Search

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 25 pages, 6 figures

IRAS source 19312+1950 (hereafter I19312) is an infrared point source with maser emissions of SiO, H$_2$O, and OH molecules. Although initial observations suggested that I19312 might be an evolved star, its characteristics are not fully consistent with this classification. This study aims to further investigate the nature of I19312 by conducting long-term monitoring of its maser emissions and comparing the results with other known astrophysical objects. We conducted long-term monitoring of SiO, H$_2$O, and OH maser emissions using single-dish radio telescopes. The results were then compared with historical maser data and the characteristics of similar objects to infer the possible origin of I19312. The SiO maser emissions from I19312 were detected over a wide velocity range and exhibited significant time variability. The OH maser lines suggest characteristics of an evolved star, while the H$_2$O maser lines indicate molecular outflows. These features suggest that I19312 could be a candidate for a Water Fountain (WF) star, though there are inconsistencies, such as the large molecular gas mass, that challenge this hypothesis. The possibility of I19312 being a Red Nova Remnant (RNR) is also considered, but this remains speculative due to the lack of direct evidence. The evolutionary stage of I19312 remains unclear, but it shares multiple characteristics with both evolved stars with peculiar properties and RNRs. Further long-term monitoring and high-resolution interferometric observations are required to better constrain the nature of this object., Comment: 36 pages, 17 figures, Accepted for publication in ApJ

Fusing small resource states into a larger, fully connected graph-state is essential for scalable photonic quantum computing. Theoretical analysis reveals that this can only be achieved when the success probability of the fusion gate surpasses a specific percolation threshold of 58.98% by using three-photon GHZ states as resource states. However, such an implementation of a fusion gate has never been experimentally realized before. Here, we successfully demonstrate a boosted fusion gate with a theoretical success probability of 75%, using deterministically generated auxiliary states. The success probability is experimentally measured to be 71.0(7)%. We further demonstrate the effectiveness of the boosted fusion gate by fusing two Bell states with a fidelity of 67(2)%. Our work paves a crucial path toward scalable linear optical quantum computing., Comment: 5 pages, 4 figures

This article critically examines the methodologies applied in ancient DNA (aDNA) research, particularly those developed by Dr. P\"a\"abo's team, which have significantly influenced the field. The focus is on the challenges of distinguishing original in situ DNA (oriDNA) from environmental DNA (eDNA) contamination in fossil samples. Recent analyses indicate that even with rigorous extraction and sequencing protocols, a considerable amount of eDNA remains present, often misinterpreted as oriDNA. This misidentification risks the accuracy of species ascription and evolutionary interpretations derived from fossil analyses. The paper explores fossil preservation's physical and chemical dynamics, which allow eDNA from similar and disparate species to infiltrate bone matrices. We propose enhancements to methodological frameworks, such as broader BLAST database usage and stringent E-value criteria, to improve species-specific aDNA identification. Additionally, the article critiques the reliance on deamination patterns as a definitive marker for aDNA, suggesting a reevaluation of this criterion due to its inconsistency and the potential for misleading sequencing results. Ultimately, our findings advocate for a more cautious and refined approach to aDNA research, ensuring more reliable and verifiable scientific outcomes, Comment: 11 pages, 2 Figures

Orbital angular momentum (OAM) can enhance the spectral efficiency by multiplying a set of orthogonal modes on the same frequency channel. To maintain the orthogonal among different OAM modes, perfect alignments between transmitters and receivers are strictly required. However, in multi-user OAM communications, the perfect alignments between the transmitter and all the receivers are impossible. The phase turbulence, caused by misaligned transmitters and receivers, leads to serious inter-mode interference, which greatly degrades the capacity of OAM transmissions. To eliminate the negative effects of phase turbulence and further enhance the transmission capacity, we introduce RIS into the system, and propose a joint precoder and reflector design for reconfigurable intelligent surface (RIS)-assisted multi-user OAM communication systems. Specifically, we propose a three-layer design at the transmitter side, which includes inter-user OAM mode interference cancellation, inter-mode self-interference elimination and the power allocation among different users. By analyzing the characteristics of the overall channels, we are able to give the specific expressions of the precoder designs, which significantly reduce the optimization complexity. We further leverage RIS to guarantee the line-ofsight (LoS) transmissions between the transmitter and users for better sum rate performance. To verify the superiority of the proposed multi-user OAM transmission system, we compare it with traditional MIMO transmission schemes, numerical results have shown that our proposed design can achieve better sum rate performance due to the well-designed orthogonality among different users and OAM modes.

Intensity correlations between neighboring pulses open a prevalent yet often overlooked security loophole in decoy-state quantum key distribution (QKD). As a solution, we present and experimentally demonstrate an intensity-correlation-tolerant QKD protocol that mitigates the negative effect that this phenomenon has on the secret key rate according to existing security analyses. Compared to previous approaches, our method significantly enhances the robustness against correlations, notably improving both the maximum transmission distances and the achievable secret key rates across different scenarios. By relaxing constraints on correlation parameters, our protocol enables practical devices to counter intensity correlations. We experimentally demonstrate this first practical solution that directly overcomes this security vulnerability, establish the feasibility and efficacy of our proposal, taking a major step towards loophole-free and high-performance QKD.

Quantum illumination is a protocol for detecting a low-reflectivity target by using two-mode entangled states composed of signal and idler modes. In this study, we extend the two-mode qubit states to three-mode qubit states, exploring the following configurations: (i) three signals, (ii) two signals and one idler, and (iii) one signal and two idlers. Each configuration considers various three-qubit states, such as GHZ state, W state, bipartite entangled states, and product states. We derive single-shot detection limits of the three-qubit states under white noise environment, evaluating the detection error probabilities for each configuration. We obtain that the performance is enhanced by the entanglement between signal and idler qubits, whereas it is degraded by the entanglement between signal qubits. In particular, the optimal probe state is not a maximally entangled state but a bipartite entangled state. Moreover, we show that the order of detection error probabilities is consistently certified by the quantum mutual information for the three-qubit probe states. Furthermore, we devote attention to whether this tendency maintains even in multipartite qudit states., Comment: 28 pages, 15 figures, 9 tables

