Search

Large Language Models for Code (LLMs4Code) excel at code generation tasks, yielding promise to release developers from huge software development burdens. Nonetheless, these models have been shown to suffer from the significant privacy risks due to the potential leakage of sensitive information embedded during training, known as the memorization problem. Addressing this issue is crucial for ensuring privacy compliance and upholding user trust, but till now there is a dearth of dedicated studies in the literature that focus on this specific direction. Recently, machine unlearning has emerged as a promising solution by enabling models to "forget" sensitive information without full retraining, offering an efficient and scalable approach compared to traditional data cleaning methods. In this paper, we empirically evaluate the effectiveness of unlearning techniques for addressing privacy concerns in LLMs4Code.Specifically, we investigate three state-of-the-art unlearning algorithms and three well-known open-sourced LLMs4Code, on a benchmark that takes into consideration both the privacy data to be forgotten as well as the code generation capabilites of these models. Results show that it is feasible to mitigate the privacy concerns of LLMs4Code through machine unlearning while maintain their code generation capabilities at the same time. We also dissect the forms of privacy protection/leakage after unlearning and observe that there is a shift from direct leakage to indirect leakage, which underscores the need for future studies addressing this risk., Comment: 11 pages

Multi-modal large language models (MLLMs) have shown remarkable abilities in various visual understanding tasks. However, MLLMs still struggle with fine-grained visual recognition (FGVR), which aims to identify subordinate-level categories from images. This can negatively impact more advanced capabilities of MLLMs, such as object-centric visual question answering and reasoning. In our study, we revisit three quintessential capabilities of MLLMs for FGVR, including object information extraction, category knowledge reserve, object-category alignment, and position of the root cause as a misalignment problem. To address this issue, we present Finedefics, an MLLM that enhances the model's FGVR capability by incorporating informative attribute descriptions of objects into the training phase. We employ contrastive learning on object-attribute pairs and attribute-category pairs simultaneously and use examples from similar but incorrect categories as hard negatives, naturally bringing representations of visual objects and category names closer. Extensive evaluations across multiple popular FGVR datasets demonstrate that Finedefics outperforms existing MLLMs of comparable parameter sizes, showcasing its remarkable efficacy. The code is available at https://github.com/PKU-ICST-MIPL/Finedefics_ICLR2025., Comment: Published as a conference paper at ICLR 2025

Simulation of urban wind environments is crucial for urban planning, pollution control, and renewable energy utilization. However, the computational requirements of high-fidelity computational fluid dynamics (CFD) methods make them impractical for real cities. To address these limitations, this study investigates the effectiveness of the Fourier Neural Operator (FNO) model in predicting flow fields under different wind directions and urban layouts. In this study, we investigate the effectiveness of the Fourier Neural Operator (FNO) model in predicting urban wind conditions under different wind directions and urban layouts. By training the model on velocity data from large eddy simulation data, we evaluate the performance of the model under different urban configurations and wind conditions. The results show that the FNO model can provide accurate predictions while significantly reducing the computational time by 99%. Our innovative approach of dividing the wind field into smaller spatial blocks for training improves the ability of the FNO model to capture wind frequency features effectively. The SDF data also provides important spatial building information, enhancing the model's ability to recognize physical boundaries and generate more realistic predictions. The proposed FNO approach enhances the AI model's generalizability for different wind directions and urban layouts.

In this study, we present a comprehensive analysis of an unidentified point-like ultra-high-energy (UHE) $\gamma$-ray source, designated as 1LHAASO J1740+0948u, situated in the vicinity of the middle-aged pulsar PSR J1740+1000. The detection significance reached 17.1$\sigma$ (9.4$\sigma$) above 25$\,$TeV (100$\,$TeV). The source energy spectrum extended up to 300$\,$TeV, which was well fitted by a log-parabola function with $N0 = (1.93\pm0.23) \times 10^{-16} \rm{TeV^{-1}\,cm^{-2}\,s^{-2}}$, $\alpha = 2.14\pm0.27$, and $\beta = 1.20\pm0.41$ at E0 = 30$\,$TeV. The associated pulsar, PSR J1740+1000, resides at a high galactic latitude and powers a bow-shock pulsar wind nebula (BSPWN) with an extended X-ray tail. The best-fit position of the gamma-ray source appeared to be shifted by $0.2^{\circ}$ with respect to the pulsar position. As the (i) currently identified pulsar halos do not demonstrate such offsets, and (ii) centroid of the gamma-ray emission is approximately located at the extension of the X-ray tail, we speculate that the UHE $\gamma$-ray emission may originate from re-accelerated electron/positron pairs that are advected away in the bow-shock tail.

Weak radiative hyperon decays represent a rich interplay between weak interactions and the internal structure of baryons, offering profound insights into Quantum Chromodynamics and weak interactions. Recent experimental observations, particularly from BESIII, have revealed deviations from theoretical predictions. These deviations could signal new physics or the need for refined theoretical models incorporating intermediate resonance effects. This review discusses recent theoretical advancements and key experimental findings, focusing on recent measurements from BESIII and their implications for strong interactions and baryon structure., Comment: 9 pages, 5 figures, 9 tables

The rapid advancement of deep learning has catalyzed the development of novel IoT applications, which often deploy pre-trained deep neural network (DNN) models across multiple edge devices for collaborative inference.

The search for the absorption of fermionic dark matter (DM) on electron is hindered by a partial or complete loss of sensitivity for the mass of DM ($m_\chi$) below $\mathcal{O} (10)$ keV. We introduce a novel search using a small but high-energy portion of DM boosted by cosmic rays, and the loss of sensitivity is circumvented because the subsequent absorption within the detector imparts to the target electron not only $m_\chi$ but also substantial $T_\chi$. The one-sided 90\% confidence level constraints on ($\sigma_e v_\chi$) are derived with the XENONnT public dataset for $m_\chi$ $\in$ (1$\times$10$^{-5}$--1) keV which has not been explored yet. Our results have placed constraints on DM at the lowest $m_\chi$ currently achievable among the DMDD experiments utilizing ionizing signals from DM interactions within the detectors.

