many NP-hard problems owing to computational complexity and high cost. Recently, quantum annealing (QA) based on quantum physics was introduced as a key enabler for solving optimization problems quickly. However, only some studies consider quantum-based approaches in wireless communications. Therefore, we investigate the performance of a QA solution to an optimization problem in wireless networks. Specifically, we aim to maximize the sum rate by jointly optimizing clustering, sub-channel assignment, and power allocation in a multi-unmanned aerial vehicle-aided wireless network. We formulate the sum rate maximization problem as a combinatorial optimization problem. Then, we divide it into two sub-problems: 1) a QA-based clustering and 2) sub-channel assignment and power allocation for a given clustering configuration. Subsequently, we obtain an optimized solution for the joint optimization problem by solving these two sub-problems. For the first sub-problem, we convert the problem into a simplified quadratic unconstrained binary optimization (QUBO) model. As for the second sub-problem, we introduce a novel QA algorithm with optimal scaling parameters to address it. Simulation results demonstrate the effectiveness of the proposed algorithm in terms of the sum rate and running time., Comment: 15 pages, 9 figures, and 2 tables. Accepted by IEEE IoT Journal

We propose and experimentally demonstrate the hybridization of radiating topological interface states, analogous to Jackiw-Rebbi states but in gain media with radiation fields. This hybridization not only modifies energy levels under a strong coupling scheme but also significantly reshapes far-field radiation characteristics. The bonding mode exhibits sub-radiant, omnidirectional emission, while the antibonding mode becomes super-radiant and highly unidirectional. Crucially, this non-Hermitian hybridization is tunable, allowing simultaneous control of energy splitting, quality factor, and far-field radiation by varying the distance between the two topological interfaces. Our findings establish hybridized radiating topological interface states as a robust platform for engineering two-level systems with tailored far-field responses, offering new possibilities for applications in beam shaping, nonlinear optics, quantum technologies, and beyond.

Synthetic data has the potential to improve the performance, training efficiency, and privacy of real training examples. Nevertheless, existing approaches for synthetic text generation are mostly heuristics and cannot generate human-readable text without compromising the privacy of real data or provide performance guarantees for training Large Language Models (LLMs). In this work, we propose the first theoretically rigorous approach for generating synthetic human-readable text that guarantees the convergence and performance of LLMs during fine-tuning on a target task. To do so, we leverage Alternating Direction Method of Multipliers (ADMM) that iteratively optimizes the embeddings of synthetic examples to match the gradient of the target training or validation data, and maps them to a sequence of text tokens with low perplexity. In doing so, the generated synthetic text can guarantee convergence of the model to a close neighborhood of the solution obtained by fine-tuning on real data. Experiments on various classification tasks confirm the effectiveness of our proposed approach., Comment: 15 pages, 5 figures, 4 tables

In this paper, we prove some Paley-Wiener theorems for function spaces consisting of slice monogenic functions such as Paley-Wiener, Hardy and Bergman spaces. As applications, we can compute the reproducing kernel functions for the related function spaces., Comment: 33 pages

We prove two theorems of Paley and Wiener in the slice regular setting. As an application, we can compute the reproducing kernel for the slice regular Paley-Wiener space, and obtain a related sampling theorem., Comment: 29 pages

We present a framework for generating universal semantic embeddings of chemical elements to advance materials inference and discovery. This framework leverages ElementBERT, a domain-specific BERT-based natural language processing model trained on 1.29 million abstracts of alloy-related scientific papers, to capture latent knowledge and contextual relationships specific to alloys. These semantic embeddings serve as robust elemental descriptors, consistently outperforming traditional empirical descriptors with significant improvements across multiple downstream tasks. These include predicting mechanical and transformation properties, classifying phase structures, and optimizing materials properties via Bayesian optimization. Applications to titanium alloys, high-entropy alloys, and shape memory alloys demonstrate up to 23% gains in prediction accuracy. Our results show that ElementBERT surpasses general-purpose BERT variants by encoding specialized alloy knowledge. By bridging contextual insights from scientific literature with quantitative inference, our framework accelerates the discovery and optimization of advanced materials, with potential applications extending beyond alloys to other material classes., Comment: 5 figures

Image deblurring aims to reconstruct a latent sharp image from its corresponding blurred one. Although existing methods have achieved good performance, most of them operate exclusively in either the spatial domain or the frequency domain, rarely exploring solutions that fuse both domains. In this paper, we propose a spatial-frequency domain adaptive fusion network (SFAFNet) to address this limitation. Specifically, we design a gated spatial-frequency domain feature fusion block (GSFFBlock), which consists of three key components: a spatial domain information module, a frequency domain information dynamic generation module (FDGM), and a gated fusion module (GFM). The spatial domain information module employs the NAFBlock to integrate local information. Meanwhile, in the FDGM, we design a learnable low-pass filter that dynamically decomposes features into separate frequency subbands, capturing the image-wide receptive field and enabling the adaptive exploration of global contextual information. Additionally, to facilitate information flow and the learning of complementary representations. In the GFM, we present a gating mechanism (GATE) to re-weight spatial and frequency domain features, which are then fused through the cross-attention mechanism (CAM). Experimental results demonstrate that our SFAFNet performs favorably compared to state-of-the-art approaches on commonly used benchmarks.

