Search

Large language models (LLMs) have demonstrated impressive capabilities in various reasoning tasks but face significant challenges with complex, knowledge-intensive multi-hop queries, particularly those involving new or long-tail knowledge. Existing benchmarks often fail to fully address these challenges. To bridge this gap, we introduce MINTQA (Multi-hop Question Answering on New and Tail Knowledge), a comprehensive benchmark to evaluate LLMs' capabilities in multi-hop reasoning across four critical dimensions: question handling strategy, sub-question generation, retrieval-augmented generation, and iterative or dynamic decomposition and retrieval. MINTQA comprises 10,479 question-answer pairs for evaluating new knowledge and 17,887 pairs for assessing long-tail knowledge, with each question equipped with corresponding sub-questions and answers. Our systematic evaluation of 22 state-of-the-art LLMs on MINTQA reveals significant limitations in their ability to handle complex knowledge base queries, particularly in handling new or unpopular knowledge. Our findings highlight critical challenges and offer insights for advancing multi-hop reasoning capabilities. The MINTQA benchmark is available at https://github.com/probe2/multi-hop/.

Multifaceted user modeling aims to uncover fine-grained patterns and learn representations from user data, revealing their diverse interests and characteristics, such as profile, preference, and personality. Recent studies on foundation model-based recommendation have emphasized the Transformer architecture's remarkable ability to capture complex, non-linear user-item interaction relationships. This paper aims to advance foundation model-based recommendersystems by introducing enhancements to multifaceted user modeling capabilities. We propose a novel Transformer layer designed specifically for recommendation, using the self-attention mechanism to capture sequential user-item interaction patterns. Specifically, we design a group gating network to identify user groups, enabling hierarchical discovery across different layers, thereby capturing the multifaceted nature of user interests through multiple Transformer layers. Furthermore, to broaden the data scope and further enhance multifaceted user modeling, we extend the framework to a federated setting, enabling the use of private datasets while ensuring privacy. Experimental validations on benchmark datasets demonstrate the superior performance of our proposed method. Code is available., Comment: Accepted as a regular paper of AAAI25

Flat electronic bands in condensed matter provide a rich avenue for exploring novel quantum phenomena. Here, we report an optical spectroscopy study of a topological hourglass semimetal Nb3SiTe6 with the electric field of the incident light parallel to its crystalline ab-plane. The ab-plane optical conductivity spectra of Nb3SiTe6 single crystals exhibit a remarkable peak-like feature around 1.20 eV, which is mainly contributed by the direct optical transitions between the two ab-initio-calculation-derived flat bands along the momentum direction Z-U. Our results pave the way for investigating exotic quantum phenomena based on the flat bands in topological hourglass semimetals., Comment: Accepted by Chinese Physics B

Lightweight design, as a key approach to mitigate disparity between computational requirements of deep learning models and hardware performance, plays a pivotal role in advancing application of deep learning technologies on mobile and embedded devices, alongside rapid development of smart home, telemedicine, and autonomous driving. With its outstanding feature extracting capabilities, Deep Convolutional Neural Networks (DCNNs) have demonstrated superior performance in computer vision tasks. However, high computational costs and large network architectures severely limit the widespread application of DCNNs on resource-constrained hardware platforms such as smartphones, robots, and IoT devices. This paper reviews lightweight design strategies for DCNNs and examines recent research progress in both lightweight architectural design and model compression. Additionally, this paper discusses current limitations in this field of research and propose prospects for future directions, aiming to provide valuable guidance and reflection for lightweight design philosophy on deep neural networks in the field of computer vision., Comment: 18 pages, 9 figures

The impressive performance of Large Language Model (LLM) has prompted researchers to develop Multi-modal LLM (MLLM), which has shown great potential for various multi-modal tasks. However, current MLLM often struggles to effectively address fine-grained multi-modal challenges. We argue that this limitation is closely linked to the models' visual grounding capabilities. The restricted spatial awareness and perceptual acuity of visual encoders frequently lead to interference from irrelevant background information in images, causing the models to overlook subtle but crucial details. As a result, achieving fine-grained regional visual comprehension becomes difficult. In this paper, we break down multi-modal understanding into two stages, from Coarse to Fine (CoF). In the first stage, we prompt the MLLM to locate the approximate area of the answer. In the second stage, we further enhance the model's focus on relevant areas within the image through visual prompt engineering, adjusting attention weights of pertinent regions. This, in turn, improves both visual grounding and overall performance in downstream tasks. Our experiments show that this approach significantly boosts the performance of baseline models, demonstrating notable generalization and effectiveness. Our CoF approach is available online at https://github.com/Gavin001201/CoF., Comment: 5 pages, Accepted by ICASSP2025, full paper

Transductive few-shot learning has recently triggered wide attention in computer vision. Yet, current methods introduce key hyper-parameters, which control the prediction statistics of the test batches, such as the level of class balance, affecting performances significantly. Such hyper-parameters are empirically grid-searched over validation data, and their configurations may vary substantially with the target dataset and pre-training model, making such empirical searches both sub-optimal and computationally intractable. In this work, we advocate and introduce the unrolling paradigm, also referred to as "learning to optimize", in the context of few-shot learning, thereby learning efficiently and effectively a set of optimized hyper-parameters. Specifically, we unroll a generalization of the ubiquitous Expectation-Maximization (EM) optimizer into a neural network architecture, mapping each of its iterates to a layer and learning a set of key hyper-parameters over validation data. Our unrolling approach covers various statistical feature distributions and pre-training paradigms, including recent foundational vision-language models and standard vision-only classifiers. We report comprehensive experiments, which cover a breadth of fine-grained downstream image classification tasks, showing significant gains brought by the proposed unrolled EM algorithm over iterative variants. The achieved improvements reach up to 10% and 7.5% on vision-only and vision-language benchmarks, respectively.

Universal Multimodal Retrieval (UMR) aims to enable search across various modalities using a unified model, where queries and candidates can consist of pure text, images, or a combination of both. Previous work has attempted to adopt multimodal large language models (MLLMs) to realize UMR using only text data. However, our preliminary experiments demonstrate that more diverse multimodal training data can further unlock the potential of MLLMs. Despite its effectiveness, the existing multimodal training data is highly imbalanced in terms of modality, which motivates us to develop a training data synthesis pipeline and construct a large-scale, high-quality fused-modal training dataset. Based on the synthetic training data, we develop the General Multimodal Embedder (GME), an MLLM-based dense retriever designed for UMR. Furthermore, we construct a comprehensive UMR Benchmark (UMRB) to evaluate the effectiveness of our approach. Experimental results show that our method achieves state-of-the-art performance among existing UMR methods. Last, we provide in-depth analyses of model scaling, training strategies, and perform ablation studies on both the model and synthetic data., Comment: 32 pages, models at https://huggingface.co/Alibaba-NLP/gme-Qwen2-VL-2B-Instruct