The optimal strategy for deploying a treatment in a population may recommend giving all in the population that treatment. Such a strategy may not be feasible, especially in resource-limited settings. One approach for determining how to allocate a treatment in such settings is the resource-constrained optimal dynamic treatment rule (RC ODTR) SuperLearner algorithm, developed by Luedtke and van der Laan. In this paper, we describe this algorithm, offer various novel approaches for presenting the RC ODTR and its value in terms of benefit and cost, and provide practical guidance on implementing the algorithm (including software). In particular, we apply this method to the Adaptive Strategies for Preventing and Treating Lapses of Retention in HIV care (NCT02338739) trial to determine how to best allocate conditional cash transfers (CCTs) for increasing HIV care adherence given varying constraints on the proportion of people who can receive CCTs in the population, providing one of the few applied illustrations of this method and novel substantive findings. We find that there are clinical and monetary benefits to deploying CCTs to a small percent (e.g., 10\%) of the population compared to administering the care standard for all; however, results suggest that these incremental benefits are only due to the loosening of constraints, rather than a presence of treatment effect heterogeneity strong enough to drive a more efficient and effective constrained allocation approach.

This paper is concerned with non-radiating elastic sources in inhomogeneous elastic media. We demonstrate that the value of non-radiating elastic sources must vanish at convex corners of their support, provided the sources exhibit H\"older continuous regularity near the corner. Additionally, their gradient must satisfy intricate algebraic relationships with the angles defining the underlying corners, assuming the sources have $C^{1,\alpha}$ regularity with $\alpha\in (0,1)$ in the neighborhood of the corners. To perform microlocal analysis around the corners, we employ the so-called complex geometrical optics (CGO) solutions as test functions within a partial differential system. These characterizations of non-radiating elastic sources in inhomogeneous elastic media at corners enable us to establish the unique identifiability results for determining the position and shape of radiating elastic sources by a single far-field measurement, both locally and globally. The uniqueness result by a single far-field measurement is a challenging problem with a colorful history in inverse scattering. Indeed, when the support of a radiating elastic source is a convex polygon, we can simultaneously determine the shape of the elastic source and its values at the corners, provided the source is H\"older continuous at the corner. Furthermore, when the source function exhibits $C^{1,\alpha}$ regularity in the neighborhood of a corner, the gradient of the source function at that corner can also be generically determined.

Using $(2712\pm14)\times10^6$ $\psi(3686)$ events collected with the BESIII detector, we perform the first amplitude analysis of the radiative decay $\psi(3686)\to \gamma K_S^0 K_S^0$ within the mass region $M_{K_S^0 K_S^0 }<2.8$ GeV/$c^2$. Employing a one-channel K-matrix approach for the description of the dynamics of the $K^0_S K^0_S$ system, the data sample is well described with four poles for the $f_0$-wave and three poles for the $f_2$-wave. The determined pole positions are consistent with those of well-established resonance states. The observed $f_0$ and $f_{2}$ states are found to be qualitatively consistent with those produced in radiative $J/\psi$ decays, indicating the similarity between the two charmonium states in their radiative decays., Comment: 20 pages, 4 figures, submitted to JHEP

Large Language Models (LLMs) are vulnerable to attacks like prompt injection, backdoor attacks, and adversarial attacks, which manipulate prompts or models to generate harmful outputs. In this paper, departing from traditional deep learning attack paradigms, we explore their intrinsic relationship and collectively term them Prompt Trigger Attacks (PTA). This raises a key question: Can we determine if a prompt is benign or poisoned? To address this, we propose UniGuardian, the first unified defense mechanism designed to detect prompt injection, backdoor attacks, and adversarial attacks in LLMs. Additionally, we introduce a single-forward strategy to optimize the detection pipeline, enabling simultaneous attack detection and text generation within a single forward pass. Our experiments confirm that UniGuardian accurately and efficiently identifies malicious prompts in LLMs., Comment: 18 Pages, 8 Figures, 5 Tables, Keywords: Attack Defending, Security, Prompt Injection, Backdoor Attacks, Adversarial Attacks, Prompt Trigger Attacks

Detecting stochastic gravitational wave background (SGWB) from cosmic strings is crucial for unveiling the evolutionary laws of the early universe and validating non-standard cosmological models. This study presents the first systematic evaluation of the detection capabilities of next-generation ground-based gravitational wave detector networks for cosmic strings. By constructing a hybrid signal model incorporating multi-source astrophysical foreground noise, including compact binary coalescences (CBCs) and compact binary hyperbolic encounters (CBHEs), we propose an innovative parameter estimation methodology based on multi-component signal separation. Numerical simulations using one-year observational data reveal three key findings: (1) The CE4020ET network, comprising the Einstein Telescope (ET-10 km) and the Cosmic Explorer (CE-40 km and CE-20 km), achieves nearly one order of magnitude improvement in constraining the cosmic string tension $G\mu$ compared to individual detectors, reaching a relative uncertainty $\Delta G\mu / G\mu < 0.5$ for $G\mu > 3.5 \times 10^{-15}$ under standard cosmological framework; (2) The network demonstrates enhanced parameter resolution in non-standard cosmological scenarios, providing a novel approach to probe pre-Big Bang Nucleosynthesis cosmic evolution; (3) Enhanced detector sensitivity amplifies CBHE foreground interference in parameter estimation, while precise modeling of such signals could further refine $G\mu$ constraints by $1-2$ orders of magnitude. This research not only quantifies the detection potential of third-generation detector networks for cosmic string models but also elucidates the intrinsic connection between foreground modeling precision and cosmological parameter estimation accuracy, offering theoretical foundations for optimizing scientific objectives of next-generation gravitational wave observatories., Comment: 15 pages, 9 figures