This paper presents \textbf{FreEformer}, a simple yet effective model that leverages a \textbf{Fre}quency \textbf{E}nhanced Trans\textbf{former} for multivariate time series forecasting. Our work is based on the assumption that the frequency spectrum provides a global perspective on the composition of series across various frequencies and is highly suitable for robust representation learning. Specifically, we first convert time series into the complex frequency domain using the Discrete Fourier Transform (DFT). The Transformer architecture is then applied to the frequency spectra to capture cross-variate dependencies, with the real and imaginary parts processed independently. However, we observe that the vanilla attention matrix exhibits a low-rank characteristic, thus limiting representation diversity. This could be attributed to the inherent sparsity of the frequency domain and the strong-value-focused nature of Softmax in vanilla attention. To address this, we enhance the vanilla attention mechanism by introducing an additional learnable matrix to the original attention matrix, followed by row-wise L1 normalization. Theoretical analysis~demonstrates that this enhanced attention mechanism improves both feature diversity and gradient flow. Extensive experiments demonstrate that FreEformer consistently outperforms state-of-the-art models on eighteen real-world benchmarks covering electricity, traffic, weather, healthcare and finance. Notably, the enhanced attention mechanism also consistently improves the performance of state-of-the-art Transformer-based forecasters.

Recent advancements in multi-view action recognition have largely relied on Transformer-based models. While effective and adaptable, these models often require substantial computational resources, especially in scenarios with multiple views and multiple temporal sequences. Addressing this limitation, this paper introduces the MV-GMN model, a state-space model specifically designed to efficiently aggregate multi-modal data (RGB and skeleton), multi-view perspectives, and multi-temporal information for action recognition with reduced computational complexity. The MV-GMN model employs an innovative Multi-View Graph Mamba network comprising a series of MV-GMN blocks. Each block includes a proposed Bidirectional State Space Block and a GCN module. The Bidirectional State Space Block introduces four scanning strategies, including view-prioritized and time-prioritized approaches. The GCN module leverages rule-based and KNN-based methods to construct the graph network, effectively integrating features from different viewpoints and temporal instances. Demonstrating its efficacy, MV-GMN outperforms the state-of-the-arts on several datasets, achieving notable accuracies of 97.3\% and 96.7\% on the NTU RGB+D 120 dataset in cross-subject and cross-view scenarios, respectively. MV-GMN also surpasses Transformer-based baselines while requiring only linear inference complexity, underscoring the model's ability to reduce computational load and enhance the scalability and applicability of multi-view action recognition technologies.

Accurate prediction of mobile traffic, \textit{i.e.,} network traffic from cellular base stations, is crucial for optimizing network performance and supporting urban development. However, the non-stationary nature of mobile traffic, driven by human activity and environmental changes, leads to both regular patterns and abrupt variations. Diffusion models excel in capturing such complex temporal dynamics due to their ability to capture the inherent uncertainties. Most existing approaches prioritize designing novel denoising networks but often neglect the critical role of noise itself, potentially leading to sub-optimal performance. In this paper, we introduce a novel perspective by emphasizing the role of noise in the denoising process. Our analysis reveals that noise fundamentally shapes mobile traffic predictions, exhibiting distinct and consistent patterns. We propose NPDiff, a framework that decomposes noise into \textit{prior} and \textit{residual} components, with the \textit{prior} derived from data dynamics, enhancing the model's ability to capture both regular and abrupt variations. NPDiff can seamlessly integrate with various diffusion-based prediction models, delivering predictions that are effective, efficient, and robust. Extensive experiments demonstrate that it achieves superior performance with an improvement over 30\%, offering a new perspective on leveraging diffusion models in this domain.

Process reward model (PRM) is critical for mathematical reasoning tasks to assign rewards for each intermediate steps. The PRM requires constructing process-wise supervision data for training, which rely on chain-of-thought (CoT) or tree-based methods to construct the reasoning steps, however, the individual reasoning steps may be redundant or containing nuanced errors that difficult to detect. We attribute these to the issue of the overlook of granularity division during process data collection. In this paper, we propose a coarse-to-fine framework to tackle this issue. Specifically, while gathering the process supervision data, we collect the coarse reasoning steps by merging adjacent steps according to preset merging granularity, then we sequentially reduce the merging granularity to collect fine-grained reasoning steps. For each synthesized new step, we relabel according to the label of last step. During training, we also traverse the collected training corpus in a coarse-to-fine manner. We conduct extensive experiments on popular mathematical reasoning datasets across diverse loss criterions, the proposed framework can consistently boost the reasoning performance.

The Wi-Fi technology (IEEE 802.11) was introduced in 1997. With the increasing use and deployment of such networks, their security has also attracted considerable attention. Current Wi-Fi networks use WPA2 (Wi-Fi Protected Access 2) for security (authentication and encryption) between access points and clients. According to the IEEE 802.11i-2004 standard, wireless networks secured with WPA2-PSK (Pre-Shared Key) are required to be protected with a passphrase between 8 to 63 ASCII characters. However, a poorly chosen passphrase significantly reduces the effectiveness of both WPA2 and WPA3-Personal Transition Mode. The objective of this paper is to empirically evaluate password choices in the wild and evaluate weakness in current common practices. We collected a total of 3,352 password hashes from Wi-Fi access points and determine the passphrases that were protecting them. We then analyze these passwords to investigate the impact of user's behavior and preference for convenience on passphrase strength in secured private Wi-Fi networks in Singapore. We characterized the predictability of passphrases that use the minimum required length of 8 numeric or alphanumeric characters, and/or symbols stipulated in wireless security standards, and the usage of default passwords, and found that 16 percent of the passwords show such behavior. Our results also indicate the prevalence of the use of default passwords by hardware manufacturers. We correlate our results with our findings and recommend methods that will improve the overall security and future of our Wi-Fi networks.