We introduce Qwen2.5-VL, the latest flagship model of Qwen vision-language series, which demonstrates significant advancements in both foundational capabilities and innovative functionalities. Qwen2.5-VL achieves a major leap forward in understanding and interacting with the world through enhanced visual recognition, precise object localization, robust document parsing, and long-video comprehension. A standout feature of Qwen2.5-VL is its ability to localize objects using bounding boxes or points accurately. It provides robust structured data extraction from invoices, forms, and tables, as well as detailed analysis of charts, diagrams, and layouts. To handle complex inputs, Qwen2.5-VL introduces dynamic resolution processing and absolute time encoding, enabling it to process images of varying sizes and videos of extended durations (up to hours) with second-level event localization. This allows the model to natively perceive spatial scales and temporal dynamics without relying on traditional normalization techniques. By training a native dynamic-resolution Vision Transformer (ViT) from scratch and incorporating Window Attention, we reduce computational overhead while maintaining native resolution. As a result, Qwen2.5-VL excels not only in static image and document understanding but also as an interactive visual agent capable of reasoning, tool usage, and task execution in real-world scenarios such as operating computers and mobile devices. Qwen2.5-VL is available in three sizes, addressing diverse use cases from edge AI to high-performance computing. The flagship Qwen2.5-VL-72B model matches state-of-the-art models like GPT-4o and Claude 3.5 Sonnet, particularly excelling in document and diagram understanding. Additionally, Qwen2.5-VL maintains robust linguistic performance, preserving the core language competencies of the Qwen2.5 LLM.

Data quantity and quality play a vital role in determining the performance of Large Language Models (LLMs). High-quality data, in particular, can significantly boost the LLM's ability to generalize on a wide range of downstream tasks. Large pre-training datasets for leading LLMs remain inaccessible to the public, whereas many open datasets are small in size (less than 5 trillion tokens), limiting their suitability for training large models. In this paper, we introduce GneissWeb, a large dataset yielding around 10 trillion tokens that caters to the data quality and quantity requirements of training LLMs. Our GneissWeb recipe that produced the dataset consists of sharded exact sub-string deduplication and a judiciously constructed ensemble of quality filters. GneissWeb achieves a favorable trade-off between data quality and quantity, producing models that outperform models trained on state-of-the-art open large datasets (5+ trillion tokens). We show that models trained using GneissWeb dataset outperform those trained on FineWeb-V1.1.0 by 2.73 percentage points in terms of average score computed on a set of 11 commonly used benchmarks (both zero-shot and few-shot) for pre-training dataset evaluation. When the evaluation set is extended to 20 benchmarks (both zero-shot and few-shot), models trained using GneissWeb still achieve a 1.75 percentage points advantage over those trained on FineWeb-V1.1.0.

Consider the compact orbits of the $\mathbb{R}^2$ action of the diagonal group on $\operatorname{SL}(3,\mathbb{R})/\operatorname{SL}(3,\mathbb{Z})$, the so-called periodic tori. For any periodic torus, the set of periods of the orbit forms a lattice in $\mathbb{R}^2$. Such a lattice, re-scaled to covolume one, gives a shape point in $\operatorname{SL}(2,\mathbb{R})/\operatorname{SL}(2,\mathbb{Z})$. We prove that the shapes of all periodic tori are dense in $\operatorname{SL}(2,\mathbb{R})/\operatorname{SL}(2,\mathbb{Z})$. This implies the density of shapes of the unit groups of totally real cubic orders.

The current generation of large language models (LLMs) is typically designed for broad, general-purpose applications, while domain-specific LLMs, especially in vertical fields like medicine, remain relatively scarce. In particular, the development of highly efficient and practical LLMs for the medical domain is challenging due to the complexity of medical knowledge and the limited availability of high-quality data. To bridge this gap, we introduce Baichuan-M1, a series of large language models specifically optimized for medical applications. Unlike traditional approaches that simply continue pretraining on existing models or apply post-training to a general base model, Baichuan-M1 is trained from scratch with a dedicated focus on enhancing medical capabilities. Our model is trained on 20 trillion tokens and incorporates a range of effective training methods that strike a balance between general capabilities and medical expertise. As a result, Baichuan-M1 not only performs strongly across general domains such as mathematics and coding but also excels in specialized medical fields. We have open-sourced Baichuan-M1-14B, a mini version of our model, which can be accessed through the following links., Comment: 33 pages, technical report

