Search

Pre-training a language model and then fine-tuning it has shown to be an efficient and effective technique for a wide range of code intelligence tasks, such as code generation, code summarization, and vulnerability detection. However, pretraining language models on a large-scale code corpus is computationally expensive. Fortunately, many off-the-shelf Pre-trained Code Models (PCMs), such as CodeBERT, CodeT5, CodeGen, and Code Llama, have been released publicly. These models acquire general code understanding and generation capability during pretraining, which enhances their performance on downstream code intelligence tasks. With an increasing number of these public pre-trained models, selecting the most suitable one to reuse for a specific task is essential. In this paper, we systematically investigate the reusability of PCMs. We first explore three intuitive model selection methods that select by size, training data, or brute-force fine-tuning. Experimental results show that these straightforward techniques either perform poorly or suffer high costs. Motivated by these findings, we explore learning-based model selection strategies that utilize pre-trained models without altering their parameters. Specifically, we train proxy models to gauge the performance of pre-trained models, and measure the distribution deviation between a model's latent features and the task's labels, using their closeness as an indicator of model transferability. We conduct experiments on 100 widely-used opensource PCMs for code intelligence tasks, with sizes ranging from 42.5 million to 3 billion parameters. The results demonstrate that learning-based selection methods reduce selection time to 100 seconds, compared to 2,700 hours with brute-force fine-tuning, with less than 6% performance degradation across related tasks., Comment: Accepted by IEEE SANER 2025

Version control systems are commonly used to manage open-source software, in which each commit may introduce new vulnerabilities or fix existing ones. Researchers have developed various tools for detecting vulnerabilities in code commits, but their performance is limited by factors such as neglecting descriptive data and challenges in accurately identifying vulnerability introductions. To overcome these limitations, we propose CommitShield, which combines the code analysis capabilities of static analysis tools with the natural language and code understanding capabilities of large language models (LLMs) to enhance the accuracy of vulnerability introduction and fix detection by generating precise descriptions and obtaining rich patch contexts. We evaluate CommitShield using the newly constructed vulnerability repair dataset, CommitVulFix, and a cleaned vulnerability introduction dataset. Experimental results indicate that CommitShield improves recall by 76%-87% over state-of-the-art methods in the vulnerability fix detection task, and its F1-score improves by 15%-27% in the vulnerability introduction detection task.

As the foundation of the Web3 trust system, blockchain technology faces increasing demands for scalability. Sharding emerges as a promising solution, but it struggles to handle highly concurrent cross-shard transactions (\textsf{CSTx}s), primarily due to simultaneous ledger operations on the same account. Hyperledger Fabric, a permissioned blockchain, employs multi-version concurrency control for parallel processing. Existing solutions use channels and intermediaries to achieve cross-sharding in Hyperledger Fabric. However, the conflict problem caused by highly concurrent \textsf{CSTx}s has not been adequately resolved. To fill this gap, we propose HiCoCS, a high concurrency cross-shard scheme for permissioned blockchains. HiCoCS creates a unique virtual sub-broker for each \textsf{CSTx} by introducing a composite key structure, enabling conflict-free concurrent transaction processing while reducing resource overhead. The challenge lies in managing large numbers of composite keys and mitigating intermediary privacy risks. HiCoCS utilizes virtual sub-brokers to receive and process \textsf{CSTx}s concurrently while maintaining a transaction pool. Batch processing is employed to merge multiple \textsf{CSTx}s in the pool, improving efficiency. We explore composite key reuse to reduce the number of virtual sub-brokers and lower system overhead. Privacy preservation is enhanced using homomorphic encryption. Evaluations show that HiCoCS improves cross-shard transaction throughput by 3.5-20.2 times compared to the baselines., Comment: Major Revised in IEEE TC