Assuming construction of Prandtl and its high order corrections for compressible fluids with small Mach number, we establish nonlinear remainder bound and justify the validity of the Prandtl layer expansion in the inviscid limit for an insulated boundary. To solve the full steady compressible Navier-Stokes system, which consists a div-curl reduction of \textit{the momentum equations} to an elliptic system for the stream function $\phi $ as well as an transport equation for the pressure $p$. The contruction of $\phi$ is based on an improved Guo-Iyer's $H^{4}$ theory for incompressible flows via development of a $H_{00}^{1/2}$ technique with a type \textit{viscous-inflow} BC, and the pressure $p$ is solved by imposing a key additional \textit{viscous-inflow} boundary condition as \begin{equation} \big\{(\mu +\lambda )u_{s}\Delta_{\varepsilon} p + 2(\frac{\mu}{\lambda}+1)p_s (1-\frac{1}{2}u_s^2)p_{x}\big\}\big|_{x=0}\backsim \text{given.} \end{equation} We reformulate the \textit{energy equation} in terms of the \textit{pseudo entropy} $S$ (not the temperature), which exhibits a crucial uniform bound for the final closure., Comment: All comments are welcome

Trajectory data play a crucial role in many applications, ranging from network optimization to urban planning. Existing studies on trajectory data are task-specific, and their applicability is limited to the specific tasks on which they have been trained, such as generation, recovery, or prediction. However, the potential of a unified model has not yet been fully explored in trajectory modeling. Although various trajectory tasks differ in inputs, outputs, objectives, and conditions, they share common mobility patterns. Based on these common patterns, we can construct a general framework that enables a single model to address different tasks. However, building a trajectory task-general framework faces two critical challenges: 1) the diversity in the formats of different tasks and 2) the complexity of the conditions imposed on different tasks. In this work, we propose a general trajectory modeling framework via masked conditional diffusion (named GenMove). Specifically, we utilize mask conditions to unify diverse formats. To adapt to complex conditions associated with different tasks, we utilize historical trajectory data to obtain contextual trajectory embeddings, which include rich contexts such as spatiotemporal characteristics and user preferences. Integrating the contextual trajectory embedding into diffusion models through a classifier-free guidance approach allows the model to flexibly adjust its outputs based on different conditions. Extensive experiments on mainstream tasks demonstrate that our model significantly outperforms state-of-the-art baselines, with the highest performance improvement exceeding 13% in generation tasks.

In this paper, we address the lifelong sequential behavior incomprehension problem in large language models (LLMs) for recommendation, where LLMs struggle to extract useful information from long user behavior sequences, even within their context limits. To tackle this, we propose ReLLaX (Retrieval-enhanced Large Language models Plus), a framework offering optimization across data, prompt, and parameter levels. At the data level, we introduce Semantic User Behavior Retrieval (SUBR) to reduce sequence heterogeneity, making it easier for LLMs to extract key information. For prompt-level enhancement, we employ Soft Prompt Augmentation (SPA) to inject collaborative knowledge, aligning item representations with recommendation tasks and improving LLMs's exploration of item relationships. Finally, at the parameter level, we propose Component Fully-interactive LoRA (CFLoRA), which enhances LoRA's expressiveness by enabling interactions between its components, allowing better capture of sequential information. Moreover, we present new perspectives to compare current LoRA-based LLM4Rec methods, i.e. from both a composite and a decomposed view. We theoretically demonstrate that the ways they employ LoRA for recommendation are degraded versions of our CFLoRA, with different constraints on atom component interactions. Extensive experiments on three public datasets demonstrate ReLLaX's superiority over existing baselines and its ability to mitigate lifelong sequential behavior incomprehension effectively., Comment: Under Review

Federated domain generalization aims to train a global model from multiple source domains and ensure its generalization ability to unseen target domains. {Due to the target domain being with unknown domain shifts, attempting to approximate these gaps by source domains may be the key to improving model generalization capability.} Existing works mainly focus on sharing and recombining local domain-specific attributes to increase data diversity and simulate potential domain shifts. {However, these methods may be insufficient since only the local attribute recombination can be hard to touch the out-of-distribution of global data.} In this paper, we propose a simple-yet-efficient framework named Federated Domain Adversarial Generation (FedDAG). {It aims to simulate the domain shift and improve the model generalization by adversarially generating novel domains different from local and global source domains.} Specifically, it generates novel-style images by maximizing the instance-level feature discrepancy between original and generated images and trains a generalizable task model by minimizing their feature discrepancy. {Further, we observed that FedDAG could cause different performance improvements for local models. It may be due to inherent data isolation and heterogeneity among clients, exacerbating the imbalance in their generalization contributions to the global model.} {Ignoring this imbalance can lead the global model's generalization ability to be sub-optimal, further limiting the novel domain generation procedure. } Thus, to mitigate this imbalance, FedDAG hierarchically aggregates local models at the within-client and across-client levels by using the sharpness concept to evaluate client model generalization contributions. {Extensive experiments across four medical benchmarks demonstrate FedDAG's ability to enhance generalization in federated medical scenarios.}

In commonly used sub-quadratic complexity modules, linear attention benefits from simplicity and high parallelism, making it promising for image synthesis tasks. However, the architectural design and learning strategy for linear attention remain underexplored in this field. In this paper, we offer a suite of ready-to-use solutions for efficient linear diffusion Transformers. Our core contributions include: (1) Simplified Linear Attention using few heads, observing the free-lunch effect of performance without latency increase. (2) Weight inheritance from a fully pre-trained diffusion Transformer: initializing linear Transformer using pre-trained diffusion Transformer and loading all parameters except for those related to linear attention. (3) Hybrid knowledge distillation objective: using a pre-trained diffusion Transformer to help the training of the student linear Transformer, supervising not only the predicted noise but also the variance of the reverse diffusion process. These guidelines lead to our proposed Linear Diffusion Transformer (LiT), an efficient text-to-image Transformer that can be deployed offline on a laptop. Experiments show that in class-conditional 256*256 and 512*512 ImageNet benchmark LiT achieves highly competitive FID while reducing training steps by 80% and 77% compared to DiT. LiT also rivals methods based on Mamba or Gated Linear Attention. Besides, for text-to-image generation, LiT allows for the rapid synthesis of up to 1K resolution photorealistic images. Project page: https://techmonsterwang.github.io/LiT/., Comment: 21 pages, 12 figures

Zero-shot visual question answering (ZS-VQA), an emerged critical research area, intends to answer visual questions without providing training samples. Existing research in ZS-VQA has proposed to leverage knowledge graphs or large language models (LLMs), respectively, as external information sources to help VQA model comprehend images and questions. However, LLMs often struggle in accurately interpreting specific question meanings. Meanwhile, although knowledge graph has rich entity relationships, it is challenging to effectively connect entities to individual image content for visual question answers. In this paper, we propose a novel design to combine knowledge graph and LLMs for zero-shot visual question answer. Our approach uses LLMs' powerful understanding capabilities to accurately interpret image content through a strategic question search mechanism. Meanwhile, the knowledge graph is used to expand and connect users' queries to the image content for better visual question answering. An optimization algorithm is further used to determine the optimal weights for the loss functions derived from different information sources, towards a globally optimal set of candidate answers. Experimental results on two benchmark datasets demonstrate that our model achieves state-of-the-art (SOTA) performance. Both source code and benchmark data will be released for public access.