Metacognitive education plays a crucial role in cultivating students' self-regulation and reflective thinking, providing essential support for those with learning difficulties through academic advising. Simulating students with insufficient learning capabilities using large language models offers a promising approach to refining pedagogical methods without ethical concerns. However, existing simulations often fail to authentically represent students' learning struggles and face challenges in evaluation due to the lack of reliable metrics and ethical constraints in data collection. To address these issues, we propose a pipeline for automatically generating and filtering high-quality simulated student agents. Our approach leverages a two-round automated scoring system validated by human experts and employs a score propagation module to obtain more consistent scores across the student graph. Experimental results demonstrate that our pipeline efficiently identifies high-quality student agents, and we discuss the traits that influence the simulation's effectiveness. By simulating students with varying degrees of learning difficulties, our work paves the way for broader applications in personalized learning and educational assessment., Comment: 12 pages, 8 figures

TianQin is a future space-based gravitational wave observatory targeting the frequency window of $10^{-4}$ Hz $\sim 1$ Hz. A large variety of gravitational wave sources are expected in this frequency band, including the merger of massive black hole binaries, the inspiral of extreme/intermediate mass ratio systems, stellar-mass black hole binaries, Galactic compact binaries, and so on. TianQin will consist of three Earth orbiting satellites on nearly identical orbits with orbital radii of about $10^5$ km. The satellites will form a normal triangle constellation whose plane is nearly perpendicular to the ecliptic plane. The TianQin project has been progressing smoothly following the ``0123" technology roadmap. In step ``0", the TianQin laser ranging station has been constructed and it has successfully ranged to all the five retro-reflectors on the Moon. In step ``1", the drag-free control technology has been tested and demonstrated using the TianQin-1 satellite. In step ``2", the inter-satellite laser interferometry technology will be tested using the pair of TianQin-2 satellites. The TianQin-2 mission has been officially approved and the satellites will be launched around 2026. In step ``3", i.e., the TianQin-3 mission, three identical satellites will be launched around 2035 to form the space-based gravitational wave detector, TianQin, and to start gravitational wave detection in space., Comment: 45 pages, 3 figures

Fingerspelling is a critical part of American Sign Language (ASL) recognition and has become an accessible optional text entry method for Deaf and Hard of Hearing (DHH) individuals. In this paper, we introduce SpellRing, a single smart ring worn on the thumb that recognizes words continuously fingerspelled in ASL. SpellRing uses active acoustic sensing (via a microphone and speaker) and an inertial measurement unit (IMU) to track handshape and movement, which are processed through a deep learning algorithm using Connectionist Temporal Classification (CTC) loss. We evaluated the system with 20 ASL signers (13 fluent and 7 learners), using the MacKenzie-Soukoref Phrase Set of 1,164 words and 100 phrases. Offline evaluation yielded top-1 and top-5 word recognition accuracies of 82.45% (9.67%) and 92.42% (5.70%), respectively. In real-time, the system achieved a word error rate (WER) of 0.099 (0.039) on the phrases. Based on these results, we discuss key lessons and design implications for future minimally obtrusive ASL recognition wearables.

Despite a strong theoretical foundation, empirical experiments reveal that existing domain generalization (DG) algorithms often fail to consistently outperform the ERM baseline. We argue that this issue arises because most DG studies focus on establishing theoretical guarantees for generalization under unrealistic assumptions, such as the availability of sufficient, diverse (or even infinite) domains or access to target domain knowledge. As a result, the extent to which domain generalization is achievable in scenarios with limited domains remains largely unexplored. This paper seeks to address this gap by examining generalization through the lens of the conditions necessary for its existence and learnability. Specifically, we systematically establish a set of necessary and sufficient conditions for generalization. Our analysis highlights that existing DG methods primarily act as regularization mechanisms focused on satisfying sufficient conditions, while often neglecting necessary ones. However, sufficient conditions cannot be verified in settings with limited training domains. In such cases, regularization targeting sufficient conditions aims to maximize the likelihood of generalization, whereas regularization targeting necessary conditions ensures its existence. Using this analysis, we reveal the shortcomings of existing DG algorithms by showing that, while they promote sufficient conditions, they inadvertently violate necessary conditions. To validate our theoretical insights, we propose a practical method that promotes the sufficient condition while maintaining the necessary conditions through a novel subspace representation alignment strategy. This approach highlights the advantages of preserving the necessary conditions on well-established DG benchmarks.