Let $M=(m_{ij})$ be an $n\times n$ matrix. The second immanant of matrix $M$ is defined by \begin{eqnarray*} d_{2}(M)=\sum_{\sigma\in S_{n}}\chi_{2}(\sigma)\prod_{s=1}^{n}m_{s\sigma(s)}, \end{eqnarray*} where $\chi_{2}$ is the irreducible character of $S_{n}$ corresponding to the partition $(2^{1},1^{n-2})$. The polynomial $d_{2}(xI-M)$ is called the second immanantal polynomial of matrix $M$. Denote by $D(G)$ (resp. $D(\overrightarrow{G})$) and $A(G)$ (resp. $A(\overrightarrow{G})$) the diagonal matrix of vertex degrees and the adjacency matrix of undirected graph $G$ (resp. digraph $\overrightarrow{G}$), respectively. In this article, we prove that $d_{2}(xI-A(G))$ (resp. $d_{2}(xI-A(\overrightarrow{G}))$) can be reconstructed from the second immanantal polynomials of the adjacency matrix of all subgraphs in $\{G-uv,G-u-v|uv\in E(G)\}$ (resp. $\{\overrightarrow{G}-e|e\in E(\overrightarrow{G})\}$). Furthermore, the polynomial $d_{2}(xI-D(\overrightarrow{G})\pm A(\overrightarrow{G}))$ can also be reconstructed by the second immanantal polynomials of the (signless) Laplacian matrixs of all subgraphs in $\{\overrightarrow{G}-e|e\in E(\overrightarrow{G})\}$, respectively.

We study the direct CP asymmetries in the decays of $B^0 \to \Xi_c^+ \overline{\Xi}_c^-$ and $B_s^0 \to \Lambda_c^+ \overline{\Lambda}_c^-$, emphasizing the critical role of final-state interactions (FSIs). In these channels, the small energy release suppresses short-distance contributions while enhancing long-distance effects. By separating the decay amplitudes into S and P waves, we show that the P-wave component, carrying only a single weak phase, contributes negligibly to CP asymmetries. In contrast, the S-wave amplitude, strongly modified by FSIs, acquires a substantial strong phase that enables interference among multiple weak phases, producing large CP-violating effects. Numerically, we find sizable asymmetries of $a_{CP}^{\text{dir}}= 0.65_{-0.34}^{+0.33}$ for $B^0 \to \Xi_c^+ \overline{\Xi}_c^-$ and $a_{CP}^{\text{dir}}= 0.26_{-0.15}^{+0.31}$ for $B_s^0 \to \Lambda_c^+ \overline{\Lambda}_c^-$, where the main uncertainties stem from poorly measured hadron couplings. These results, combined with the relatively large branching fractions, make these modes prime targets for experimental searches at LHCb and Belle II, potentially offering new insights into strong dynamics and nonperturbative QCD in heavy-hadron decays., Comment: 10 pages, 4 figures

Integrated sensing and communication (ISAC) uses the same software and hardware resources to achieve both communication and sensing functionalities. Thus, it stands as one of the core technologies of 6G and has garnered significant attention in recent years. In ISAC systems, a variety of machine learning models are trained to analyze and identify signal patterns, thereby ensuring reliable sensing and communications. However, considering factors such as communication rates, costs, and privacy, collecting sufficient training data from various ISAC scenarios for these models is impractical. Hence, this paper introduces a generative AI (GenAI) enabled robust data augmentation scheme. The scheme first employs a conditioned diffusion model trained on a limited amount of collected CSI data to generate new samples, thereby expanding the sample quantity. Building on this, the scheme further utilizes another diffusion model to enhance the sample quality, thereby facilitating the data augmentation in scenarios where the original sensing data is insufficient and unevenly distributed. Moreover, we propose a novel algorithm to estimate the acceleration and jerk of signal propagation path length changes from CSI. We then use the proposed scheme to enhance the estimated parameters and detect the number of targets based on the enhanced data. The evaluation reveals that our scheme improves the detection performance by up to 70%, demonstrating reliability and robustness, which supports the deployment and practical use of the ISAC network., Comment: 13 pages, 10 figures

Prior studies have shown that distinguishing text generated by large language models (LLMs) from human-written one is highly challenging, and often no better than random guessing. To verify the generalizability of this finding across languages and domains, we perform an extensive case study to identify the upper bound of human detection accuracy. Across 16 datasets covering 9 languages and 9 domains, 19 annotators achieved an average detection accuracy of 87.6%, thus challenging previous conclusions. We find that major gaps between human and machine text lie in concreteness, cultural nuances, and diversity. Prompting by explicitly explaining the distinctions in the prompts can partially bridge the gaps in over 50% of the cases. However, we also find that humans do not always prefer human-written text, particularly when they cannot clearly identify its source.