The pseudoscalar-pole contributions to hadronic light-by-light scattering are determined by the respective transition form factors (TFFs) into two virtual photons. These TFFs constitute complicated functions of the photon virtualities that, in turn, can be reconstructed in a dispersive approach from their discontinuities. In this work, we present such an analysis for the $\eta^{(\prime)}$ TFFs, implementing a number of constraints from both experiment and theory: normalizations from the $\eta^{(\prime)}\to\gamma\gamma$ decay widths, unitarity constraints from the $\eta^{(\prime)}\to\pi^+\pi^-\gamma$ spectra, chiral symmetry for the $\eta^{(\prime)}\to2(\pi^+\pi^-)$ amplitudes, vector-meson couplings, singly-virtual data from $e^+e^-\to e^+e^-\eta^{(\prime)}$, and the asymptotic behavior predicted by the light-cone expansion. In particular, we account for the leading left-hand-cut singularity by including effects from the $a_2$ resonance, necessitating the solution of an inhomogeneous Muskhelishvili-Omn\`es problem via a carefully chosen path deformation. The resulting TFFs allow us to evaluate the $\eta^{(\prime)}$-pole contributions to the anomalous magnetic moment of the muon, $a_\mu^{\eta\text{-pole}}=14.7(9)\times 10^{-11}$ and $a_\mu^{\eta'\text{-pole}}=13.5(7)\times 10^{-11}$, completing a dedicated program for the lowest-lying pseudoscalar intermediate states in a dispersive approach to hadronic light-by-light scattering, $a_\mu^{\text{PS-poles}}=91.2^{+2.9}_{-2.4}\times 10^{-11}$., Comment: 58 pages, 20 figures, results for the space-like $\eta$ and $\eta'$ transition form factors included as ancillary material

The very high-energy afterglow in GRB 221009A, known as the ``brightest of all time'' (BOAT), has been thoroughly analyzed in previous studies. In this paper, we conducted a statistical analysis of the waiting time behavior of 172 TeV photons from the BOAT observed by LHAASO-KM2A. The following results were obtained: (I) The waiting time distribution (WTD) of these photons deviates from the exponential distribution. (II) The behavior of these photons exhibits characteristics resembling those of a self-organized critical system, such as a power-law distribution and scale-invariance features in the WTD. The power-law distribution of waiting times is consistent with the prediction of a nonstationary process. (III) The relationship between the power-law slopes of the WTD and the scale-invariant characteristics of the Tsallis q-Gaussian distribution deviates from existing theory. We suggest that this deviation is due to the photons not being completely independent of each other. In summary, the power-law and scale-free characteristics observed in these photons imply a self-organized critical process in the generation of teraelectronvolt photons from GRB 221009A. Based on other relevant research, we propose that the involvement of a partially magnetically dominated component and the continuous energy injection from the central engine can lead to deviations in the generation of teraelectronvolt afterglow from the simple external shock-dominated process, thereby exhibiting the self-organized critical characteristics mentioned above., Comment: 5 pages, 3 figures; accepted by Astronomy and Astrophysics

The continuous development of bright x/gamma-ray sources has opened up new frontiers of science and advanced applications. Currently, there is still a lack of efficient approaches to produce gamma-rays with photon energies up to GeV and with high peak brilliance comparable to modern free-electron lasers. Here we report a novel mechanism called beam fast pinching radiation burst to generate such gamma-ray sources. It is achieved by injecting a GeV electron beam into a submillimeter plasma with an upramp density profile, enabling violent beam pinching to occur rapidly. During this process, a burst of collimated gamma-rays is efficiently produced with photon energy up to GeV, energy conversion efficiency exceeding $30\%$, and peak brilliance exceeding $10^{28}$ photons s$^{-1}$ mm$^{-2}$ mrad$^{-2}$ per $0.1\%$ bandwidth. All of these are several orders of magnitude higher than existing gamma-ray sources. This opens a novel avenue for the development of extremely bright gamma-ray sources for both fundamental research and cutting-edge applications.

Industrial control network (ICN) is characterized by real-time responsiveness and reliability, which plays a key role in increasing production speed, rational and efficient processing, and managing the production process. Despite tremendous advantages, ICN inevitably struggles with some challenges, such as malicious user intrusion and hacker attack. To detect malicious intrusions in ICN, intrusion detection systems have been deployed. However, in ICN, network traffic data is equipped with characteristics of large scale, irregularity, multiple features, temporal correlation and high dimensionality, which greatly affect the efficiency and performance. To properly solve the above problems, we design a new intrusion detection method for ICN. Specifically, we first design a novel neural network model called associative recurrent network (ARN), which can properly handle the relationship between past moment hidden state and current moment information. Then, we adopt ARN to design a new intrusion detection method that can efficiently and accurately detect malicious intrusions in ICN. Subsequently, we demonstrate the high efficiency of our proposed method through theoretical computational complexity analysis. Finally, we develop a prototype implementation to evaluate the accuracy. The experimental results prove that our proposed method has sate-of-the-art performance on both the ICN dataset SWaT and the conventional network traffic dataset UNSW-NB15. The accuracies on the SWaT dataset and the UNSW-NB15 dataset reach 95.48% and 97.61%, respectively.

We investigate the correlations between net baryon number and electric charge up to sixth order related to the interactions of nuclear matter at low temperature, and explore their relationship with the nuclear liquid-gas phase transition (LGPT) within the framework of the nonlinear Walecka model. The calculation shows that strong correlations between the baryon number and electric charge exist in the vicinity of LGPT, and the higher order correlations are more sensitive than the lower order ones near the phase transition. However, in the high-temperature region away from the LGPT the rescaled lower order correlations are relatively larger than most of the higher order ones. Besides, some of the fifth- and sixth-order correlations possibly change the sign from negative to positive along the chemical freeze-out line with the decrease of temperature. In combination with the future experimental projects at lower collision energies, the derived results can be referred to study the phase structure of strongly interacting matter and analyze the related experimental signals., Comment: 9 pages, 6 figures

Predicting human mobility is crucial for urban planning, traffic control, and emergency response. Mobility behaviors can be categorized into individual and collective, and these behaviors are recorded by diverse mobility data, such as individual trajectory and crowd flow. As different modalities of mobility data, individual trajectory and crowd flow have a close coupling relationship. Crowd flows originate from the bottom-up aggregation of individual trajectories, while the constraints imposed by crowd flows shape these individual trajectories. Existing mobility prediction methods are limited to single tasks due to modal gaps between individual trajectory and crowd flow. In this work, we aim to unify mobility prediction to break through the limitations of task-specific models. We propose a universal human mobility prediction model (named UniMob), which can be applied to both individual trajectory and crowd flow. UniMob leverages a multi-view mobility tokenizer that transforms both trajectory and flow data into spatiotemporal tokens, facilitating unified sequential modeling through a diffusion transformer architecture. To bridge the gap between the different characteristics of these two data modalities, we implement a novel bidirectional individual and collective alignment mechanism. This mechanism enables learning common spatiotemporal patterns from different mobility data, facilitating mutual enhancement of both trajectory and flow predictions. Extensive experiments on real-world datasets validate the superiority of our model over state-of-the-art baselines in trajectory and flow prediction. Especially in noisy and scarce data scenarios, our model achieves the highest performance improvement of more than 14% and 25% in MAPE and Accuracy@5.