The integration of density functional theory (DFT) with machine learning enables efficient \textit{ab initio} electronic structure calculations for ultra-large systems. In this work, we develop a transfer learning framework tailored for long-wavelength moir\'e systems. To balance efficiency and accuracy, we adopt a two-step transfer learning strategy: (1) the model is pre-trained on a large dataset of computationally inexpensive non-twisted structures until convergence, and (2) the network is then fine-tuned using a small set of computationally expensive twisted structures. Applying this method to twisted MoTe$_2$, the neural network model generates the resulting Hamiltonian for a 1000-atom system in 200 seconds, achieving a mean absolute error below 0.1 meV. To demonstrate $O(N)$ scalability, we model nanoribbon systems with up to 0.25 million atoms ($\sim9$ million orbitals), accurately capturing edge states consistent with predicted Chern numbers. This approach addresses the challenges of accuracy, efficiency, and scalability, offering a viable alternative to conventional DFT and enabling the exploration of electronic topology in large scale moir\'e systems towards simulating realistic device architectures., Comment: 5+14 pages, 4+ 11 figures

We investigate nucleon tensor form factors in the large-$N_{c}$ limit. In this picture, the nucleon emerges as a state of the $N_c$ valence quarks, which were bound by pion mean fields that were created by the presence of the valence quarks self-consistently. We find that the tensor charge ($g^{u-d}_{T}=0.99$) and the anomalous tensor magnetic moment ($\kappa^{u+d}_{T}=7.61$) are dominated by valence quarks, while the tensor quadrupole moment ($Q^{u-d}_{T}=-7.02$) shows significant sea quark effects. We examine how these quantities vary as the average size of the pion mean field is changed, showing interpolation between non-relativistic quark and Skyrme limits. We also observe that $g^{u-d}_{T}$ and $\kappa^{u+d}_{T}$ depend weakly on the pion mass. In contrast, $Q^{u-d}_{T}$ exhibits strong enhancement near the chiral limit. The numerical results are in good agreement with available lattice QCD data and provide predictions for unmeasured quantities., Comment: 20 pages, 3 figures

We present Hunyuan3D 2.0, an advanced large-scale 3D synthesis system for generating high-resolution textured 3D assets. This system includes two foundation components: a large-scale shape generation model -- Hunyuan3D-DiT, and a large-scale texture synthesis model -- Hunyuan3D-Paint. The shape generative model, built on a scalable flow-based diffusion transformer, aims to create geometry that properly aligns with a given condition image, laying a solid foundation for downstream applications. The texture synthesis model, benefiting from strong geometric and diffusion priors, produces high-resolution and vibrant texture maps for either generated or hand-crafted meshes. Furthermore, we build Hunyuan3D-Studio -- a versatile, user-friendly production platform that simplifies the re-creation process of 3D assets. It allows both professional and amateur users to manipulate or even animate their meshes efficiently. We systematically evaluate our models, showing that Hunyuan3D 2.0 outperforms previous state-of-the-art models, including the open-source models and closed-source models in geometry details, condition alignment, texture quality, and etc. Hunyuan3D 2.0 is publicly released in order to fill the gaps in the open-source 3D community for large-scale foundation generative models. The code and pre-trained weights of our models are available at: https://github.com/Tencent/Hunyuan3D-2, Comment: GitHub link: https://github.com/Tencent/Hunyuan3D-2

Background: RETFound, a self-supervised, retina-specific foundation model (FM), showed potential in downstream applications. However, its comparative performance with traditional deep learning (DL) models remains incompletely understood. This study aimed to evaluate RETFound against three ImageNet-pretrained supervised DL models (ResNet50, ViT-base, SwinV2) in detecting ocular and systemic diseases. Methods: We fine-tuned/trained RETFound and three DL models on full datasets, 50%, 20%, and fixed sample sizes (400, 200, 100 images, with half comprising disease cases; for each DR severity class, 100 and 50 cases were used. Fine-tuned models were tested internally using the SEED (53,090 images) and APTOS-2019 (3,672 images) datasets and externally validated on population-based (BES, CIEMS, SP2, UKBB) and open-source datasets (ODIR-5k, PAPILA, GAMMA, IDRiD, MESSIDOR-2). Model performance was compared using area under the receiver operating characteristic curve (AUC) and Z-tests with Bonferroni correction (P<0.05/3). Interpretation: Traditional DL models are mostly comparable to RETFound for ocular disease detection with large datasets. However, RETFound is superior in systemic disease detection with smaller datasets. These findings offer valuable insights into the respective merits and limitation of traditional models and FMs.

Testing and evaluation is an important step before the large-scale application of the autonomous driving systems (ADSs). Based on the three level of scenario abstraction theory, a testing can be performed within a logical scenario, followed by an evaluation stage which is inputted with the testing results of each concrete scenario generated from the logical parameter space. During the above process, abundant testing information is produced which is beneficial for comprehensive and accurate evaluations. To make full use of testing information, this paper proposes an Integrated accelerated Testing and Evaluation Method (ITEM). Based on a Monte Carlo Tree Search (MCTS) paradigm and a dual surrogates testing framework proposed in our previous work, this paper applies the intermediate information (i.e., the tree structure, including the affiliation of each historical sampled point with the subspaces and the parent-child relationship between subspaces) generated during the testing stage into the evaluation stage to achieve accurate hazardous domain identification. Moreover, to better serve this purpose, the UCB calculation method is improved to allow the search algorithm to focus more on the hazardous domain boundaries. Further, a stopping condition is constructed based on the convergence of the search algorithm. Ablation and comparative experiments are then conducted to verify the effectiveness of the improvements and the superiority of the proposed method. The experimental results show that ITEM could well identify the hazardous domains in both low- and high-dimensional cases, regardless of the shape of the hazardous domains, indicating its generality and potential for the safety evaluation of ADSs., Comment: 15 pages, 11 figures

Active Speaker Detection (ASD) aims to identify speaking individuals in complex visual scenes. While humans can easily detect speech by matching lip movements to audio, current ASD models struggle to establish this correspondence, often misclassifying non-speaking instances when audio and lip movements are unsynchronized. To address this limitation, we propose Lip landmark Assisted Speaker dEtection for Robustness (LASER). Unlike models that rely solely on facial frames, LASER explicitly focuses on lip movements by integrating lip landmarks in training. Specifically, given a face track, LASER extracts frame-level visual features and the 2D coordinates of lip landmarks using a lightweight detector. These coordinates are encoded into dense feature maps, providing spatial and structural information on lip positions. Recognizing that landmark detectors may sometimes fail under challenging conditions (e.g., low resolution, occlusions, extreme angles), we incorporate an auxiliary consistency loss to align predictions from both lip-aware and face-only features, ensuring reliable performance even when lip data is absent. Extensive experiments across multiple datasets show that LASER outperforms state-of-the-art models, especially in scenarios with desynchronized audio and visuals, demonstrating robust performance in real-world video contexts. Code is available at \url{https://github.com/plnguyen2908/LASER_ASD}.