Complicated time-dependent curved spacetime and electric field are involved in many astrophysical situations, including the early universe, Hawking radiation, the Schwinger effect, and gravitational pair production. In this Letter, a generalized quantum two-level model (GQTLM) is developed, which is applicable to arbitrary time-dependent curved spacetime and electric field. The model is found to be consistent with quantum kinetic theory, and is characterized by its simplicity and versatility. The momentum distribution of particles and the effects of gravitational distortions can be correctly described. Quantum properties concerning vortex structures, such as the intrinsic orbital angular momentum of particles and antiparticles can also be conveniently calculated. The model is expected to significantly advance the quantum exploration of the universe. It could refine the prediction of primordial gravitational waves and relevant non-Gaussian signals, extend the calculation of Hawking radiation to general black hole configurations, help to distinguish neutron stars from strange quark stars, and elucidate the gravitational pair production mechanism., Comment: 13 pages,1 figure

Utilizing 4.5 $fb^-$ of $e^+e^-$ annihilation data collected at center-of-mass energies ranging from 4599.53 MeV to 4698.82 MeV by the BESIII detector at the BEPCII collider, we search for the singly Cabibbo-suppressed hadronic decays $\Lambda_{c}^{+}\to\Sigma^{0} K^{+}\pi^{0}$ and $\Lambda_{c}^{+}\to\Sigma^{0}K^{+}\pi^+\pi^-$ with a single-tag method. No significant signals are observed for both decays. The upper limits on the branching fractions at the $90\%$ confidence level are determined to be $5.0\times 10^{-4}$ for $\Lambda_{c}^{+}\to\Sigma^{0} K^{+}\pi^{0}$ and $6.5\times 10^{-4}$ for $\Lambda_c^{+}\to\Sigma^0K^{+}\pi^{+}\pi^{-}$., Comment: 12 pages, 6 figures

Stock trend prediction involves forecasting the future price movements by analyzing historical data and various market indicators. With the advancement of machine learning, graph neural networks (GNNs) have been extensively employed in stock prediction due to their powerful capability to capture spatiotemporal dependencies of stocks. However, despite the efforts of various GNN stock predictors to enhance predictive performance, the improvements remain limited, as they focus solely on analyzing historical spatiotemporal dependencies, overlooking the correlation between historical and future patterns. In this study, we propose a novel distillation-based future-aware GNN framework (DishFT-GNN) for stock trend prediction. Specifically, DishFT-GNN trains a teacher model and a student model, iteratively. The teacher model learns to capture the correlation between distribution shifts of historical and future data, which is then utilized as intermediate supervision to guide the student model to learn future-aware spatiotemporal embeddings for accurate prediction. Through extensive experiments on two real-world datasets, we verify the state-of-the-art performance of DishFT-GNN.

Integrated sensing and communication (ISAC) has garnered substantial research interest owing to its pivotal role in advancing the development of next-generation (6G) wireless networks. However, achieving a performance balance between communication and sensing in the dual-function radar communication (DFRC)-based ISAC system remains a significant challenge. In this paper, an aerial intelligent reflecting surface (IRS)-assisted ISAC system is explored, where a base station (BS) supports dual-functional operations, enabling both data transmission for multiple users and sensing for a blocked target, with the channel quality enhanced by an IRS mounted on the unmanned aerial vehicle (UAV). Moreover, we formulate an integrated communication, sensing, and energy efficiency multi-objective optimization problem (CSEMOP), which aims to maximize the communication rate of the users and the echo rate of the target, while minimizing UAV propulsion energy consumption by jointly optimizing the BS beamforming matrix, IRS phase shifts, the flight velocity and angle of the UAV. Considering the non-convexity, trade-off, and dynamic nature of the formulated CSEMOP, we propose a generative diffusion model-based deep deterministic policy gradient (GDMDDPG) method to solve the problem. Specifically, the diffusion model is incorporated into the actor network of DDPG to improve the action quality, with noise perturbation mechanism for better exploration and recent prioritized experience replay (RPER) sampling mechanism for enhanced training efficiency. Simulation results indicate that the GDMDDPG method delivers superior performance compared to the existing methods.

Recent advancements in intelligent reflecting surfaces (IRS) and mobile edge computing (MEC) offer new opportunities to enhance the performance of vehicular networks. However, meeting the computation-intensive and latency-sensitive demands of vehicles remains challenging due to the energy constraints and dynamic environments. To address this issue, we study an IRS-assisted MEC architecture for vehicular networks. We formulate a multi-objective optimization problem aimed at minimizing the total task completion delay and total energy consumption by jointly optimizing task offloading, IRS phase shift vector, and computation resource allocation. Given the mixed-integer nonlinear programming (MINLP) and NP-hard nature of the problem, we propose a generative diffusion model (GDM)-based Stackelberg game (GDMSG) approach. Specifically, the problem is reformulated within a Stackelberg game framework, where generative GDM is integrated to capture complex dynamics to efficiently derive optimal solutions. Simulation results indicate that the proposed GDMSG achieves outstanding performance compared to the benchmark approaches., Comment: 6 pages, 5 figures, published to ICC

Federated Foundation Models (FedFMs) represent a distributed learning paradigm that fuses general competences of foundation models as well as privacy-preserving capabilities of federated learning. This combination allows the large foundation models and the small local domain models at the remote clients to learn from each other in a teacher-student learning setting. This paper provides a comprehensive summary of the ten challenging problems inherent in FedFMs, encompassing foundational theory, utilization of private data, continual learning, unlearning, Non-IID and graph data, bidirectional knowledge transfer, incentive mechanism design, game mechanism design, model watermarking, and efficiency. The ten challenging problems manifest in five pivotal aspects: ``Foundational Theory," which aims to establish a coherent and unifying theoretical framework for FedFMs. ``Data," addressing the difficulties in leveraging domain-specific knowledge from private data while maintaining privacy; ``Heterogeneity," examining variations in data, model, and computational resources across clients; ``Security and Privacy," focusing on defenses against malicious attacks and model theft; and ``Efficiency," highlighting the need for improvements in training, communication, and parameter efficiency. For each problem, we offer a clear mathematical definition on the objective function, analyze existing methods, and discuss the key challenges and potential solutions. This in-depth exploration aims to advance the theoretical foundations of FedFMs, guide practical implementations, and inspire future research to overcome these obstacles, thereby enabling the robust, efficient, and privacy-preserving FedFMs in various real-world applications.