Parallel Tempering (PT) is a classical MCMC algorithm designed for leveraging parallel computation to sample efficiently from high-dimensional, multimodal or otherwise complex distributions via annealing. One limitation of the standard formulation of PT is the growth of computational resources required to generate high-quality samples, as measured by effective sample size or round trip rate, for increasingly challenging distributions. To address this issue, we propose the framework: Generalised Parallel Tempering (GePT) which allows for the incorporation of recent advances in modern generative modelling, such as normalising flows and diffusion models, within Parallel Tempering, while maintaining the same theoretical guarantees as MCMC-based methods. For instance, we show that this allows us to utilise diffusion models in a parallelised manner, bypassing the usual computational cost of a large number of steps to generate quality samples. Further, we empirically demonstrate that GePT can improve sample quality and reduce the growth of computational resources required to handle complex distributions over the classical algorithm.

The deployment of advanced deep learning models for medical image segmentation is often constrained by the requirement for extensively annotated datasets. Weakly-supervised learning, which allows less precise labels, has become a promising solution to this challenge. Building on this approach, we propose QMaxViT-Unet+, a novel framework for scribble-supervised medical image segmentation. This framework is built on the U-Net architecture, with the encoder and decoder replaced by Multi-Axis Vision Transformer (MaxViT) blocks. These blocks enhance the model's ability to learn local and global features efficiently. Additionally, our approach integrates a query-based Transformer decoder to refine features and an edge enhancement module to compensate for the limited boundary information in the scribble label. We evaluate the proposed QMaxViT-Unet+ on four public datasets focused on cardiac structures, colorectal polyps, and breast cancer: ACDC, MS-CMRSeg, SUN-SEG, and BUSI. Evaluation metrics include the Dice similarity coefficient (DSC) and the 95th percentile of Hausdorff distance (HD95). Experimental results show that QMaxViT-Unet+ achieves 89.1\% DSC and 1.316mm HD95 on ACDC, 88.4\% DSC and 2.226mm HD95 on MS-CMRSeg, 71.4\% DSC and 4.996mm HD95 on SUN-SEG, and 69.4\% DSC and 50.122mm HD95 on BUSI. These results demonstrate that our method outperforms existing approaches in terms of accuracy, robustness, and efficiency while remaining competitive with fully-supervised learning approaches. This makes it ideal for medical image analysis, where high-quality annotations are often scarce and require significant effort and expense. The code is available at: https://github.com/anpc849/QMaxViT-Unet

Large Multimodal Models (LMMs) exhibit major shortfalls when interpreting images and, by some measures, have poorer spatial cognition than small children or animals. Despite this, they attain high scores on many popular visual benchmarks, with headroom rapidly eroded by an ongoing surge of model progress. To address this, there is a pressing need for difficult benchmarks that remain relevant for longer. We take this idea to its limit by introducing ZeroBench-a lightweight visual reasoning benchmark that is entirely impossible for contemporary frontier LMMs. Our benchmark consists of 100 manually curated questions and 334 less difficult subquestions. We evaluate 20 LMMs on ZeroBench, all of which score 0.0%, and rigorously analyse the errors. To encourage progress in visual understanding, we publicly release ZeroBench., Comment: 20 pages, 13 figures

Recent advances in conditional diffusion models have shown promise for generating realistic TalkingFace videos, yet challenges persist in achieving consistent head movement, synchronized facial expressions, and accurate lip synchronization over extended generations. To address these, we introduce the \textbf{M}otion-priors \textbf{C}onditional \textbf{D}iffusion \textbf{M}odel (\textbf{MCDM}), which utilizes both archived and current clip motion priors to enhance motion prediction and ensure temporal consistency. The model consists of three key elements: (1) an archived-clip motion-prior that incorporates historical frames and a reference frame to preserve identity and context; (2) a present-clip motion-prior diffusion model that captures multimodal causality for accurate predictions of head movements, lip sync, and expressions; and (3) a memory-efficient temporal attention mechanism that mitigates error accumulation by dynamically storing and updating motion features. We also release the \textbf{TalkingFace-Wild} dataset, a multilingual collection of over 200 hours of footage across 10 languages. Experimental results demonstrate the effectiveness of MCDM in maintaining identity and motion continuity for long-term TalkingFace generation. Code, models, and datasets will be publicly available.

Splitting methods constitute a well-established class of numerical schemes for solving convection-diffusion-reaction problems. They have been shown to be effective in solving problems with periodic boundary conditions. However, in the case of Dirichlet boundary conditions, order reduction has been observed even with homogeneous boundary conditions. In this paper, we propose a novel splitting approach, the so-called `initial-corrected splitting method', which succeeds in overcoming order reduction. A convergence analysis is performed to demonstrate second-order convergence of this modified Strang splitting method. Furthermore, we conduct numerical experiments to illustrate the performance of the newly developed splitting approach.