Realizing universal fault-tolerant quantum computation is a key goal in quantum information science. By encoding quantum information into logical qubits utilizing quantum error correcting codes, physical errors can be detected and corrected, enabling substantial reduction in logical error rates. However, the set of logical operations that can be easily implemented on such encoded qubits is often constrained, necessitating the use of special resource states known as 'magic states' to implement universal, classically hard circuits. A key method to prepare high-fidelity magic states is to perform 'distillation', creating them from multiple lower fidelity inputs. Here we present the experimental realization of magic state distillation with logical qubits on a neutral-atom quantum computer. Our approach makes use of a dynamically reconfigurable architecture to encode and perform quantum operations on many logical qubits in parallel. We demonstrate the distillation of magic states encoded in d=3 and d=5 color codes, observing improvements of the logical fidelity of the output magic states compared to the input logical magic states. These experiments demonstrate a key building block of universal fault-tolerant quantum computation, and represent an important step towards large-scale logical quantum processors., Comment: 8+11 pages, 4+4 figures

Automatic Heuristic Design (AHD) is an active research area due to its utility in solving complex search and NP-hard combinatorial optimization problems in the real world. The recent advancements in Large Language Models (LLMs) introduce new possibilities by coupling LLMs with evolutionary computation to automatically generate heuristics, known as LLM-based Evolutionary Program Search (LLM-EPS). While previous LLM-EPS studies obtained great performance on various tasks, there is still a gap in understanding the properties of heuristic search spaces and achieving a balance between exploration and exploitation, which is a critical factor in large heuristic search spaces. In this study, we address this gap by proposing two diversity measurement metrics and perform an analysis on previous LLM-EPS approaches, including FunSearch, EoH, and ReEvo. Results on black-box AHD problems reveal that while EoH demonstrates higher diversity than FunSearch and ReEvo, its objective score is unstable. Conversely, ReEvo's reflection mechanism yields good objective scores but fails to optimize diversity effectively. With this finding in mind, we introduce HSEvo, an adaptive LLM-EPS framework that maintains a balance between diversity and convergence with a harmony search algorithm. Through experimentation, we find that HSEvo achieved high diversity indices and good objective scores while remaining cost-effective. These results underscore the importance of balancing exploration and exploitation and understanding heuristic search spaces in designing frameworks in LLM-EPS., Comment: 18 pages, 12 figures

Large-scale text-to-image diffusion models, (e.g., DALL-E, SDXL) are capable of generating famous persons by simply referring to their names. Is it possible to make such models generate generic identities as simple as the famous ones, e.g., just use a name? In this paper, we explore the existence of a "Name Space", where any point in the space corresponds to a specific identity. Fortunately, we find some clues in the feature space spanned by text embedding of celebrities' names. Specifically, we first extract the embeddings of celebrities' names in the Laion5B dataset with the text encoder of diffusion models. Such embeddings are used as supervision to learn an encoder that can predict the name (actually an embedding) of a given face image. We experimentally find that such name embeddings work well in promising the generated image with good identity consistency. Note that like the names of celebrities, our predicted name embeddings are disentangled from the semantics of text inputs, making the original generation capability of text-to-image models well-preserved. Moreover, by simply plugging such name embeddings, all variants (e.g., from Civitai) derived from the same base model (i.e., SDXL) readily become identity-aware text-to-image models. Project homepage: \url{https://magicfusion.github.io/MagicNaming/}., Comment: Accepted by AAAI 2025

We uncover a phenomenon largely overlooked by the scientific community utilizing AI: neural networks exhibit high susceptibility to minute perturbations, resulting in significant deviations in their outputs. Through an analysis of five diverse application areas -- weather forecasting, chemical energy and force calculations, fluid dynamics, quantum chromodynamics, and wireless communication -- we demonstrate that this vulnerability is a broad and general characteristic of AI systems. This revelation exposes a hidden risk in relying on neural networks for essential scientific computations, calling further studies on their reliability and security., Comment: 26 pages, 6 figures

Controllable artistic image stylization and generation aims to render the content provided by text or image with the learned artistic style, where content and style decoupling is the key to achieve satisfactory results. However, current methods for content and style disentanglement primarily rely on image information for supervision, which leads to two problems: 1) models can only support one modality for style or content input;2) incomplete disentanglement resulting in semantic interference from the reference image. To address the above issues, this paper proposes a content-style representation disentangling method for controllable artistic image stylization and generation. We construct a WikiStyle+ dataset consists of artworks with corresponding textual descriptions for style and content. Based on the multimodal dataset, we propose a disentangled content and style representations guided diffusion model. The disentangled representations are first learned by Q-Formers and then injected into a pre-trained diffusion model using learnable multi-step cross-attention layers for better controllable stylization. This approach allows model to accommodate inputs from different modalities. Experimental results show that our method achieves a thorough disentanglement of content and style in reference images under multimodal supervision, thereby enabling a harmonious integration of content and style in the generated outputs, successfully producing style-consistent and expressive stylized images.

In models of warm dark matter, there is an appreciable population of high momentum particles in the early universe, which free stream out of primordial over/under densities, thereby prohibiting the growth of structure on small length scales. The distance that a dark matter particle travels without obstruction, known as the free streaming length, depends on the particle's mass and momentum, but also on the cosmological expansion rate. In this way, measurements of the linear matter power spectrum serve to probe warm dark matter as well as the cosmological expansion history. In this work, we focus on ultra-light wave wave dark matter (WWDM) characterized by a typical comoving momentum $q_\ast$ and mass $m$. We first derive constraints on the WWDM parameter space $(q_\ast, m)$ using Lyman-$\alpha$ forest observations due to a combination of the free-streaming effect and the white-noise effect. We next assess how the free streaming of WWDM is affected by three modified expansion histories: early matter domination, early dark energy, and very early dark energy., Comment: 17 pages + references, 8 figures

A measurement of the $CP$-violating parameters in $B_s^0 \to D_s^{\mp} K^{\pm}$ decays is reported, based on the analysis of proton-proton collision data collected by the LHCb experiment corresponding to an integrated luminosity of $6\,\mathrm{fb}^{-1}$ at a centre-of-mass energy of $13 \,\mathrm{TeV}$. The measured parameters are $C_f = 0.791 \pm 0.061 \pm 0.022$, $A_f^{\Delta\Gamma} = -0.051 \pm 0.134 \pm 0.058$, $A_{\overline{f}}^{\Delta\Gamma} = -0.303 \pm 0.125 \pm 0.055$, $S_f = -0.571 \pm 0.084 \pm 0.023$ and $S_{\overline{f}} = -0.503 \pm 0.084 \pm 0.025$, where the first uncertainty is statistical and the second systematic. Together with the value of the Bs mixing phase $-2\beta_s$, these parameters are used to obtain a measurement of the CKM angle $\gamma$ equal to $ (74\pm12)^\circ$ modulo $180^{\circ}$, where the uncertainty contains both statistical and systematic contributions. This result is combined with the previous LHCb measurement in this channel using $3\,\mathrm{fb}^{-1}$ resulting in a determination of $\gamma= (81^{+12}_{-11})^\circ$., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/3575/ (LHCb public pages)