Satisfactory progress has been achieved recently in universal segmentation of CT images. Following the success of vision-language methods, there is a growing trend towards utilizing text prompts and contrastive learning to develop universal segmentation models. However, there exists a significant imbalance in information density between 3D images and text prompts. Moreover, the standard fully connected layer segmentation approach faces significant challenges in handling multiple classes and exhibits poor generalizability. To address these challenges, we propose the VOxel Interacting with LAnguage method (VOILA) for universal CT image segmentation. Initially, we align voxels and language into a shared representation space and classify voxels on the basis of cosine similarity. Subsequently, we develop the Voxel-Language Interaction framework to mitigate the impact of class imbalance caused by foreground-background discrepancies and variations in target volumes. Furthermore, a Complexity-Aware Sampling method is proposed to focus on region hard to segment, achieved by generating pseudo-heatmaps from a trainable Gaussian mixture distribution. Our results indicate the proposed VOILA is capable to achieve improved performance with reduced parameters and computational cost during training. Furthermore, it demonstrates significant generalizability across diverse datasets without additional fine-tuning., Comment: Accepted by AAAI 2025

Predicting plant species composition in specific spatiotemporal contexts plays an important role in biodiversity management and conservation, as well as in improving species identification tools. Our work utilizes 88,987 plant survey records conducted in specific spatiotemporal contexts across Europe. We also use the corresponding satellite images, time series data, climate time series, and other rasterized environmental data such as land cover, human footprint, bioclimatic, and soil variables as training data to train the model to predict the outcomes of 4,716 plant surveys. We propose a feature construction and result correction method based on the graph structure. Through comparative experiments, we select the best-performing backbone networks for feature extraction in both temporal and image modalities. In this process, we built a backbone network based on the Swin-Transformer Block for extracting temporal Cubes features. We then design a hierarchical cross-attention mechanism capable of robustly fusing features from multiple modalities. During training, we adopt a 10-fold cross-fusion method based on fine-tuning and use a Threshold Top-K method for post-processing. Ablation experiments demonstrate the improvements in model performance brought by our proposed solution pipeline., Comment: CVPR GeolifeCLEF

Based on $(2712.4 \pm 14.3)\times 10^6$ $\psi(3686)$ events collected at the BESIII detector operating at the BEPCII collider, we present the first observation of the decay $\psi(3686) \to K^{-}\Lambda(1520)\bar{\Xi}^{+} + c.c.$. The product branching fraction ${\cal B}[\psi(3686) \to K^{-}\Lambda(1520)\bar{\Xi}^{+} + c.c.] \times {\cal B}[\Lambda(1520) \to pK^{-}]$ is measured to be $(9.5 \pm 0.8 \pm 1.1) \times 10^{-7}$, where the first uncertainty is statistical and the second systematic.

The advancement of magnonics has facilitated the utilization of hybrid magnetic systems in quantum technologies. A hybrid magnetic lattice (HML), comprising an array of superconducting loops and magnetic particles, has been devised as a quantum bus to disseminate quantum resources among magnetic quantum entities (MQEs) serving as nodes of a quantum network. However, the HML also exerts a decoherence effect on the MQEs, which has the potential to impair its practical performance. By studying the non-Markovian dynamics of two MQEs comprised of either nitrogen-vacancy centers or magnon modes coupled to two independent HMLs, we propose a Floquet-engineering scheme by applying periodic driving on the MQEs to overcome the unwanted effect. It is revealed that the decoherence can be suppressed and a significant degree of entanglement can be maintained in the steady state, provided that a FBS exists within the quasienergy spectrum of the total system of each periodically driven MQE and its HML. This result enhances our ability to control hybrid magnetic systems and is beneficial for the application of HML in quantum networks.

A subgroup $H\le G$ is a quasi-retract of $G$ if there exists a quasi-homomorphism from $G$ to $H$ which restricts to the identity on $H$. In this paper, we study under what conditions a subgroup is a quasi-retract of the whole group. As an intermediate result, we show that for any short exact sequence of groups $1\to H\to G\to G/H\to 1$, the following are equivalent: (i) it is left quasi-split (or equivalently, $H$ is a quasi-retract of $G$). (ii) it is right quasi-split and $H$ almost commutes with a right coset transversal of $H$ in $G$. (iii) $G$ is strictly quasi-isomorphic to $H\times G/H$. This is an analogue to the classical result on split short exact sequence of groups. As a corollary, we present a method to detect whether a finite-index subgroup is a quasi-retract. Besides, we show that any quasi-retract of a non-elementary torsion-free group acting properly on a metric space with contracting elements is core-free., Comment: 16 pages