We utilized archived data from the Five-hundred-meter Aperture Spherical Radio Telescope (FAST) to analyze the single-pulse profile morphology of PSR J1935$+$1616 (B1933$+$16). The results show that PSR J1935$+$1616 exhibits significant micropulses as well as various changes in single-pulse profile morphology. In the FAST archived data, a total of 969 single pulses with microstructure were identified, accounting for 9.69$\%$ of the total pulse sample, with characteristic widths of $127.63^{+70.74}_{-46.25}$ $\mu$s. About half of these pulses display quasiperiodic micropulses, with a periodicity of 231.77 $\pm$ 9.90 $\mu$s. Among the 520 single pulses with quasiperiodic microstructure, 208 also exhibit quasiperiodicity in circular polarization, with a characteristic period of $244.70^{+45.66}_{-21.05}$ $\mu$s. The micropulse characteristic width in circular polarization is 106.52 $\pm$ 46.14 $\mu$s. Compared to normal pulses, the relative energy (E/) of single pulse with microstructure follows a double Gaussian distribution, while that of normal pulses follows a single Gaussian distribution. Based on the intensity of the leading and trailing components in the single-pulse profile morphology of PSR J1935+1616, we classified the pulses into four morphological modes (A, B, C, and D). The relative energy distribution of pulses in mode A is significantly different from the others, following a double Gaussian distribution, while the relative energy distributions in modes B, C, and D follow a single Gaussian distribution. Our study also suggests a possible correlation between micropulses and single-pulse profile morphology. Single pulse with micropulses are most likely to occur in mode A, while their occurrence is least likely in mode D.

Bringing increasingly complex polyatomic molecules within reach of precision measurement experiments offers fascinating and far-reaching prospects ranging from Earth sciences and astrophysics, to metrology and quantum sciences. Here, we demonstrate sub-Doppler spectroscopic measurements in the mid-IR fingerprint region of, to our knowledge, the largest molecule to date. To this end, we use a high-resolution ~10.3 $\mu$m spectrometer based on a sub-Hz quantum cascade laser remotely calibrated against state-of-the-art primary frequency standards via a metrology-grade fibre link. We perform saturated absorption spectroscopy in the v5 CO stretching mode of 1,3,5-trioxane, (H2CO)3, at a resolution of ~100 kHz, allowing us to measure the absolute frequency of hundreds of rovibrational transitions at unprecedented uncertainties for such a complex species, as low as ~5 kHz. Our work demonstrates the extension of frequency metrology methodologies to ever larger molecular system, confirming the potential of the technologies we develop for bringing increasingly complex species within reach of ultra-precise measurement experiments.

Ride-sharing services are revolutionizing urban mobility while simultaneously raising significant concerns regarding fairness and driver equity. This study employs Chicago Trip Network Provider dataset to investigate disparities in ride-sharing earnings between 2018 and 2023. Our analysis reveals marked temporal shifts, including an earnings surge in early 2021 followed by fluctuations and a decline in inflation-adjusted income, as well as pronounced spatial disparities, with drivers in Central and airport regions earning substantially more than those in peripheral areas. Recognizing the limitations of trip-level data, we introduce a novel trip-driver assignment algorithm to reconstruct plausible daily work patterns, uncovering distinct driver clusters with varied earning profiles. Notably, drivers operating during late-evening and overnight hours secure higher per-trip and hourly rates, while emerging groups in low-demand regions face significant earnings deficits. Our findings call for more transparent pricing models and a re-examination of platform design to promote equitable driver outcomes.

Diffusion bridges have shown potential in paired image-to-image (I2I) translation tasks. However, existing methods are limited by their unidirectional nature, requiring separate models for forward and reverse translations. This not only doubles the computational cost but also restricts their practicality. In this work, we introduce the Bidirectional Diffusion Bridge Model (BDBM), a scalable approach that facilitates bidirectional translation between two coupled distributions using a single network. BDBM leverages the Chapman-Kolmogorov Equation for bridges, enabling it to model data distribution shifts across timesteps in both forward and backward directions by exploiting the interchangeability of the initial and target timesteps within this framework. Notably, when the marginal distribution given endpoints is Gaussian, BDBM's transition kernels in both directions possess analytical forms, allowing for efficient learning with a single network. We demonstrate the connection between BDBM and existing bridge methods, such as Doob's h-transform and variational approaches, and highlight its advantages. Extensive experiments on high-resolution I2I translation tasks demonstrate that BDBM not only enables bidirectional translation with minimal additional cost but also outperforms state-of-the-art bridge models. Our source code is available at [https://github.com/kvmduc/BDBM||https://github.com/kvmduc/BDBM]., Comment: Source code: https://github.com/kvmduc/BDBM