Diffusion language models have emerged as a promising approach for text generation. One would naturally expect this method to be an efficient replacement for autoregressive models since multiple tokens can be sampled in parallel during each diffusion step. However, its efficiency-accuracy trade-off is not yet well understood. In this paper, we present a rigorous theoretical analysis of a widely used type of diffusion language model, the Masked Diffusion Model (MDM), and find that its effectiveness heavily depends on the target evaluation metric. Under mild conditions, we prove that when using perplexity as the metric, MDMs can achieve near-optimal perplexity in sampling steps regardless of sequence length, demonstrating that efficiency can be achieved without sacrificing performance. However, when using the sequence error rate--which is important for understanding the "correctness" of a sequence, such as a reasoning chain--we show that the required sampling steps must scale linearly with sequence length to obtain "correct" sequences, thereby eliminating MDM's efficiency advantage over autoregressive models. Our analysis establishes the first theoretical foundation for understanding the benefits and limitations of MDMs. All theoretical findings are supported by empirical studies., Comment: 32 pages, 3 figures

With a statistical analysis of arXiv paper abstracts, we report a marked drop in the frequency of several words previously identified as overused by ChatGPT, such as "delve", starting soon after they were pointed out in early 2024. The frequency of certain other words favored by ChatGPT, such as "significant", has instead kept increasing. These phenomena suggest that some authors of academic papers have adapted their use of large language models (LLMs), for example, by selecting outputs or applying modifications to the LLM-generated content. Such coevolution and cooperation of humans and LLMs thus introduce additional challenges to the detection of machine-generated text in real-world scenarios. Estimating the impact of LLMs on academic writing by examining word frequency remains feasible, and more attention should be paid to words that were already frequently employed, including those that have decreased in frequency due to LLMs' disfavor.

In 1995, D. Jerison and C. Kenig in \cite{JK-1995} considered the the inhomogeneous Dirichlet problem $\Delta u= f$ on $\Omega$, $u=0$ on $\partial\Omega$ in Lipschitz domains. One of their main results shows that the $W^{1,p}$ estimate holds for the sharp range $\frac{3}{2}-\varepsilon

Retrieval-augmented generation (RAG) is a key technique for leveraging external knowledge and reducing hallucinations in large language models (LLMs). However, RAG still struggles to fully prevent hallucinated responses. To address this, it is essential to identify samples prone to hallucination or guide LLMs toward correct responses, which experts then annotate to develop high-quality datasets for refining LLMs. However, the growing scarcity of such datasets makes their creation challenging. This paper proposes using the vast amount of conversations from widespread LLM usage to build these datasets, training LLMs to avoid hallucination-prone questions while accurately responding to manageable ones. Given the impracticality of expert-annotating all conversation records, the paper introduces AL4RAG, which uses active learning to select the most suitable conversation samples for annotation, optimizing performance within an annotation budget. Additionally, recognizing that traditional active learning methods are not fully compatible with RAG due to unsuitable distance metrics, we develop a novel sample distance measurement for RAG active learning. Extensive experiments show that our method consistently outperforms baselines across multiple metrics.

This paper proposes a UAV-assisted forwarding system based on distributed beamforming to enhance age of information (AoI) in Internet of Things (IoT). Specifically, UAVs collect and relay data between sensor nodes (SNs) and the remote base station (BS). However, flight delays increase the AoI and degrade the network performance. To mitigate this, we adopt distributed beamforming to extend the communication range, reduce the flight frequency and ensure the continuous data relay and efficient energy utilization. Then, we formulate an optimization problem to minimize AoI and UAV energy consumption, by jointly optimizing the UAV trajectories and communication schedules. The problem is non-convex and with high dynamic, and thus we propose a deep reinforcement learning (DRL)-based algorithm to solve the problem, thereby enhancing the stability and accelerate convergence speed. Simulation results show that the proposed algorithm effectively addresses the problem and outperforms other benchmark algorithms., Comment: 6 pages, 4 figures, ICC2025

By analyzing a $\psi(3686)$ data sample containing $(107.7\pm0.6)\times10^{6}$ events taken with the BESIII detector at the BEPCII storage ring in 2009, the $\chi_{c0}$ resonance parameters are precisely measured using $\chi_{c0,c2} \to \pi^+\pi^-/K^+K^-$ events. The mass of $\chi_{c0}$ is determined to be $M(\chi_{c0})=(3415.67\pm0.07\pm0.06\pm0.07$)~MeV/$c^2$, and its full width is $\Gamma(\chi_{c0})=(12.44\pm0.12\pm0.12)~{\rm MeV}$, where the first uncertainty is statistical, the second systematic, and the third for mass comes from $\chi_{c2}$ mass uncertainty. These measurements improve the precision of $\chi_{c0}$ mass by a factor of four and width by one order of magnitude over the previous individual measurements, and significantly boost our knowledge about the charmonium spectrum. Together with additional $(345.4\pm2.6)\times10^{6}$ $\psi(3686)$ data events taken in 2012, the decay branching fractions of $\chi_{c0,c2}\to\pi^+\pi^-/K^+K^-$ are measured as well, with precision improved by a factor of three compared to previous measurements. These $\chi_{c0}$ decay branching fractions provide important inputs for the study of glueballs., Comment: 9 pages, 1 figure