A search for $\eta_c(2S)\to p\bar{p}K^+K^-$, together with measurement of branching fractions of $\chi_{cJ(J=0,1,2)}\to p\bar{p}K^+K^-$ in the $\psi(3686) \to \gamma \eta_c(2S)$ and the $\psi(3686) \to \gamma \chi_{cJ}$ radiative decays, is performed with $(2712.4\pm14.3)\times 10^6$ $\psi(3686)$ events collected with the BESIII detector at the BEPCII collider. An evidence for $\eta_c(2S)\to p\bar{p}K^+K^-$ is found, with a significance of $3.3\sigma$. The product branching fraction of $\mathcal{B}[\psi(3686)\to\gamma\eta_c(2S)]\cdot\mathcal{B}[\eta_c(2S)\to p\bar{p}K^+K^-]$ is determined to be $(1.98\mkern 2mu\pm\mkern 2mu0.41_{\text{stat.}}\mkern 2mu\pm\mkern 2mu0.99_{\text{syst.}})\times 10^{-7}$. The product branching fractions of $\mathcal{B}[\psi(3686)\to\gamma\chi_{cJ}]\cdot\mathcal{B}[\chi_{cJ}\to p\bar{p}K^+K^-]$ are measured to be $(2.49\mkern 2mu\pm\mkern 2mu 0.03_{\text{stat.}}\mkern 2mu\pm\mkern 2mu 0.15_{\text{syst.}})\times 10^{-5}$, $(1.83\mkern 2mu \pm\mkern 2mu 0.02_{\text{stat.}}\mkern 2mu \pm\mkern 2mu 0.11_{\text{syst.}})\times 10^{-5}$, and $(2.43\mkern 2mu\pm\mkern 2mu 0.02_{\text{stat.}}\mkern 2mu\pm\mkern 2mu 0.15_{\text{syst.}})\times 10^{-5}$, for $J=0,\ 1$, and 2, respectively., Comment: 12 pages, 2 figures

Graph convolutional networks (GCNs) have demonstrated superiority in graph-based learning tasks. However, training GCNs on full graphs is particularly challenging, due to the following two challenges: (1) the associated feature tensors can easily explode the memory and block the communication bandwidth of modern accelerators, and (2) the computation workflow in training GCNs alternates between sparse and dense matrix operations, complicating the efficient utilization of computational resources. Existing solutions for scalable distributed full-graph GCN training mostly adopt partition parallelism, which is unsatisfactory as they only partially address the first challenge while incurring scaled-out communication volume. To this end, we propose MixGCN aiming to simultaneously address both the aforementioned challenges towards GCN training. To tackle the first challenge, MixGCN integrates mixture of parallelism. Both theoretical and empirical analysis verify its constant communication volumes and enhanced balanced workload; For handling the second challenge, we consider mixture of accelerators (i.e., sparse and dense accelerators) with a dedicated accelerator for GCN training and a fine-grain pipeline. Extensive experiments show that MixGCN achieves boosted training efficiency and scalability., Comment: 15 pages, 12 figures, 5 tables

Remote sensing (RS) change detection incurs a high cost because of false negatives, which are more costly than false positives. Existing frameworks, struggling to improve the Precision metric to reduce the cost of false positive, still have limitations in focusing on the change of interest, which leads to missed detections and discontinuity issues. This work tackles these issues by enhancing feature learning capabilities and integrating the frequency components of feature information, with a strategy to incrementally boost the Recall value. We propose an enhanced hybrid of CNN and Transformer network (EHCTNet) for effectively mining the change information of interest. Firstly, a dual branch feature extraction module is used to extract the multi scale features of RS images. Secondly, the frequency component of these features is exploited by a refined module I. Thirdly, an enhanced token mining module based on the Kolmogorov Arnold Network is utilized to derive semantic information. Finally, the semantic change information's frequency component, beneficial for final detection, is mined from the refined module II. Extensive experiments validate the effectiveness of EHCTNet in comprehending complex changes of interest. The visualization outcomes show that EHCTNet detects more intact and continuous changed areas and perceives more accurate neighboring distinction than state of the art models.