In abstract visual reasoning, monolithic deep learning models suffer from limited interpretability and generalization, while existing neuro-symbolic approaches fall short in capturing the diversity and systematicity of attributes and relation representations. To address these challenges, we propose a Systematic Abductive Reasoning model with diverse relation representations (Rel-SAR) in Vector-symbolic Architecture (VSA) to solve Raven's Progressive Matrices (RPM). To derive attribute representations with symbolic reasoning potential, we introduce not only various types of atomic vectors that represent numeric, periodic and logical semantics, but also the structured high-dimentional representation (SHDR) for the overall Grid component. For systematic reasoning, we propose novel numerical and logical relation functions and perform rule abduction and execution in a unified framework that integrates these relation representations. Experimental results demonstrate that Rel-SAR achieves significant improvement on RPM tasks and exhibits robust out-of-distribution generalization. Rel-SAR leverages the synergy between HD attribute representations and symbolic reasoning to achieve systematic abductive reasoning with both interpretable and computable semantics.

The nature of Dark Matter (DM) remains mysterious despite the substantial evidence from astrophysical and cosmological observations. While the majority of DM in our universe is non-relativistic, collisionless and its equation of state (EoS) is approximately pressureless $p\simeq 0$, DM becomes relativistic near the massive black holes in galactic center. Yet its EoS is seldom discussed in the relativistic regime. Here we initially explore the possible equation of state for DM in the vicinity of Schwarzschild black holes. We work in a spherical and quasi-static background spacetime, and describe DM as a perfect fluid in equilibrium. Through numerically solving the TOV equations with physical boundary conditions, we show that DM can have static profiles near black holes and its pressure should be negative in order to support the viable density profiles $\rho$. We illustrate with two simple general equations of state, namely the power law $p \propto \rho^\gamma$ and the radius-dependent $p \propto r\cdot \rho$, and compare them with the observations of the Milky Way. Our findings provide insights into the model-building of DM, which should incorporate the possibility of negative pressure in the relativistic regime around black holes., Comment: 7 pages, 6 figures

We report the fabrication of an 8-meter-long thin-film lithium niobate (TFLN) optical true delay line (OTDL) using the photolithography-assisted chemomechanical etching (PLACE) technique, showing a low transmission loss of 0.036 dB/cm in the conventional telecom band.

The intrinsic integration of Rydberg atomic receivers into wireless communication systems is proposed, by harnessing the principles of quantum physics in wireless communications. More particularly, we conceive a pair of Rydberg atomic receivers, one incorporates a local oscillator (LO), referred to as an LO-dressed receiver, while the other operates without an LO and is termed an LO-free receiver. The appropriate wireless model is developed for each configuration, elaborating on the receiver's responses to the radio frequency (RF) signal, on the potential noise sources, and on the system performance. Next, we investigate the association distortion effects that might occur, specifically demonstrating the boundaries of linear dynamic regions, which provides critical insights into its practical implementations in wireless systems. Extensive simulation results are provided for characterizing the performance of wireless systems, harnessing this pair of Rydberg atomic receivers. Our results demonstrate that they deliver complementary benefits: LO-free systems excel in proximity operations, while LO-dressed systems are eminently suitable for long-distance sensing at extremely low power levels. More specifically, LO-dressed systems achieve a significant signal-to-noise ratio (SNR) gain of approximately 44 dB over conventional RF receivers, exhibiting an effective coverage range extension over conventional RF receivers by a factor of 150. Furthermore, LO-dressed systems support higher-order quadrature amplitude modulation (QAM) at reduced symbol error rates (SER) compared to conventional RF receivers, hence significantly enhancing wireless communication performance., Comment: This manuscript has been submitted to IEEE journal, with 13 pages of body and 2 pages of supplementary material

The technology for generating music from textual descriptions has seen rapid advancements. However, evaluating text-to-music (TTM) systems remains a significant challenge, primarily due to the difficulty of balancing performance and cost with existing objective and subjective evaluation methods. In this paper, we propose an automatic assessment task for TTM models to align with human perception. To address the TTM evaluation challenges posed by the professional requirements of music evaluation and the complexity of the relationship between text and music, we collect MusicEval, the first generative music assessment dataset. This dataset contains 2,748 music clips generated by 31 advanced and widely used models in response to 384 text prompts, along with 13,740 ratings from 14 music experts. Furthermore, we design a CLAP-based assessment model built on this dataset, and our experimental results validate the feasibility of the proposed task, providing a valuable reference for future development in TTM evaluation. The dataset is available at https://www.aishelltech.com/AISHELL_7A., Comment: Accepted by ICASSP 2025

Context. The DESI Legacy Imaging Survey DR9, with its extensive dataset of galaxy locations and photometric redshifts, presents an opportunity to study Baryon Acoustic Oscillations (BAO) in the region covered by the ongoing DESI spectroscopic survey. Aims. We aim to investigate differences between different parts of the DR9 footprint. Furthermore, we want to measure the BAO scale for luminous red galaxies within them. Our selected redshift range of 0.6 to 0.8 corresponds to the bin in which a tension between DESI Y1 and eBOSS was found. Methods. We calculated the anisotropic two-point correlation function in a modified binning scheme to detect the BAO in DR9 data. We then use template fits based on simulations to measure the BAO scale in the imaging data. Results. Our analysis revealed the expected correlation function shape in most of the footprint areas, showing a BAO scale consistent with Planck's observations. Aside from identified mask-related data issues in the southern region of the South Galactic Cap, we also find a notable variance between the different footprints. Conclusions. We find that this variance is consistent with the difference between the DESI Y1 and eBOSS data and it supports the argument that that tension is caused by sample variance. Additionally, we also uncovered systematic biases not previously accounted for in photometric BAO studies. We emphasize the necessity of adjusting for the systematic shift in the BAO scale associated with typical photometric redshift uncertainties to ensure accurate measurements., Comment: 20 pages, 14 figures, 3 tables, accepted for publication in Astronomy and Astrophysics