In deep-inelastic positron-proton scattering, the lepton-jet azimuthal angular asymmetry is measured using data collected with the H1 detector at HERA. When the average transverse momentum of the lepton-jet system, $\lvert \vec{P}_\perp \rvert $, is much larger than the total transverse momentum of the system, $\lvert \vec{q}_\perp \rvert$, the asymmetry between parallel and antiparallel configurations, $\vec{P}_\perp$ and $\vec{q}_\perp$, is expected to be generated by initial and final state soft gluon radiation and can be predicted using perturbation theory. Quantifying the angular properties of the asymmetry therefore provides an additional test of the strong force. Studying the asymmetry is important for future measurements of intrinsic asymmetries generated by the proton's constituents through Transverse Momentum Dependent (TMD) Parton Distribution Functions (PDFs), where this asymmetry constitutes a dominant background. Moments of the azimuthal asymmetries are measured using a machine learning method for unfolding that does not require binning., Comment: 15 pages, 4 figures, 2 tables, updated affiliations and acknowledgements

A search for $CP$ violation in $\Lambda_b^0\rightarrow pK^-$ and $\Lambda_b^0\rightarrow p\pi^-$ decays is presented using the full Run 1 and Run 2 data samples of $pp$ collisions collected with the LHCb detector, corresponding to an integrated luminosity of 9 $\mathrm{fb}^{-1}$ at center-of-mass energies of 7, 8, and 13 TeV. For the Run 2 data sample, the $CP$-violating asymmetries are measured to be $A_{CP}^{pK^-} = (-1.4 \pm 0.7 \pm 0.4)\%$ and $A_{CP}^{p\pi^-} = (0.4 \pm 0.9 \pm 0.4)\%$, where the first uncertainty is statistical and the second is systematic. Following significant improvements in the evaluation of systematic uncertainties compared to the previous LHCb measurement, the Run 1 dataset is reanalyzed to update the corresponding results. When combining the Run 2 and updated Run 1 measurements, the final results are found to be $A_{CP}^{pK^-} = (-1.1 \pm 0.7 \pm 0.4)\%$ and $A_{CP}^{p\pi^-} = (0.2 \pm 0.8 \pm 0.4)\%$, constituting the most precise measurements of these asymmetries to date., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/3533/ (LHCb public pages)

Medical video generation has transformative potential for enhancing surgical understanding and pathology insights through precise and controllable visual representations. However, current models face limitations in controllability and authenticity. To bridge this gap, we propose SurgSora, a motion-controllable surgical video generation framework that uses a single input frame and user-controllable motion cues. SurgSora consists of three key modules: the Dual Semantic Injector (DSI), which extracts object-relevant RGB and depth features from the input frame and integrates them with segmentation cues to capture detailed spatial features of complex anatomical structures; the Decoupled Flow Mapper (DFM), which fuses optical flow with semantic-RGB-D features at multiple scales to enhance temporal understanding and object spatial dynamics; and the Trajectory Controller (TC), which allows users to specify motion directions and estimates sparse optical flow, guiding the video generation process. The fused features are used as conditions for a frozen Stable Diffusion model to produce realistic, temporally coherent surgical videos. Extensive evaluations demonstrate that SurgSora outperforms state-of-the-art methods in controllability and authenticity, showing its potential to advance surgical video generation for medical education, training, and research.

We present MobiFuse, a high-precision depth perception system on mobile devices that combines dual RGB and Time-of-Flight (ToF) cameras. To achieve this, we leverage physical principles from various environmental factors to propose the Depth Error Indication (DEI) modality, characterizing the depth error of ToF and stereo-matching. Furthermore, we employ a progressive fusion strategy, merging geometric features from ToF and stereo depth maps with depth error features from the DEI modality to create precise depth maps. Additionally, we create a new ToF-Stereo depth dataset, RealToF, to train and validate our model. Our experiments demonstrate that MobiFuse excels over baselines by significantly reducing depth measurement errors by up to 77.7%. It also showcases strong generalization across diverse datasets and proves effectiveness in two downstream tasks: 3D reconstruction and 3D segmentation. The demo video of MobiFuse in real-life scenarios is available at the de-identified YouTube link(https://youtu.be/jy-Sp7T1LVs).

Object detection is a critical task in computer vision, with applications in various domains such as autonomous driving and urban scene monitoring. However, deep learning-based approaches often demand large volumes of annotated data, which are costly and difficult to acquire, particularly in complex and unpredictable real-world environments. This dependency significantly hampers the generalization capability of existing object detection techniques. To address this issue, we introduce a novel single-domain object detection generalization method, named GoDiff, which leverages a pre-trained model to enhance generalization in unseen domains. Central to our approach is the Pseudo Target Data Generation (PTDG) module, which employs a latent diffusion model to generate pseudo-target domain data that preserves source domain characteristics while introducing stylistic variations. By integrating this pseudo data with source domain data, we diversify the training dataset. Furthermore, we introduce a cross-style instance normalization technique to blend style features from different domains generated by the PTDG module, thereby increasing the detector's robustness. Experimental results demonstrate that our method not only enhances the generalization ability of existing detectors but also functions as a plug-and-play enhancement for other single-domain generalization methods, achieving state-of-the-art performance in autonomous driving scenarios.

In this paper, the problem of distributively searching the Stackelberg equilibria of aggregative games with bilevel structures is studied. Different from the traditional aggregative games, here the aggregation is determined by the minimizer of the objective function in the inner level, which depends on players' actions in the outer level. Moreover, the global objective function in the inner level is formed by the sum of some local bifunctions, each of which is strongly convex with respect to the second argument and is only available to a specific player. To handle this problem, first, we propose a second order gradient-based distributed algorithm, where the Hessain matrices associated with the objective functions in the inner level is involved. By the algorithm, players update their actions in the outer level while cooperatively minimizing the sum of the bifunctions in the inner level to estimate the aggregation by communicating with their neighbors via a connected graph. Under mild assumptions on the graph and cost functions, we prove that the actions of players and the estimate on the aggregation asymptotically converge to the Stackelberg equilibrium. Then, for the case where the Hessain matrices associated with the objective functions in the inner level are not available, we propose a first order gradient-based distributed algorithm, where a distributed two-point estimate strategy is developed to estimate the gradients of cost functions in the outer level. Under the same conditions, we prove that the convergence errors of players' actions and the estimate on the aggregation to the Stackelberg equilibrium are linear with respect to the estimate parameters. Finally, simulations are provided to demonstrate the effectiveness of our theoretical results.