Pathology plays a critical role in diagnosing a wide range of diseases, yet existing approaches often rely heavily on task-specific models trained on extensive, well-labeled datasets. These methods face sustainability challenges due to the diversity of pathologies and the labor-intensive nature of data collection. To address these limitations, we highlight the need for universal multimodal embeddings that can support multiple downstream tasks. Previous approaches often involve fine-tuning CLIP-based models, which handle images and text separately, limiting their ability to capture complex multimodal relationships. Additionally, these models are evaluated across diverse datasets without a unified benchmark for assessing multimodal embeddings in pathology. To address these challenges, we propose MLLM4PUE, a novel framework that leverages Multimodal Large Language Models (MLLMs) to generate Pathology Universal Embeddings. The MLLM4PUE framework not only facilitates robust integration of images and text but also enhances understanding and fusion capabilities across various tasks. We further introduce the Pathology Multimodal Embedding Benchmark (PMEB), a comprehensive benchmark designed to assess the quality of pathology multimodal embeddings. PMEB comprises 15 original tasks drawn from 14 datasets, organized into three meta-tasks: retrieval, classification, and composed retrieval. Experimental results demonstrate the superiority of MLLM4PUE, illustrating MLLM-based models can effectively support a wide range of downstream tasks and unify the research direction for foundation models in pathology.

This study presents the Music Informatics for Radio Across the GlobE (MIRAGE) online dashboard, which allows users to access, interact with, and export metadata (e.g., artist name, track title) and musicological features (e.g., instrument list, voice type, key/mode) for 1 million events streaming on 10,000 internet radio stations across the globe. Users can search for stations or events according to several criteria, display, analyze, and listen to the selected station/event lists using interactive visualizations that include embedded links to streaming services, and finally export relevant metadata and visualizations for further study., Comment: 7 pages, 5 figures, 1 table

We detail the emission behaviors of three long-period pulsars detected using the Five-hundred-meter Aperture Spherical radio Telescope (FAST) during the CRAFTS survey. Their rotational periods range from 1.83 s to 4.75 s, and the null fractions measure between 28% and 53%. PSR J1945+1211 and PSR J2323+1214 exhibited quasi-periodic nulls, with duration of around 57 seconds. The longest null was observed in PSR J1945+1211, lasting 76 seconds. PSR J2323+1214 displayed varying null fractions between its leading and trailing components. For the first time in PSR J2323+1214, we detected five dwarf pulses, which are much weaker and narrower pulses than typical burst pulses. In addition, we investigate the microstructure of PSR J1900-0134 for the first time, revealing intricate pulses of up to 2.05 milliseconds and noting its complex emission characteristics. Bright pulses occur in all of these sources at different rates. These observations suggest complex magnetospheric processes, potentially related to magnetic reconnections, and provide insights into the origins of bright and microstructure pulses, as well as their distinctions from ordinary pulses., Comment: 36 pages, 23 figures

Astrophysical observations of the cosmos allow us to probe extreme physics and answer foundational questions on our universe. Modern astronomy is increasingly operating under a holistic approach, probing the same question with multiple diagnostics including how sources vary over time, how they appear across the electromagnetic spectrum, and through their other signatures, including gravitational waves, neutrinos, cosmic rays, and dust on Earth. Astrophysical observations are now reaching the point where approximate physics models are insufficient. Key sources of interest are explosive transients, whose understanding requires multidisciplinary studies at the intersection of astrophysics, gravity, nuclear science, plasma physics, fluid dynamics and turbulence, computation, particle physics, atomic, molecular, and optical science, condensed matter and materials science, radiation transport, and high energy density physics. This white paper provides an overview of the major scientific advances that lay at the intersection of physics and astronomy and are best probed through time-domain and multimessenger astrophysics, an exploration of how multidisciplinary science can be fostered, and introductory descriptions of the relevant scientific disciplines and key astrophysical sources of interest., Comment: This white paper is the product of the Third Time-Domain And Multimessenger Astrophysics workshop: Multidisciplinary Science in the Multimessenger Era, hosted by Louisiana State University with additional support from DOE, NSF, and NASA

Understanding the generalization properties of optimization algorithms under heavy-tailed noise has gained growing attention. However, the existing theoretical results mainly focus on stochastic gradient descent (SGD) and the analysis of heavy-tailed optimizers beyond SGD is still missing. In this work, we establish generalization bounds for SGD with momentum (SGDm) under heavy-tailed gradient noise. We first consider the continuous-time limit of SGDm, i.e., a Levy-driven stochastic differential equation (SDE), and establish quantitative Wasserstein algorithmic stability bounds for a class of potentially non-convex loss functions. Our bounds reveal a remarkable observation: For quadratic loss functions, we show that SGDm admits a worse generalization bound in the presence of heavy-tailed noise, indicating that the interaction of momentum and heavy tails can be harmful for generalization. We then extend our analysis to discrete-time and develop a uniform-in-time discretization error bound, which, to our knowledge, is the first result of its kind for SDEs with degenerate noise. This result shows that, with appropriately chosen step-sizes, the discrete dynamics retain the generalization properties of the limiting SDE. We illustrate our theory on both synthetic quadratic problems and neural networks., Comment: 64 pages, 2 figures