This paper proposes a Theme-Explanation Structure-based Table Summarization (Tabular-TX) pipeline designed to efficiently process table data. Tabular-TX preprocesses table data by focusing on highlighted cells and then generates summary sentences structured with a Theme Part in the form of adverbial phrases followed by an Explanation Part in the form of clauses. In this process, customized analysis is performed by considering the structural characteristics and comparability of the table. Additionally, by utilizing In-Context Learning, Tabular-TX optimizes the analytical capabilities of large language models (LLMs) without the need for fine-tuning, effectively handling the structural complexity of table data. Results from applying the proposed Tabular-TX to generate table-based summaries demonstrated superior performance compared to existing fine-tuning-based methods, despite limitations in dataset size. Experimental results confirmed that Tabular-TX can process complex table data more effectively and established it as a new alternative for table-based question answering and summarization tasks, particularly in resource-constrained environments., Comment: 6 pages, in Korean language. The 2024 Joint Conference on Human and Cognitive Language Technology, Korean Association for Corpus Linguistics (HCLT-KACL 2024)

In this paper we construct a class of Degenerate Higher-Order Scalar-Tensor (DHOST) theories with an extra scalar field, which admits viable solutions of bouncing universe satisfying the following requirements: (i) absence of Belinski-Khalatnikov-Lifshitz (BKL) instability, ghost and gradient instability, (ii) absence of superluminality, (iii) generation of nearly scale-invariant curvature perturbations and very small tensor-to-scalar ratio, and (iv) conventional asymptotics in the distant past and future, where gravity sector is described by General Relativity and the DHOST scalar has a canonical form of Lagrangian. We also expect our models to have sufficiently small non-Gaussianities of primordial curvature perturbations to be compatible with observations. As such, this work exemplifies for the first time the fully viable two-field DHOST bouncing cosmology, which is free of instability and superluminality problems as well as compatible with observations., Comment: 28 pages, two appendices, 12 figures

Multi-turn interaction in the dialogue system research refers to a system's ability to maintain context across multiple dialogue turns, enabling it to generate coherent and contextually relevant responses. Recent advancements in large language models (LLMs) have significantly expanded the scope of multi-turn interaction, moving beyond chatbots to enable more dynamic agentic interactions with users or environments. In this paper, we provide a focused review of the multi-turn capabilities of LLMs, which are critical for a wide range of downstream applications, including conversational search and recommendation, consultation services, and interactive tutoring. This survey explores four key aspects: (1) the core model capabilities that contribute to effective multi-turn interaction, (2) how multi-turn interaction is evaluated in current practice, (3) the general algorithms used to enhance multi-turn interaction, and (4) potential future directions for research in this field., Comment: Draft Version, 14 pages, Ongoing refinement over time

Quantum backflow, a counterintuitive interference phenomenon where particles with positive momentum can propagate backward, is important in applications involving light-matter interactions. To date, experimental demonstrations of backflow have been restricted to classical optical systems, where momentum is measured using the slit scanning technique or the Shack-Hartmann wavefront sensor technique. However, these techniques have low spatial resolution due to limitations in slit width and Fourier transform lenslet array density. Here, by adopting the technique of weak measurement, we report an observation of azimuthal backflow both theoretically and experimentally. Our results show that a heralded single photon, prepared in specific superposition states with solely negative orbital angular momentum (OAM), exhibits positive OAM. The effects of mode ratio, propagation distance and OAM index on the azimuthal backflow are systematically investigated. Our method avoids using slits and lenslet arrays, allowing for the accurate extraction of photon momentum at each pixel. This work provides new insights and techniques for observing and manipulating backflow in quantum systems., Comment: 5 pages, 3 figures

The non-Hermitian Aharonov-Bohm (AB) cage is a unique localization phenomenon that confines all possible excitations. This confinement leads to fully flat spectrum in momentum space, which are typically accompanied with the degeneracy with various types. Classifying the degeneracy type is crucial for studying the dynamical properties of the non-Hermitian AB cage, but the methods for such classification and their physical connections remain not very clear. Here, we construct a non-Hermitian AB cage in a bosonic Bogoliubov-de Gennes (BdG) system with various type of degenerate flat bands (DFBs). Using the transfer matrix, we demonstrate the localization mechanism for the formation of AB cage and derive the minimal polynomial in mathematics for classifying the degeneracy types of DFBs, thus providing comprehensive understanding of the correspondence among the degeneracy type of DFBs, the minimal polynomial, and the transfer matrix. With such correspondence, we propose a scheme to realize highly degenerate flat bands., Comment: 10 pages, 4 figures, one table

Machine writing with large language models often relies on retrieval-augmented generation. However, these approaches remain confined within the boundaries of the model's predefined scope, limiting the generation of content with rich information. Specifically, vanilla-retrieved information tends to lack depth, utility, and suffers from redundancy, which negatively impacts the quality of generated articles, leading to shallow, repetitive, and unoriginal outputs. To address these issues, we propose OmniThink, a machine writing framework that emulates the human-like process of iterative expansion and reflection. The core idea behind OmniThink is to simulate the cognitive behavior of learners as they progressively deepen their knowledge of the topics. Experimental results demonstrate that OmniThink improves the knowledge density of generated articles without compromising metrics such as coherence and depth. Human evaluations and expert feedback further highlight the potential of OmniThink to address real-world challenges in the generation of long-form articles.