Domain generalization aims to learn a model from multiple training domains and generalize it to unseen test domains. Recent theory has shown that seeking the deep models, whose parameters lie in the flat minima of the loss landscape, can significantly reduce the out-of-domain generalization error. However, existing methods often neglect the consistency of loss landscapes in different domains, resulting in models that are not simultaneously in the optimal flat minima in all domains, which limits their generalization ability. To address this issue, this paper proposes an iterative Self-Feedback Training (SFT) framework to seek consistent flat minima that are shared across different domains by progressively refining loss landscapes during training. It alternatively generates a feedback signal by measuring the inconsistency of loss landscapes in different domains and refines these loss landscapes for greater consistency using this feedback signal. Benefiting from the consistency of the flat minima within these refined loss landscapes, our SFT helps achieve better out-of-domain generalization. Extensive experiments on DomainBed demonstrate superior performances of SFT when compared to state-of-the-art sharpness-aware methods and other prevalent DG baselines. On average across five DG benchmarks, SFT surpasses the sharpness-aware minimization by 2.6% with ResNet-50 and 1.5% with ViT-B/16, respectively. The code will be available soon.

Reentrant localization has recently been observed in systems with quasi-periodic nearest-neighbor hopping, where the interplay between dimerized hopping and staggered disorder is identified as the driving mechanism. However, the robustness of reentrant localization in the presence of long-range hopping remains an open question. In this work, we investigate the phenomenon of reentrant localization in systems incorporating long-range hopping. Our results reveal that long-range hopping induces reentrant localization regardless of whether the disorder is staggered or uniform. We demonstrate that long-range hopping does not inherently disrupt localization; instead, under specific conditions, it facilitates the emergence of reentrant localization. Furthermore, by analyzing critical exponents, we show that the inclusion of long-range hopping modifies the critical behavior, leading to transitions that belong to distinct universality classes.

Noisy labels are both inevitable and problematic in machine learning methods, as they negatively impact models' generalization ability by causing overfitting. In the context of learning with noise, the transition matrix plays a crucial role in the design of statistically consistent algorithms. However, the transition matrix is often considered unidentifiable. One strand of methods typically addresses this problem by assuming that the transition matrix is instance-independent; that is, the probability of mislabeling a particular instance is not influenced by its characteristics or attributes. This assumption is clearly invalid in complex real-world scenarios. To better understand the transition relationship and relax this assumption, we propose to study the data generation process of noisy labels from a causal perspective. We discover that an unobservable latent variable can affect either the instance itself, the label annotation procedure, or both, which complicates the identification of the transition matrix. To address various scenarios, we have unified these observations within a new causal graph. In this graph, the input instance is divided into a noise-resistant component and a noise-sensitive component based on whether they are affected by the latent variable. These two components contribute to identifying the ``causal transition matrix'', which approximates the true transition matrix with theoretical guarantee. In line with this, we have designed a novel training framework that explicitly models this causal relationship and, as a result, achieves a more accurate model for inferring the clean label.

The capability to measure the state of a quantum system is vital to a practical quantum network, for applications including distributed quantum computing and long-distance quantum communication. As a thriving platform for quantum information technology, single neutral atoms suffer from low achievable photon scattering rate and shallow trapping potential, which limits the fidelity and speed of state readout process. Here, by coupling an single neutral atom with a high-finesse fiber-based Fabry-P\'erot microcavity (FFPC) in Purcell regime, we realize strong enhancement of the atomic photoemission rate, which enables ultrafast and high-fidelity discrimination of bright and dark hyperfine states of the atom. The readout fidelity can reach 99.1(2)% within 200 ns and 99.985(8)% within 9 $\mu$s. Furthermore, we demonstrate that state preparation via optical pumping can be efficiently accelerated by real-time decision protocol based on ultrafast state readout. This work paves the way to the implementation of quantum networking protocols with high communication rate and high fidelity., Comment: 9 pages, 9 figures, 53 references

Molecular structure elucidation is a fundamental step in understanding chemical phenomena, with applications in identifying molecules in natural products, lab syntheses, forensic samples, and the interstellar medium. We consider the task of predicting a molecule's all-atom 3D structure given only its molecular formula and moments of inertia, motivated by the ability of rotational spectroscopy to measure these moments. While existing generative models can conditionally sample 3D structures with approximately correct moments, this soft conditioning fails to leverage the many digits of precision afforded by experimental rotational spectroscopy. To address this, we first show that the space of $n$-atom point clouds with a fixed set of moments of inertia is embedded in the Stiefel manifold $\mathrm{St}(n, 4)$. We then propose Stiefel Flow Matching as a generative model for elucidating 3D structure under exact moment constraints. Additionally, we learn simpler and shorter flows by finding approximate solutions for equivariant optimal transport on the Stiefel manifold. Empirically, enforcing exact moment constraints allows Stiefel Flow Matching to achieve higher success rates and faster sampling than Euclidean diffusion models, even on high-dimensional manifolds corresponding to large molecules in the GEOM dataset.

The tearing mode, a large-scale MHD instability in tokamak, typically disrupts the equilibrium magnetic surfaces, leads to the formation of magnetic islands, and reduces core electron temperature and density, thus resulting in significant energy losses and may even cause discharge termination. This process is unacceptable for ITER. Therefore, the accurate identification of a magnetic island in real time is crucial for the effective control of the tearing mode in ITER in the future. In this study, based on the characteristics induced by tearing modes, an attention-aware convolutional neural network (AM-CNN) is proposed to identify the presence of magnetic islands in tearing mode discharge utilizing the data from ECE diagnostics in the EAST tokamak. A total of 11 ECE channels covering the range of core is used in the tearing mode dataset, which includes 2.5*10^9 data collected from 68 shots from 2016 to 2021 years. We split the dataset into training, validation, and test sets (66.5%, 5.7%, and 27.8%), respectively. An attention mechanism is designed to couple with the convolutional neural networks to improve the capability of feature extraction of signals. During the model training process, we utilized adaptive learning rate adjustment and early stopping mechanisms to optimize performance of AM-CNN. The model results show that a classification accuracy of 91.96% is achieved in tearing mode identification. Compared to CNN without AM, the attention-aware convolutional neural networks demonstrate great performance across accuracy, recall metrics, and F1 score. By leveraging the deep learning model, which incorporates a physical understanding of the tearing process to identify tearing mode behaviors, the combination of physical mechanisms and deep learning is emphasized, significantly laying an important foundation for the future intelligent control of tearing mode dynamics.