Let $(\mathcal X, d,\mu)$ be an RD-space, and let $\rho$ be an admissible function on $\mathcal X$. We establish necessary and sufficient conditions for the boundedness of a new class of generalized Calder\'on-Zygmund operators of log-Dini type on the Hardy space $H^1_\rho(\mathcal X)$, introduced by Yang and Zhou. Our results extend and unify some recent results, providing further insights into the study of singular integral operators in this setting., Comment: 28 pages, Banach J. Math. Anal. (to appear)

Representation Misdirection (RM) and variants are established large language model (LLM) unlearning methods with state-of-the-art performance. In this paper, we show that RM methods inherently reduce models' robustness, causing them to misbehave even when a single non-adversarial forget-token is in the retain-query. Toward understanding underlying causes, we reframe the unlearning process as backdoor attacks and defenses: forget-tokens act as backdoor triggers that, when activated in retain-queries, cause disruptions in RM models' behaviors, similar to successful backdoor attacks. To mitigate this vulnerability, we propose Random Noise Augmentation -- a model and method agnostic approach with theoretical guarantees for improving the robustness of RM methods. Extensive experiments demonstrate that RNA significantly improves the robustness of RM models while enhancing the unlearning performances., Comment: 12 pages, 4 figures, 1 table

The increasing complexity of modern software systems necessitates robust autonomic self-management capabilities. While Large Language Models (LLMs) demonstrate potential in this domain, they often face challenges in adapting their general knowledge to specific service contexts. To address this limitation, we propose ServiceOdyssey, a self-learning agent system that autonomously manages microservices without requiring prior knowledge of service-specific configurations. By leveraging curriculum learning principles and iterative exploration, ServiceOdyssey progressively develops a deep understanding of operational environments, reducing dependence on human input or static documentation. A prototype built with the Sock Shop microservice demonstrates the potential of this approach for autonomic microservice management.

Automating the generation of Planning Domain Definition Language (PDDL) with Large Language Model (LLM) opens new research topic in AI planning, particularly for complex real-world tasks. This paper introduces Image2PDDL, a novel framework that leverages Vision-Language Models (VLMs) to automatically convert images of initial states and descriptions of goal states into PDDL problems. By providing a PDDL domain alongside visual inputs, Imasge2PDDL addresses key challenges in bridging perceptual understanding with symbolic planning, reducing the expertise required to create structured problem instances, and improving scalability across tasks of varying complexity. We evaluate the framework on various domains, including standard planning domains like blocksworld and sliding tile puzzles, using datasets with multiple difficulty levels. Performance is assessed on syntax correctness, ensuring grammar and executability, and content correctness, verifying accurate state representation in generated PDDL problems. The proposed approach demonstrates promising results across diverse task complexities, suggesting its potential for broader applications in AI planning. We will discuss a potential use case in robot-assisted teaching of students with Autism Spectrum Disorder.

The advancement of large language models has intensified the need to modernize enterprise applications and migrate legacy systems to secure, versatile languages. However, existing code translation benchmarks primarily focus on individual functions, overlooking the complexities involved in translating entire repositories, such as maintaining inter-module coherence and managing dependencies. While some recent repository-level translation benchmarks attempt to address these challenges, they still face limitations, including poor maintainability and overly coarse evaluation granularity, which make them less developer-friendly. We introduce Skeleton-Guided-Translation, a framework for repository-level Java to C# code translation with fine-grained quality evaluation. It uses a two-step process: first translating the repository's structural "skeletons", then translating the full repository guided by these skeletons. Building on this, we present TRANSREPO-BENCH, a benchmark of high quality open-source Java repositories and their corresponding C# skeletons, including matching unit tests and build configurations. Our unit tests are fixed and can be applied across multiple or incremental translations without manual adjustments, enhancing automation and scalability in evaluations. Additionally, we develop fine-grained evaluation metrics that assess translation quality at the individual test case level, addressing traditional binary metrics' inability to distinguish when build failures cause all tests to fail. Evaluations using TRANSREPO-BENCH highlight key challenges and advance more accurate repository level code translation.