In physics, Lagrangians provide a systematic way to describe laws governing physical systems. In the context of particle physics, they encode the interactions and behavior of the fundamental building blocks of our universe. By treating Lagrangians as complex, rule-based constructs similar to linguistic expressions, we trained a transformer model -- proven to be effective in natural language tasks -- to predict the Lagrangian corresponding to a given list of particles. We report on the transformer's performance in constructing Lagrangians respecting the Standard Model $\mathrm{SU}(3)\times \mathrm{SU}(2)\times \mathrm{U}(1)$ gauge symmetries. The resulting model is shown to achieve high accuracies (over 90\%) with Lagrangians up to six matter fields, with the capacity to generalize beyond the training distribution, albeit within architectural constraints. We show through an analysis of input embeddings that the model has internalized concepts such as group representations and conjugation operations as it learned to generate Lagrangians. We make the model and training datasets available to the community. An interactive demonstration can be found at: \url{https://huggingface.co/spaces/JoseEliel/generate-lagrangians}., Comment: 32 pages, 11 figues, 18 tables

The rare physical property of negative thermal expansion (NTE) is intriguing because materials with large NTE over a wide temperature range can serve as high-performance thermal expansion compensators. However, applications of NTE are hindered by the fact that most of the available NTE materials show small magnitudes of NTE, and/or NTE occurs only in a narrow temperature range. Herein, for the first time, we investigated the effect of anion substitution instead of general Pb/Ti-site substitutions on the thermal expansion properties of a typical ferroelectric NTE material, PbTiO3. Intriguingly, the substitution of S for O in PbTiO3 further increases the tetragonality of PbTiO3. Consequently, an unusually enhanced NTE with an average volumetric coefficient of thermal expansion $\bar{\alpha}_V$ = -2.50 $\times$ 10$^{-5}$/K was achieved over a wide temperature range (300 -- 790 K), which is contrasted to that of pristine PbTiO3 ($\bar{\alpha}_V$ = -1.99 $\times$ 10$^{-5}$/K RT -- 763 K). The intensified NTE is attributed to the enhanced hybridization between Pb/Ti and O/S atoms by the substitution of S, as evidenced by our theoretical investigations. We therefore demonstrate a new technique for introducing mixed anions to achieve large NTE over a wide temperature range in PbTiO3-based ferroelectrics., Comment: 14 pages, 4 figures

Language has long been conceived as an essential tool for human reasoning. The breakthrough of Large Language Models (LLMs) has sparked significant research interest in leveraging these models to tackle complex reasoning tasks. Researchers have moved beyond simple autoregressive token generation by introducing the concept of "thought" -- a sequence of tokens representing intermediate steps in the reasoning process. This innovative paradigm enables LLMs' to mimic complex human reasoning processes, such as tree search and reflective thinking. Recently, an emerging trend of learning to reason has applied reinforcement learning (RL) to train LLMs to master reasoning processes. This approach enables the automatic generation of high-quality reasoning trajectories through trial-and-error search algorithms, significantly expanding LLMs' reasoning capacity by providing substantially more training data. Furthermore, recent studies demonstrate that encouraging LLMs to "think" with more tokens during test-time inference can further significantly boost reasoning accuracy. Therefore, the train-time and test-time scaling combined to show a new research frontier -- a path toward Large Reasoning Model. The introduction of OpenAI's o1 series marks a significant milestone in this research direction. In this survey, we present a comprehensive review of recent progress in LLM reasoning. We begin by introducing the foundational background of LLMs and then explore the key technical components driving the development of large reasoning models, with a focus on automated data construction, learning-to-reason techniques, and test-time scaling. We also analyze popular open-source projects at building large reasoning models, and conclude with open challenges and future research directions., Comment: 36 pages, 5 figures

Deep model merging represents an emerging research direction that combines multiple fine-tuned models to harness their specialized capabilities across different tasks and domains. Current model merging techniques focus on merging all available models simultaneously, with weight interpolation-based methods being the predominant approaches. However, these conventional approaches are not well-suited for scenarios where models become available sequentially, and they often suffer from high memory requirements and potential interference between tasks. In this study, we propose a training-free projection-based continual merging method that processes models sequentially through orthogonal projections of weight matrices and adaptive scaling mechanisms. Our method operates by projecting new parameter updates onto subspaces orthogonal to existing merged parameter updates while using an adaptive scaling mechanism to maintain stable parameter distances, enabling efficient sequential integration of task-specific knowledge. Our approach maintains constant memory complexity to the number of models, minimizes interference between tasks through orthogonal projections, and retains the performance of previously merged models through adaptive task vector scaling. Extensive experiments on CLIP-ViT models demonstrate that our method achieves a 5-8% average accuracy improvement while maintaining robust performance in different task orderings.

Object detection plays a crucial role in smart video analysis, with applications ranging from autonomous driving and security to smart cities. However, achieving real-time object detection on edge devices presents significant challenges due to their limited computational resources and the high demands of deep neural network (DNN)-based detection models, particularly when processing high-resolution video. Conventional strategies, such as input down-sampling and network up-scaling, often compromise detection accuracy for faster performance or lead to higher inference latency. To address these issues, this paper introduces RE-POSE, a Reinforcement Learning (RL)-Driven Partitioning and Edge Offloading framework designed to optimize the accuracy-latency trade-off in resource-constrained edge environments. Our approach features an RL-Based Dynamic Clustering Algorithm (RL-DCA) that partitions video frames into non-uniform blocks based on object distribution and the computational characteristics of DNNs. Furthermore, a parallel edge offloading scheme is implemented to distribute these blocks across multiple edge servers for concurrent processing. Experimental evaluations show that RE-POSE significantly enhances detection accuracy and reduces inference latency, surpassing existing methods.

While large language models (LLMs) present significant potential for supporting numerous real-world applications and delivering positive social impacts, they still face significant challenges in terms of the inherent risk of privacy leakage, hallucinated outputs, and value misalignment, and can be maliciously used for generating toxic content and unethical purposes after been jailbroken. Therefore, in this survey, we present a comprehensive review of recent advancements aimed at mitigating these issues, organized across the four phases of LLM development and usage: data collecting and pre-training, fine-tuning and alignment, prompting and reasoning, and post-processing and auditing. We elaborate on the recent advances for enhancing the performance of LLMs in terms of privacy protection, hallucination reduction, value alignment, toxicity elimination, and jailbreak defenses. In contrast to previous surveys that focus on a single dimension of responsible LLMs, this survey presents a unified framework that encompasses these diverse dimensions, providing a comprehensive view of enhancing LLMs to better serve real-world applications.