We employ vacuum-to-$B$ meson correlation functions with interpolating currents $\bar{q}'\slashed{n}q$ and $\bar{q}'\slashed{n}\gamma_{\perp}q$ to construct light-cone sum rules (LCSR) for calculating the $B\to K^*$ form factors in the large recoil region. We investigate several subleading-power corrections to these form factors at tree level, including the next-to-leading power contributions from the hard-collinear propagator, the subleading power corrections from the effective current $\bar{q}\Gamma[i\slashed{D}_{\perp}/(2m_b)]h_v$, and four-body higher-twist effects. By utilizing the available leading-power results at $\mathcal{O}(\alpha_s)$ and the power corrections from higher-twist $B$-meson light-cone distribution amplitudes from our previous work, we further derive the improved numerical predictions for $B\to K^*$ form factors by applying the three-parameter model for $B$-meson light-cone distribution amplitudes (LCDAs). The subleading-power contribution is about $30\%$ relative to the corresponding leading-power result. Taking the well-adopted Bourrely-Caprini-Lellouch (BCL) parametrization, we then provide the numerical results of $B\to K^*$ form factors in the entire kinematic range, by adopting the combined fit to both LCSR predictions and lattice QCD results. Taking advantage of the newly obtained $B\to K^*$ form factors, we analyse the rare exclusive decay $B \to K^* \nu_\ell\bar{\nu}_\ell$ and estimate the Standard Model predictions of $\mathcal{BR}(\bar{B}^0 \to \bar{K}^{*0} \nu_\ell\bar{\nu}_\ell)=7.54(86)\times 10^{-6}$, $\mathcal{BR}(\bar{B}^+ \to \bar{K}^{*+} \nu_\ell\bar{\nu}_\ell)=9.35(94)\times 10^{-6}$ and longitudinal $K^*$ polarization fraction $F_L=0.44(4)$., Comment: 33 pages, 9 figures

Event-based semantic segmentation has great potential in autonomous driving and robotics due to the advantages of event cameras, such as high dynamic range, low latency, and low power cost. Unfortunately, current artificial neural network (ANN)-based segmentation methods suffer from high computational demands, the requirements for image frames, and massive energy consumption, limiting their efficiency and application on resource-constrained edge/mobile platforms. To address these problems, we introduce SLTNet, a spike-driven lightweight transformer-based network designed for event-based semantic segmentation. Specifically, SLTNet is built on efficient spike-driven convolution blocks (SCBs) to extract rich semantic features while reducing the model's parameters. Then, to enhance the long-range contextural feature interaction, we propose novel spike-driven transformer blocks (STBs) with binary mask operations. Based on these basic blocks, SLTNet employs a high-efficiency single-branch architecture while maintaining the low energy consumption of the Spiking Neural Network (SNN). Finally, extensive experiments on DDD17 and DSEC-Semantic datasets demonstrate that SLTNet outperforms state-of-the-art (SOTA) SNN-based methods by at least 7.30% and 3.30% mIoU, respectively, with extremely 5.48x lower energy consumption and 1.14x faster inference speed., Comment: Submitted to IEEE ICRA 2025

Semantic segmentation and 3D reconstruction are two fundamental tasks in remote sensing, typically treated as separate or loosely coupled tasks. Despite attempts to integrate them into a unified network, the constraints between the two heterogeneous tasks are not explicitly modeled, since the pioneering studies either utilize a loosely coupled parallel structure or engage in only implicit interactions, failing to capture the inherent connections. In this work, we explore the connections between the two tasks and propose a new network that imposes semantic constraints on the stereo matching task, both implicitly and explicitly. Implicitly, we transform the traditional parallel structure to a new cascade structure termed Semantic-Guided Cascade structure, where the deep features enriched with semantic information are utilized for the computation of initial disparity maps, enhancing semantic guidance. Explicitly, we propose a Semantic Selective Refinement (SSR) module and a Left-Right Semantic Consistency (LRSC) module. The SSR refines the initial disparity map under the guidance of the semantic map. The LRSC ensures semantic consistency between two views via reducing the semantic divergence after transforming the semantic map from one view to the other using the disparity map. Experiments on the US3D and WHU datasets demonstrate that our method achieves state-of-the-art performance for both semantic segmentation and stereo matching., Comment: 9 pages, 6 figures, AAAI 2025

The insufficient generalization of adaptive moment estimation (Adam) has hindered its broader application. Recent studies have shown that flat minima in loss landscapes are highly associated with improved generalization. Inspired by the filtering effect of integration operations on high-frequency signals, we propose multiple integral Adam (MIAdam), a novel optimizer that integrates a multiple integral term into Adam. This multiple integral term effectively filters out sharp minima encountered during optimization, guiding the optimizer towards flatter regions and thereby enhancing generalization capability. We provide a theoretical explanation for the improvement in generalization through the diffusion theory framework and analyze the impact of the multiple integral term on the optimizer's convergence. Experimental results demonstrate that MIAdam not only enhances generalization and robustness against label noise but also maintains the rapid convergence characteristic of Adam, outperforming Adam and its variants in state-of-the-art benchmarks.

The first test of lepton flavour universality between muons and electrons using $B^+ \to K^+\pi^+\pi^-\ell^+\ell^-$ ($\ell=e,\mu$) decays is presented. The measurement is performed with data from proton-proton collisions collected by the LHCb experiment at centre-of-mass energies of 7, 8 and 13 TeV, corresponding to an integrated luminosity of $9\mathrm{fb}^{-1}$. The ratio of branching fractions between $B^+ \to K^+\pi^+\pi^-e^+e^-$ and $B^+ \to K^+\pi^+\pi^-\mu^+\mu^-$decays is measured in the dilepton invariant-mass-squared range $1.1 < q^2 < 7.0~\mathrm{GeV}^2/c^4$ and is found to be $R_{K\pi\pi}^{-1} = 1.31^{+0.18}_{-0.17} \;(\mathrm{stat})\;^{+0.12}_{-0.09} \;(\mathrm{syst})$, in agreement with the Standard Model prediction. The first observation of the $B^+ \to K^+\pi^+\pi^-e^+e^-$ decay is also reported., Comment: All figures and tables, along with machine-readable versions and any supplementary material and additional information, are available at https://lbfence.cern.ch/alcm/public/analysis/full-details/1606/ (LHCb public pages)

Single atoms trapped in optical cavities exhibit immense potential as key nodes in future quantum information processing. They have already demonstrated significant advancement in various quantum technologies, particularly regarding the generation of nonclassical light. Here, we efficiently produce genuine photonic Bell states through the inelastic scattering process of single two-level intracavity atoms. An experimental violation of the Bell inequality, arising from the interference between the probability amplitudes of two photons, validates the intrinsic nature of energy-time entanglement. Coupling atoms with an optical cavity in the Purcell regime substantially enhances the two-photon scattering. This Bell state generation process does not require atomic spin control, thereby rendering it inherently immune to decoherence effects. This work advances the comprehension of resonance fluorescence and has the potential to broaden the landscape of quantum technologies and facilitate the application of photonic Bell states., Comment: 9 pages, 8 figures, 47 reference