Entanglement islands are subregions in a gravitational universe whose information is fully encoded in a disconnected non-gravitational system away from it. In the context of the black hole information paradox, entanglement islands state that the information about the black hole interior is encoded in the early-time Hawking radiation. Nevertheless, it was unclear how this seemingly nonlocal information encoding emerges from a manifestly local theory. In this paper, we provide an answer to this question by uncovering the mechanism behind this information encoding scheme. As we will see, the early understanding that graviton is massive in island models plays an essential role in this mechanism. As an example, we will discuss how this mechanism works in detail in the Karch-Randall braneworld. This study also suggests the potential importance of this mechanism to the ER=EPR conjecture., Comment: 39 pages, 6 figures

We report the position reconstruction methods and surface background model for the PandaX-4T dark matter direct search experiment. This work develops two position reconstruction algorithms: template matching (TM) method and photon acceptance function (PAF) method. Both methods determine the horizontal position of events based on the light pattern of secondary scintillation collected by the light sensors. After a comprehensive evaluation of resolution, uniformity, and robustness, the PAF method was selected for position reconstruction, while the TM method was employed for verification. The PAF method achieves a bulk event resolution of 1.0 mm and a surface event resolution of 4.4 mm for a typical $S2$ signal with a bottom charge of 1500 PE (about 14 keV). The uniformity is around 20\%. Robustness studies reveal average deviations of 5.1 mm and 8.8 mm for the commissioning run (Run0) and the first science run (Run1), respectively, due to the deactivation of certain PMTs. A data-driven surface background model is developed based on the PAF method. The surface background is estimated to be $0.09 \pm 0.06$ events for Run0 (0.54 tonne$\cdot$year) and $0.17 \pm 0.11$ events for Run1 (1.00 tonne$\cdot$year)., Comment: 22 pages, 15 figures, 2 tables

Sequential multiple assignment randomized trials (SMARTs) have grown in popularity in recent years, and many of their study protocols propose conducting a cost effectiveness analysis of the adaptive strategies embedded within them. The cost effectiveness of these regimes is often proposed to be assessed using incremental cost effectiveness ratios (ICERs). In this paper, we present an estimation and inference procedure for such cost effectiveness measures for the embedded dynamic treatment regimes within a SMART design. In particular, we describe a targeted maximum likelihood estimator for the ICER of a SMART's embedded regimes with influence curve-based inference. We illustrate the performance of these methods using simulations. Throughout, we use as illustration a cost effectiveness analysis for the Adaptive Strategies for Preventing and Treating Lapses of Retention in HIV Care (ADAPT-R; NCT02338739) trial, presenting estimated ICERs (with inference) for embedded regimes aimed at increasing HIV care adherence. This manuscript is one of the first to present cost effectiveness analysis results from a SMART.

Distributed Learning (DL) enables the training of machine learning models across multiple devices, yet it faces challenges like non-IID data distributions and device capability disparities, which can impede training efficiency. Communication bottlenecks further complicate traditional Federated Learning (FL) setups. To mitigate these issues, we introduce the Personalized Federated Learning with Decentralized Selection Training (PFedDST) framework. PFedDST enhances model training by allowing devices to strategically evaluate and select peers based on a comprehensive communication score. This score integrates loss, task similarity, and selection frequency, ensuring optimal peer connections. This selection strategy is tailored to increase local personalization and promote beneficial peer collaborations to strengthen the stability and efficiency of the training process. Our experiments demonstrate that PFedDST not only enhances model accuracy but also accelerates convergence. This approach outperforms state-of-the-art methods in handling data heterogeneity, delivering both faster and more effective training in diverse and decentralized systems.

Using $e^+e^-$ collision data at 13 center-of-mass energies ranging from 4.600 to 4.950 GeV collected with the BESIII detector, we search for the unmeasured $e^+e^-\to K_S^0 K_S^0 h_c$ process . No significant signal is observed, and the upper limits of the Born cross sections at each center-of-mass energy are presented.

Unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) and data collection (DC) have been popular research issues. Different from existing works that consider MEC and DC scenarios separately, this paper investigates a multi-UAV-assisted joint MEC-DC system. Specifically, we formulate a joint optimization problem to minimize the MEC latency and maximize the collected data volume. This problem can be classified as a non-convex mixed integer programming problem that exhibits long-term optimization and dynamics. Thus, we propose a deep reinforcement learning-based approach that jointly optimizes the UAV movement, user transmit power, and user association in real time to solve the problem efficiently. Specifically, we reformulate the optimization problem into an action space-reduced Markov decision process (MDP) and optimize the user association by using a two-phase matching-based association (TMA) strategy. Subsequently, we propose a soft actor-critic (SAC)-based approach that integrates the proposed TMA strategy (SAC-TMA) to solve the formulated joint optimization problem collaboratively. Simulation results demonstrate that the proposed SAC-TMA is able to coordinate the two subsystems and can effectively reduce the system latency and improve the data collection volume compared with other benchmark algorithms., Comment: 18 pages, 10 figures

We show that it is impossible to experimentally observe the quantum intrinsic orbital angular momentum (IOAM) effect without its axial current. Broadly speaking, we argue that the spiral or interference characteristics of the axial current density determine the occurrence of nonlinear or tunneling effects in any spacetimedependent quantum systems. Our findings offer a comprehensive theoretical framework that addresses the limitations of Keldysh theory and provides new insights into the angular momentum properties of quantum systems, particularly in tunneling-dominated regimes. Using Wigner function methods, fermionic generalized two-level model, and Berry phase simulations, we predict that IOAM effect can persist even in pure quantum tunneling processes. These results open the door for experimental verification of IOAM effects in future high-intensity QED experiments, such as those using X-ray free electron lasers., Comment: 5 pages, 2 figures

This study investigates 180 MX5Y5 (X, Y = B, C, or N) clathrate-like structures derived from MH10 superhydrides using high-throughput Density Functional Theory (DFT) geometry optimizations and phonon calculations. A wide variety of electropositive and electronegative encapsulated atoms were considered. From all of the studied systems only 31 MB5N5 phases were found to be dynamically stable at ambient pressure. The highest 1-atmosphere superconducting critical transition temperature was predicted for FB5N5. However, ab initio molecular dynamics simulations revealed that all of the identified superconducting phases decompose by 300~K at 1~atm, while only nine semiconducting phases remained thermally stable. Our findings underscore the critical role of kinetic and thermal stability in predicting viable superconductors. The electronic structure of the MB5N5 compounds were rationalized in terms of electron donating and withdrawing intercalants, and machine-learning based predictions of their mechanical properties were compared with those of an empty boronitride cage.