We introduce Qwen2.5-1M, a series of models that extend the context length to 1 million tokens. Compared to the previous 128K version, the Qwen2.5-1M series have significantly enhanced long-context capabilities through long-context pre-training and post-training. Key techniques such as long data synthesis, progressive pre-training, and multi-stage supervised fine-tuning are employed to effectively enhance long-context performance while reducing training costs. To promote the use of long-context models among a broader user base, we present and open-source our inference framework. This framework includes a length extrapolation method that can expand the model context lengths by at least four times, or even more, without additional training. To reduce inference costs, we implement a sparse attention method along with chunked prefill optimization for deployment scenarios and a sparsity refinement method to improve precision. Additionally, we detail our optimizations in the inference engine, including kernel optimization, pipeline parallelism, and scheduling optimization, which significantly enhance overall inference performance. By leveraging our inference framework, the Qwen2.5-1M models achieve a remarkable 3x to 7x prefill speedup in scenarios with 1 million tokens of context. This framework provides an efficient and powerful solution for developing applications that require long-context processing using open-source models. The Qwen2.5-1M series currently includes the open-source models Qwen2.5-7B-Instruct-1M and Qwen2.5-14B-Instruct-1M, as well as the API-accessed model Qwen2.5-Turbo. Evaluations show that Qwen2.5-1M models have been greatly improved in long-context tasks without compromising performance in short-context scenarios. Specifically, the Qwen2.5-14B-Instruct-1M model significantly outperforms GPT-4o-mini in long-context tasks and supports contexts eight times longer.

We introduce Baichuan-Omni-1.5, an omni-modal model that not only has omni-modal understanding capabilities but also provides end-to-end audio generation capabilities. To achieve fluent and high-quality interaction across modalities without compromising the capabilities of any modality, we prioritized optimizing three key aspects. First, we establish a comprehensive data cleaning and synthesis pipeline for multimodal data, obtaining about 500B high-quality data (text, audio, and vision). Second, an audio-tokenizer (Baichuan-Audio-Tokenizer) has been designed to capture both semantic and acoustic information from audio, enabling seamless integration and enhanced compatibility with MLLM. Lastly, we designed a multi-stage training strategy that progressively integrates multimodal alignment and multitask fine-tuning, ensuring effective synergy across all modalities. Baichuan-Omni-1.5 leads contemporary models (including GPT4o-mini and MiniCPM-o 2.6) in terms of comprehensive omni-modal capabilities. Notably, it achieves results comparable to leading models such as Qwen2-VL-72B across various multimodal medical benchmarks.

Recently, emotional speech generation and speaker cloning have garnered significant interest in text-to-speech (TTS). With the open-sourcing of codec language TTS models trained on massive datasets with large-scale parameters, adapting these general pre-trained TTS models to generate speech with specific emotional expressions and target speaker characteristics has become a topic of great attention. Common approaches, such as full and adapter-based fine-tuning, often overlook the specific contributions of model parameters to emotion and speaker control. Treating all parameters uniformly during fine-tuning, especially when the target data has limited content diversity compared to the pre-training corpus, results in slow training speed and an increased risk of catastrophic forgetting. To address these challenges, we propose a characteristic-specific partial fine-tuning strategy, short as CSP-FT. First, we use a weighted-sum approach to analyze the contributions of different Transformer layers in a pre-trained codec language TTS model for emotion and speaker control in the generated speech. We then selectively fine-tune the layers with the highest and lowest characteristic-specific contributions to generate speech with target emotional expression and speaker identity. Experimental results demonstrate that our method achieves performance comparable to, or even surpassing, full fine-tuning in generating speech with specific emotional expressions and speaker identities. Additionally, CSP-FT delivers approximately 2x faster training speeds, fine-tunes only around 8% of parameters, and significantly reduces catastrophic forgetting. Furthermore, we show that codec language TTS models perform competitively with self-supervised models in speaker identification and emotion classification tasks, offering valuable insights for developing universal speech processing models., Comment: 13 pages

Highly-irradiated gas giant exoplanets are predicted to show circulation patterns dominated by day-to-night heat transport and a spatial distribution of clouds that is driven by advection and local heating. Hot-Jupiters have been extensively studied from broadband phase-curve observations at infrared and optical wavelengths, but spectroscopic observations in the reflected light are rare and the regime of smaller and higher-metallicity ultra-hot planets, such as hot-Neptunes, remains largely unexplored to date. Here we present the phase-resolved reflected-light and thermal-emission spectroscopy of the ultra-hot Neptune LTT 9779b, obtained through observing its full phase-curve from 0.6 to 2.8 $\mu$m with JWST NIRISS/SOSS. We detect an asymmetric dayside in reflected light (3.1$\sigma$ significance) with highly-reflective white clouds on the western dayside (A = 0.79$\pm$0.15) and a much lower-albedo eastern dayside (A = 0.41$\pm$0.10), resulting in an overall dayside albedo of A = 0.50$\pm$0.07. The thermal phase curve is symmetric about the substellar point, with a dayside effective temperature of T$_\mathrm{eff,day}$ = 2,260$^{+40}_{-50}$ K and a cold nightside (T$_\mathrm{eff,night}$ <1,330 K at 3-$\sigma$ confidence), indicative of short radiative timescales. We propose an atmospheric circulation and cloud distribution regime in which heat is transported eastward from the dayside towards the cold nightside by an equatorial jet, leading to a colder western dayside where temperatures are sufficiently low for the condensation of silicate clouds., Comment: Accepted for publication