Bipartite graphs are commonly used to model relationships between two distinct entities in real-world applications, such as user-product interactions, user-movie ratings and collaborations between authors and publications. A butterfly (a 2x2 bi-clique) is a critical substructure in bipartite graphs, playing a significant role in tasks like community detection, fraud detection, and link prediction. As more real-world data is presented in a streaming format, efficiently counting butterflies in streaming bipartite graphs has become increasingly important. However, most existing algorithms typically assume that duplicate edges are absent, which is hard to hold in real-world graph streams, as a result, they tend to sample edges that appear multiple times, leading to inaccurate results. The only algorithm designed to handle duplicate edges is FABLE, but it suffers from significant limitations, including high variance, substantial time complexity, and memory inefficiency due to its reliance on a priority queue. To overcome these limitations, we introduce DEABC (Duplicate-Edge-Aware Butterfly Counting), an innovative method that uses bucket-based priority sampling to accurately estimate the number of butterflies, accounting for duplicate edges. Compared to existing methods, DEABC significantly reduces memory usage by storing only the essential sampled edge data while maintaining high accuracy. We provide rigorous proofs of the unbiasedness and variance bounds for DEABC, ensuring they achieve high accuracy. We compare DEABC with state-of-the-art algorithms on real-world streaming bipartite graphs. The results show that our DEABC outperforms existing methods in memory efficiency and accuracy, while also achieving significantly higher throughput.

Unsupervised visible-infrared person re-identification (USL-VI-ReID) is of great research and practical significance yet remains challenging due to the absence of annotations. Existing approaches aim to learn modality-invariant representations in an unsupervised setting. However, these methods often encounter label noise within and across modalities due to suboptimal clustering results and considerable modality discrepancies, which impedes effective training. To address these challenges, we propose a straightforward yet effective solution for USL-VI-ReID by mitigating universal label noise using neighbor information. Specifically, we introduce the Neighbor-guided Universal Label Calibration (N-ULC) module, which replaces explicit hard pseudo labels in both homogeneous and heterogeneous spaces with soft labels derived from neighboring samples to reduce label noise. Additionally, we present the Neighbor-guided Dynamic Weighting (N-DW) module to enhance training stability by minimizing the influence of unreliable samples. Extensive experiments on the RegDB and SYSU-MM01 datasets demonstrate that our method outperforms existing USL-VI-ReID approaches, despite its simplicity. The source code is available at: https://github.com/tengxiao14/Neighbor-guided-USL-VI-ReID.

The residuals in factor models prevalent in asset pricing presents opportunities to exploit the mis-pricing from unexplained cross-sectional variation for arbitrage. We performed a replication of the methodology of Guijarro-Ordonez et al. (2019) (G-P-Z) on Deep Learning Statistical Arbitrage (DLSA), originally applied to U.S. equity data from 1998 to 2016, using a more recent out-of-sample period from 2016 to 2024. Adhering strictly to point-in-time (PIT) principles and ensuring no information leakage, we follow the same data pre-processing, factor modeling, and deep learning architectures (CNNs and Transformers) as outlined by G-P-Z. Our replication yields unusually strong performance metrics in certain tests, with out-of-sample Sharpe ratios occasionally exceeding 10. While such results are intriguing, they may indicate model overfitting, highly specific market conditions, or insufficient accounting for transaction costs and market impact. Further examination and robustness checks are needed to align these findings with the more modest improvements reported in the original study. (This work was conducted as the final project for IEOR 4576: Data-Driven Methods in Finance at Columbia University.), Comment: Investment universe selection error discovered in the replication process, should've used dynamic S&P 500 pool instead of a static S&P 500 list

Metal-organic frameworks (MOFs) are porous, crystalline materials with high surface area, adjustable porosity, and structural tunability, making them ideal for diverse applications. However, traditional experimental and computational methods have limited scalability and interpretability, hindering effective exploration of MOF structure-property relationships. To address these challenges, we introduce, for the first time, a category-specific topological learning (CSTL), which combines algebraic topology with chemical insights for robust property prediction. The model represents MOF structures as simplicial complexes and incorporates elemental categorizations to enable balanced, interpretable machine learning study. By integrating category-specific persistent homology, CSTL captures both global and local structural characteristics, rendering multi-dimensional, category-specific descriptors that support a predictive model with high accuracy and robustness across eight MOF datasets, outperforming all previous results. This alignment of topological and chemical features enhances the predictive power and interpretability of CSTL, advancing understanding of structure-property relationships of MOFs and promoting efficient material discovery., Comment: 29 pages, 5 figures in main text, and 10 figures in appendiex

Magnon blockade is a fundamental quantum phenomenon for generating single-magnon state, which gradually becomes one of the candidates for quantum information processing. In this paper, we propose a theoretical scheme to generate the magnon blockade in a hybrid system consisting of a YIG micromagnet and a skyrmion. Considering weak probing of the magnon and driving of the skyrmion qubit, the second-order correlation function is analytically derived, and the optimal condition for realizing the magnon blockade is identified. Under the optimal condition, we systematically analyze the behavior of the second-order correlation function $g^{(2)}(0)$ under different parameter regimes. Our analysis shows that with appropriate driving and probing field intensities, the magnon blockade effect can be significantly enhanced, effectively suppressing multi-magnon states and facilitating the generation of high-purity single-magnon states exhibiting pronounced antibunching. Furthermore, we explore the physical mechanisms underlying the magnon blockade, revealing the coexistence and interplay of conventional and unconventional magnon blockade. This scheme provides a versatile all-magnetic platform for generating high purity single-magnon sources.

Device-to-device (D2D)-assisted mobile edge computing (MEC) is one of the critical technologies of future sixth generation (6G) networks. The core of D2D-assisted MEC is to reduce system latency for network edge UEs by supporting cloud computing services, thereby achieving high-speed transmission. Due to the sensitivity of communication signals to obstacles, relaying is adopted to enhance the D2D-assisted MEC system's performance and its coverage area. However, relay nodes and the base station (BS) are typically equipped with large-scale antenna arrays. This increases the cost of relay-assisted D2D MEC systems and limits their deployment. Movable antenna (MA) technology is used to work around this limitation without compromising performance. Specifically, the core of MA technology lies in optimizing the antenna positions to increase system capacity. Therefore, this paper proposes a novel resource allocation scheme for MA-enhanced relay-assisted D2D MEC systems. Specifically, the MA positions and beamforming of user equipments (UEs), relay, and BS as well as the allocation of resources and the computation task offloading rate at the MEC server, all are optimized herein with the objective of minimizing the maximum latency while satisfying computation and communication rate constraints. Since this is a multivariable non-convex problem, a parallel and distributed penalty dual decomposition (PDD) based algorithm is developed and combined with successive convex approximation (SCA) to solve this non-convex problem. The results of extensive numerical analyses show that the proposed algorithm significantly improves the performance of the MA-enhanced relay-assisted D2D communication system compared to a counterpart where relays and the BS are equiped with traditional fixed-position antenna (FPA).