Unmanned aerial vehicles (UAVs) have emerged as the potential aerial base stations (BSs) to improve terrestrial communications. However, the limited onboard energy and antenna power of a UAV restrict its communication range and transmission capability. To address these limitations, this work employs collaborative beamforming through a UAV-enabled virtual antenna array to improve transmission performance from the UAV to terrestrial mobile users, under interference from non-associated BSs and dynamic channel conditions. Specifically, we introduce a memory-based random walk model to more accurately depict the mobility patterns of terrestrial mobile users. Following this, we formulate a multi-objective optimization problem (MOP) focused on maximizing the transmission rate while minimizing the flight energy consumption of the UAV swarm. Given the NP-hard nature of the formulated MOP and the highly dynamic environment, we transform this problem into a multi-objective Markov decision process and propose an improved evolutionary multi-objective reinforcement learning algorithm. Specifically, this algorithm introduces an evolutionary learning approach to obtain the approximate Pareto set for the formulated MOP. Moreover, the algorithm incorporates a long short-term memory network and hyper-sphere-based task selection method to discern the movement patterns of terrestrial mobile users and improve the diversity of the obtained Pareto set. Simulation results demonstrate that the proposed method effectively generates a diverse range of non-dominated policies and outperforms existing methods. Additional simulations demonstrate the scalability and robustness of the proposed CB-based method under different system parameters and various unexpected circumstances.

Recent studies have verified that semi-supervised learning (SSL) is vulnerable to data poisoning backdoor attacks. Even a tiny fraction of contaminated training data is sufficient for adversaries to manipulate up to 90\% of the test outputs in existing SSL methods. Given the emerging threat of backdoor attacks designed for SSL, this work aims to protect SSL against such risks, marking it as one of the few known efforts in this area. Specifically, we begin by identifying that the spurious correlations between the backdoor triggers and the target class implanted by adversaries are the primary cause of manipulated model predictions during the test phase. To disrupt these correlations, we utilize three key techniques: Gaussian Filter, complementary learning and trigger mix-up, which collectively filter, obstruct and dilute the influence of backdoor attacks in both data pre-processing and feature learning. Experimental results demonstrate that our proposed method, Backdoor Invalidator (BI), significantly reduces the average attack success rate from 84.7\% to 1.8\% across different state-of-the-art backdoor attacks. It is also worth mentioning that BI does not sacrifice accuracy on clean data and is supported by a theoretical guarantee of its generalization capability.

We present the first lattice result of the near threshold $\Lambda_c\Lambda_c$ scattering with $I(J^P) = 0(0^+)$. The calculation is performed on two $N_f = 2+1$ Wilson-Clover ensembles with pion mass $m_\pi \approx 303 $MeV and lattice spacing $a\approx 0.07746$fm. The L\"uscher's finite volume method is utilized to extract the scattering parameters from the finite-volume spectrum. The coupled channel $\Xi_{cc}N$ is ignored in the scattering analysis based on the observation that the energy levels computed from the $\Lambda_c\Lambda_c$ and $\Xi_{cc}N$ operators do not mix. The $\Sigma_c\Sigma_c$ channel is not included either since the energy range explored in this study is well below its threshold. Our results indicate that the interaction in the $\Lambda_c\Lambda_c$ single channel is repulsive, and the scattering length is determined to be $a_0 = -0.21(4)(7) \mathrm{fm}$, where the first error is the statistical error and the second is the systematic error., Comment: 9 pages, 6 figures

Machine learning is widely believed to be one of the most promising practical applications of quantum computing. Existing quantum machine learning schemes typically employ a quantum-classical hybrid approach that relies crucially on gradients of model parameters. Such an approach lacks provable convergence to global minima and will become infeasible as quantum learning models scale up. Here, we introduce quantum automated learning, where no variational parameter is involved and the training process is converted to quantum state preparation. In particular, we encode training data into unitary operations and iteratively evolve a random initial state under these unitaries and their inverses, with a target-oriented perturbation towards higher prediction accuracy sandwiched in between. Under reasonable assumptions, we rigorously prove that the evolution converges exponentially to the desired state corresponding to the global minimum of the loss function. We show that such a training process can be understood from the perspective of preparing quantum states by imaginary time evolution, where the data-encoded unitaries together with target-oriented perturbations would train the quantum learning model in an automated fashion. We further prove that the quantum automated learning paradigm features good generalization ability with the generalization error upper bounded by the ratio between a logarithmic function of the Hilbert space dimension and the number of training samples. In addition, we carry out extensive numerical simulations on real-life images and quantum data to demonstrate the effectiveness of our approach and validate the assumptions. Our results establish an unconventional quantum learning strategy that is gradient-free with provable and explainable trainability, which would be crucial for large-scale practical applications of quantum computing in machine learning scenarios., Comment: 21 pages, 7 figures