Large language models (LLMs), while driving a new wave of interactive AI applications across numerous domains, suffer from high inference costs and heavy cloud dependency. Motivated by the redundancy phenomenon in linguistics, we propose a progressive inference paradigm over cloud and edge, i.e., firstly generating the sketch of the answer by LLMs at cloud, and then conducting parallel extension to fill in details by small models (SLMs) at edge. Progressive inference offers potential benefits to improve throughput and reduce inference latency while facing key implementation challenges, including decreased response quality from SLMs, a tradeoff between the brevity and comprehensiveness of sketches, as well as increased latency caused by network transmission and edge inference. In this work, we propose and implement PICE, an LLM serving system with semantic-level cloud-edge collaboration, enhancing inference throughput and quality through dynamic inference task scheduling, ensemble learning, and parallel edge inference. Extensive testbed experiments illustrate that our approach achieves $1.5-2\times$ throughput enhancement and up to 43% latency reduction, while also potentially enhancing the quality compared to SOTA systems.

Black holes have long served as a testing ground for probing theories of gravity and quantum mechanics. Notably, fundamental fields in the neighborhood of black holes exhibit rich phenomena that could yield astrophysical observable signatures. However, exploring these structures typically requires computationally intensive numerical calculations. In this work, the dynamics of a massive Dirac field outside a Schwarzschild black hole is revisited. We propose a novel matching scheme that enables the analytical solution of the coupled first-order Dirac equation, as opposed to the conventional second-order approach. This method yields a compact and unified analytical expression for the energy spectrum, which shows improved agreement with numerical results. The improvement is due to high-order correction of angular parameter that has been ignored previously., Comment: 10 pages, 2 figures

Multi-modal document retrieval is designed to identify and retrieve various forms of multi-modal content, such as figures, tables, charts, and layout information from extensive documents. Despite its significance, there is a notable lack of a robust benchmark to effectively evaluate the performance of systems in multi-modal document retrieval. To address this gap, this work introduces a new benchmark, named as MMDocIR, encompassing two distinct tasks: page-level and layout-level retrieval. The former focuses on localizing the most relevant pages within a long document, while the latter targets the detection of specific layouts, offering a more fine-grained granularity than whole-page analysis. A layout can refer to a variety of elements such as textual paragraphs, equations, figures, tables, or charts. The MMDocIR benchmark comprises a rich dataset featuring expertly annotated labels for 1,685 questions and bootstrapped labels for 173,843 questions, making it a pivotal resource for advancing multi-modal document retrieval for both training and evaluation. Through rigorous experiments, we reveal that (i) visual retrievers significantly outperform their text counterparts, (ii) MMDocIR train set can effectively benefit the training process of multi-modal document retrieval and (iii) text retrievers leveraging on VLM-text perform much better than those using OCR-text. These findings underscores the potential advantages of integrating visual elements for multi-modal document retrieval., Comment: https://huggingface.co/MMDocIR

Configuration translation is a critical and frequent task in network operations. When a network device is damaged or outdated, administrators need to replace it to maintain service continuity. The replacement devices may originate from different vendors, necessitating configuration translation to ensure seamless network operation. However, translating configurations manually is a labor-intensive and error-prone process. In this paper, we propose an intent-based framework for translating network configuration with Large Language Model (LLM) Agents. The core of our approach is an Intent-based Retrieval Augmented Generation (IRAG) module that systematically splits a configuration file into fragments, extracts intents, and generates accurate translations. We also design a two-stage verification method to validate the syntax and semantics correctness of the translated configurations. We implement and evaluate the proposed method on real-world network configurations. Experimental results show that our method achieves 97.74% syntax correctness, outperforming state-of-the-art methods in translation accuracy.

ABACUS (Atomic-orbital Based Ab-initio Computation at USTC) is an open-source software for first-principles electronic structure calculations and molecular dynamics simulations. It mainly features density functional theory (DFT) and is compatible with both plane-wave basis sets and numerical atomic orbital basis sets. ABACUS serves as a platform that facilitates the integration of various electronic structure methods, such as Kohn-Sham DFT, stochastic DFT, orbital-free DFT, and real-time time-dependent DFT, etc. In addition, with the aid of high-performance computing, ABACUS is designed to perform efficiently and provide massive amounts of first-principles data for generating general-purpose machine learning potentials, such as DPA models. Furthermore, ABACUS serves as an electronic structure platform that interfaces with several AI-assisted algorithms and packages, such as DeePKS-kit, DeePMD, DP-GEN, DeepH, DeePTB, etc.

State-of-the-art supervised stereo matching methods have achieved amazing results on various benchmarks. However, these data-driven methods suffer from generalization to real-world scenarios due to the lack of real-world annotated data. In this paper, we propose StereoGen, a novel pipeline for high-quality stereo image generation. This pipeline utilizes arbitrary single images as left images and pseudo disparities generated by a monocular depth estimation model to synthesize high-quality corresponding right images. Unlike previous methods that fill the occluded area in warped right images using random backgrounds or using convolutions to take nearby pixels selectively, we fine-tune a diffusion inpainting model to recover the background. Images generated by our model possess better details and undamaged semantic structures. Besides, we propose Training-free Confidence Generation and Adaptive Disparity Selection. The former suppresses the negative effect of harmful pseudo ground truth during stereo training, while the latter helps generate a wider disparity distribution and better synthetic images. Experiments show that models trained under our pipeline achieve state-of-the-art zero-shot generalization results among all published methods. The code will be available upon publication of the paper.

Stereo matching recovers depth from image correspondences. Existing methods struggle to handle ill-posed regions with limited matching cues, such as occlusions and textureless areas. To address this, we propose MonSter, a novel method that leverages the complementary strengths of monocular depth estimation and stereo matching. MonSter integrates monocular depth and stereo matching into a dual-branch architecture to iteratively improve each other. Confidence-based guidance adaptively selects reliable stereo cues for monodepth scale-shift recovery. The refined monodepth is in turn guides stereo effectively at ill-posed regions. Such iterative mutual enhancement enables MonSter to evolve monodepth priors from coarse object-level structures to pixel-level geometry, fully unlocking the potential of stereo matching. As shown in Fig.1, MonSter ranks 1st across five most commonly used leaderboards -- SceneFlow, KITTI 2012, KITTI 2015, Middlebury, and ETH3D. Achieving up to 49.5% improvements (Bad 1.0 on ETH3D) over the previous best method. Comprehensive analysis verifies the effectiveness of MonSter in ill-posed regions. In terms of zero-shot generalization, MonSter significantly and consistently outperforms state-of-the-art across the board. The code is publicly available at: https://github.com/Junda24/MonSter.