Confinement is a well-known phenomenon in the infrared regime of (supersymmetric) Yang-Mills theory. Although experiments and numerical simulations have solidly confirmed confinement, its physical origin remains mysterious today, and finding a theoretical explanation for it is a long-standing and challenging problem in physics and mathematics. Inspired by the recent progress in quantum Jackiw-Teitelboim gravity, we compute the Wilson loop correlators of the large-$N$ limit of $\mathscr{N}=4$ super-Yang-Mills theory holographically in an extremal AdS$_5$ Reissner-Nordstr\"om black brane background. The quantum gravity fluctuations of the near-horizon region are considered, which consequently affect the holographic Wilson loop correlators. Within this framework, the results suggest that the confinement of the super-Yang-Mills theory is induced by the near-horizon quantum gravity fluctuations of the bulk extremal AdS$_5$ black brane., Comment: 5 pages + 3 appendices, 4 figures

The rapid advancement of deep learning has intensified the need for comprehensive data for use by autonomous driving algorithms. High-quality datasets are crucial for the development of effective data-driven autonomous driving solutions. Next-generation autonomous driving datasets must be multimodal, incorporating data from advanced sensors that feature extensive data coverage, detailed annotations, and diverse scene representation. To address this need, we present OmniHD-Scenes, a large-scale multimodal dataset that provides comprehensive omnidirectional high-definition data. The OmniHD-Scenes dataset combines data from 128-beam LiDAR, six cameras, and six 4D imaging radar systems to achieve full environmental perception. The dataset comprises 1501 clips, each approximately 30-s long, totaling more than 450K synchronized frames and more than 5.85 million synchronized sensor data points. We also propose a novel 4D annotation pipeline. To date, we have annotated 200 clips with more than 514K precise 3D bounding boxes. These clips also include semantic segmentation annotations for static scene elements. Additionally, we introduce a novel automated pipeline for generation of the dense occupancy ground truth, which effectively leverages information from non-key frames. Alongside the proposed dataset, we establish comprehensive evaluation metrics, baseline models, and benchmarks for 3D detection and semantic occupancy prediction. These benchmarks utilize surround-view cameras and 4D imaging radar to explore cost-effective sensor solutions for autonomous driving applications. Extensive experiments demonstrate the effectiveness of our low-cost sensor configuration and its robustness under adverse conditions. Data will be released at https://www.2077ai.com/OmniHD-Scenes.

The emergence of Software-Defined Vehicles (SDVs) signifies a shift from a distributed network of electronic control units (ECUs) to a centralized computing architecture within the vehicle's electrical and electronic systems. This transition addresses the growing complexity and demand for enhanced functionality in traditional E/E architectures, with containerization and virtualization streamlining software development and updates within the SDV framework. While widely used in cloud computing, their performance and suitability for intelligent vehicles have yet to be thoroughly evaluated. In this work, we conduct a comprehensive performance evaluation of containerization and virtualization on embedded and high-performance AMD64 and ARM64 systems, focusing on CPU, memory, network, and disk metrics. In addition, we assess their impact on real-world automotive applications using the Autoware framework and further integrate a microservice-based architecture to evaluate its start-up time and resource consumption. Our extensive experiments reveal a slight 0-5% performance decline in CPU, memory, and network usage for both containerization and virtualization compared to bare-metal setups, with more significant reductions in disk operations-5-15% for containerized environments and up to 35% for virtualized setups. Despite these declines, experiments with actual vehicle applications demonstrate minimal impact on the Autoware framework, and in some cases, a microservice architecture integration improves start-up time by up to 18%., Comment: 15 pages, 15 figures

Anisotropic stochastic gravitational wave background (SGWB) serves as a potential probe of the large-scale structure (LSS) of the universe. In this work, we explore the anisotropic SGWB from local ($z < \sim 0.085$) merging stellar mass compact binaries, specifically focusing on merging stellar binary black holes, merging neutron-star-black-hole binaries, and merging binary neutron stars. The analysis employs seven all-sky mock lightcone gravitational wave event catalogues, which are derived from the Millennium simulation combined with a semi-analytic model of galaxy formation and a binary population synthesis model. We calculate the angular power spectra $\mathrm{C}_\ell$ at multipole moments $\ell$, expressed as $\text{log}_{10} [\ell(\ell+1)\mathrm{C}_\ell/(2\pi)]$, based on the skymaps of the overdensity $\delta_\mathrm{GW}$ in the anisotropic SGWB. The spectra for all three source types exhibit an approximately linear increase with $\text{log}_{10} \ell$ at higher $\ell$ (e.g., $\ell > \sim 30 - 300$) in seven catalogues, with a characteristic slope of $\sim 2$. The spectra of seven catalogues exhibit considerable variations, arising from fluctuations in spatial distribution, primarily in the radial distribution, of nearby sources (e.g., $< 50$ Mpc/h). After subtracting these nearby sources, the variations become much smaller and the spectra for the three source types become closely aligned (within discrepancies of a factor of $\sim 2$ across $\ell = 1 - 1000$ for all catalogues)., Comment: 11 pages 5 figures

Cognitive impairment detection through spontaneous speech offers potential for early diagnosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI). The PROCESS Grand Challenge, part of ICASSP 2025, focuses on advancing this field with innovative solutions for classification and regression tasks. In this work, we integrate interpretable features with temporal features extracted from pre-trained models through a multimodal fusion strategy. For the classification task, our model achieved an F1-score of 0.649 in predicting cognitive states (healthy, MCI, dementia). For the regression task, which involves MMSE score prediction, we obtained a root-mean-square error (RMSE) of 2.628. These results led to our team securing the top overall ranking in the competition., Comment: ICASSP 2025

Deep hash-based retrieval techniques are widely used in facial retrieval systems to improve the efficiency of facial matching. However, it also carries the danger of exposing private information. Deep hash models are easily influenced by adversarial examples, which can be leveraged to protect private images from malicious retrieval. The existing adversarial example methods against deep hash models focus on universality and transferability, lacking the research on its robustness in online social networks (OSNs), which leads to their failure in anti-retrieval after post-processing. Therefore, we provide the first in-depth discussion on robustness adversarial perturbation in universal transferable anti-facial retrieval and propose Three-in-One Adversarial Perturbation (TOAP). Specifically, we construct a local and global Compression Generator (CG) to simulate complex post-processing scenarios, which can be used to mitigate perturbation. Then, we propose robust optimization objectives based on the discovery of the variation patterns of model's distribution after post-processing, and generate adversarial examples using these objectives and meta-learning. Finally, we iteratively optimize perturbation by alternately generating adversarial examples and fine-tuning the CG, balancing the performance of perturbation while enhancing CG's ability to mitigate them. Numerous experiments demonstrate that, in addition to its advantages in universality and transferability, TOAP significantly outperforms current state-of-the-art methods in multiple robustness metrics. It further improves universality and transferability by 5% to 28%, and achieves up to about 33% significant improvement in several simulated post-processing scenarios as well as mainstream OSNs, demonstrating that TOAP can effectively protect private images from malicious retrieval in real-world scenarios.