There has been significant interest in exploring topological disclination states, which effectively probe the band topology of the host material beyond the conventional bulk-edge correspondence. While most studies in this area have primarily focused on Hermitian systems, recent theoretical work predicts that non-Hermiticity can drive topological phase transitions and host topological disclination states associated with fractional charge. However, no experimental observations have been reported to date. Here, we report the first experimental observation of topological disclination states in electric circuits, induced solely by gain and loss. Through admittance matrix measurements and eigenstate analysis, we confirm their emergence and compute the corresponding fractional charge. Moreover, the disclination mode profile and localization effect can be directly visualized via monochromatic field excitation. Additionally, we demonstrate the emergence of degenerate zero-energy topological disclination states, devoid of fractional charge, in distinct non-Hermitian geometries. Our findings open the possibility of non-Hermiticity-induced fractional charges in two-dimensional non-Hermitian lattices, which may pave the way for advancements in active topological photonic devices., Comment: 16 pages, 4 figures

The core-collapse supernova remnant (SNR) Cassiopeia A (Cas A) is one of the brightest galactic radio sources with an angular radius of $\sim$ 2.5 $\arcmin$. Although no extension of this source has been detected in the $\gamma$-ray band, using more than 1000 days of LHAASO data above $\sim 0.8$ TeV, we find that its spectrum is significantly softer than those obtained with Imaging Air Cherenkov Telescopes (IACTs) and its flux near $\sim 1$ TeV is about two times higher. In combination with analyses of more than 16 years of \textit{Fermi}-LAT data covering $0.1 \, \mathrm{GeV} - 1 \, \mathrm{TeV}$, we find that the spectrum above 30 GeV deviates significantly from a single power-law, and is best described by a smoothly broken power-law with a spectral index of $1.90 \pm 0.15_\mathrm{stat}$ ($3.41 \pm 0.19_\mathrm{stat}$) below (above) a break energy of $0.63 \pm 0.21_\mathrm{stat} \, \mathrm{TeV}$. Given differences in the angular resolution of LHAASO-WCDA and IACTs, TeV $\gamma$-ray emission detected with LHAASO may have a significant contribution from regions surrounding the SNR illuminated by particles accelerated earlier, which, however, are treated as background by IACTs. Detailed modelling can be used to constrain acceleration processes of TeV particles in the early stage of SNR evolution.

By analyzing 7.93 $\rm fb^{-1}$ of $e^+e^-$ collision data collected at the center-of-mass energy of 3.773 GeV with the BESIII detector operated at the BEPCII collider, we report the observation of the semimuonic decays of $D^+\to \bar K_1(1270)^0\mu^+\nu_\mu$ and $D^0\to K_1(1270)^-\mu^+\nu_\mu$ with statistical significances of $12.5\sigma$ and $6.0\sigma$, respectively. Their decay branching fractions are determined to be ${\mathcal B}[D^{+}\to \bar{K}_1(1270)^0 \mu^{+}\nu_{\mu}]=(2.36\pm0.20^{+0.18}_{-0.27}\pm 0.48)\times10^{-3}$ and ${\mathcal B}[D^{0}\to K_1(1270)^{-} \mu^{+}\nu_{\mu}]=(0.78\pm0.11^{+0.05}_{-0.09}\pm 0.15)\times10^{-3}$, where the first and second uncertainties are statistical and systematic, respectively, and the third originates from the input branching fraction of $\bar K_{1}(1270)^0\to K^- \pi^+\pi^0$ or $K_1(1270)^-\to K^-\pi^+\pi^-$. Combining our branching fractions with the previous measurements of ${\mathcal B}[D^+\to \bar K_1(1270)^0e^+\nu_{e}]$ and ${\mathcal B}[D^0\to K_1(1270)^-e^+\nu_{e}]$, we determine the branching fraction ratios to be ${\mathcal B}[D^+\to \bar K_1(1270)^0\mu^+\nu_{\mu}]/{\mathcal B}[D^+\to \bar K_1(1270)^0e^+\nu_{e}]=1.03 \pm 0.14 \substack{+0.11\\-0.15}$ and ${\mathcal B}[D^0\to K_1(1270)^-\mu^+\nu_{\mu}]/{\mathcal B}[D^0\to K_1(1270)^-e^+\nu_{e}]=0.74\pm 0.13 \substack{+0.08\\-0.13}$. Using the branching fractions measured in this work and the world-average lifetimes of the $D^+$ and $D^0$ mesons, we determine the semimuonic partial decay width ratio to be $\Gamma [D^+\to \bar K_1(1270)^0 \mu^+\nu_\mu]/\Gamma [D^0\to K_1(1270)^- \mu^+\nu_\mu]=1.22\pm 0.10\substack{+0.06\\-0.09}$, which is consistent with unity as predicted by isospin conservation., Comment: 10 pages, 2 figures

Large language models (LLMs) excel across various tasks, but standard first-order (FO) fine-tuning demands considerable memory, significantly limiting real-world deployment. Recently, zeroth-order (ZO) optimization stood out as a promising memory-efficient training paradigm, avoiding backward passes and relying solely on forward passes for gradient estimation, making it attractive for resource-constrained scenarios. However, ZO method lags far behind FO method in both convergence speed and accuracy. To bridge the gap, we introduce a novel layer-wise divergence analysis that uncovers the distinct update pattern of FO and ZO optimization. Aiming to resemble the learning capacity of FO method from the findings, we propose \textbf{Di}vergence-driven \textbf{Z}eroth-\textbf{O}rder (\textbf{DiZO}) optimization. DiZO conducts divergence-driven layer adaptation by incorporating projections to ZO updates, generating diverse-magnitude updates precisely scaled to layer-wise individual optimization needs. Our results demonstrate that DiZO significantly reduces the needed iterations for convergence without sacrificing throughput, cutting training GPU hours by up to 48\% on various datasets. Moreover, DiZO consistently outperforms the representative ZO baselines in fine-tuning RoBERTa-large, OPT-series, and Llama-series on downstream tasks and, in some cases, even surpasses memory-intensive FO fine-tuning.