To address the growing size of AI model training data and the lack of a universal data selection methodology-factors that significantly drive up training costs -- this paper presents the General Information Metrics Evaluation (GIME) method. GIME leverages general information metrics from Objective Information Theory (OIT), including volume, delay, scope, granularity, variety, duration, sampling rate, aggregation, coverage, distortion, and mismatch to optimize dataset selection for training purposes. Comprehensive experiments conducted across diverse domains, such as CTR Prediction, Civil Case Prediction, and Weather Forecasting, demonstrate that GIME effectively preserves model performance while substantially reducing both training time and costs. Additionally, applying GIME within the Judicial AI Program led to a remarkable 39.56% reduction in total model training expenses, underscoring its potential to support efficient and sustainable AI development.

Evaluating and ranking the capabilities of different LLMs is crucial for understanding their performance and alignment with human preferences. Due to the high cost and time-consuming nature of human evaluations, an automatic LLM bencher (i.e., an automatic evaluation framework that aims to rank LLMs based on their alignment with human preferences) is indispensable. An automatic LLM bencher consists of four components: the input set (e.g., a user instruction), the evaluation model (e.g., an LLM), the evaluation type (e.g., pairwise comparison), and the aggregation method (e.g., the ELO rating system). However, previous work has not thoroughly explored how to select these components or how their different combinations influence the results. In this work, through controlled experiments, we provide a series of recommendations on how to choose each component to better automate the evaluation of LLMs. Furthermore, we discovered that when evaluating LLMs with similar performance, the performance of the automatic LLM bencher declines sharply, underscoring the limitations of current benchers and calling for future work. Lastly, we found that the evaluation models' performance at the instance level (e.g., the accuracy of selecting the best output) does not always align with their effectiveness when used as a component of a bencher, highlighting the importance of dedicated system-level evaluation of benchers.

This paper addresses the scarcity of large-scale datasets for accurate object-in-hand pose estimation, which is crucial for robotic in-hand manipulation within the ``Perception-Planning-Control" paradigm. Specifically, we introduce VinT-6D, the first extensive multi-modal dataset integrating vision, touch, and proprioception, to enhance robotic manipulation. VinT-6D comprises 2 million VinT-Sim and 0.1 million VinT-Real splits, collected via simulations in MuJoCo and Blender and a custom-designed real-world platform. This dataset is tailored for robotic hands, offering models with whole-hand tactile perception and high-quality, well-aligned data. To the best of our knowledge, the VinT-Real is the largest considering the collection difficulties in the real-world environment so that it can bridge the gap of simulation to real compared to the previous works. Built upon VinT-6D, we present a benchmark method that shows significant improvements in performance by fusing multi-modal information. The project is available at https://VinT-6D.github.io/.

This work studies the proof theory of left (right) skew monoidal closed categories and skew monoidal bi-closed categories from the perspective of non-associative Lambek calculus. Skew monoidal closed categories represent a relaxed version of monoidal closed categories, where the structural laws are not invertible; instead, they are natural transformations with a specific orientation. Uustalu et al. used sequents with stoup (the leftmost position of an antecedent that can be either empty or a single formula) to deductively model left skew monoidal closed categories, yielding results regarding proof identities and categorical coherence. However, their syntax does not work well when modeling right skew monoidal closed and skew monoidal bi-closed categories. We solve the problem by constructing cut-free sequent calculi for left skew monoidal closed and skew monoidal bi-closed categories, reminiscent of non-associative Lambek calculus, with trees as antecedents. Each calculus is respectively equivalent to the sequent calculus with stoup (for left skew monoidal categories) and the axiomatic calculus (for skew monoidal bi-closed categories). Moreover, we prove that the latter calculus is sound and complete with respect to its relational models. We also prove a correspondence between frame conditions and structural laws, providing an algebraic way to understand the relationship between the left and right skew monoidal (closed) categories., Comment: In Proceedings NCL'24, arXiv:2412.20053

Let $\mathfrak{B}_{m|n}(\underline{\boldsymbol{z}})$ be the Gaudin algebra of the general linear Lie superalgebra $\mathfrak{gl}_{m|n}$ with respect to a sequence $\underline{\boldsymbol{z}} \in \mathbb{C}^\ell$ of pairwise distinct complex numbers, and let $M$ be any $\ell$-fold tensor product of irreducible polynomial modules over $\mathfrak{gl}_{m|n}$. We show that the singular space $M^{\rm sing}$ of $M$ is a cyclic $\mathfrak{B}_{m|n}(\underline{\boldsymbol{z}})$-module and the Gaudin algebra $\mathfrak{B}_{m|n}(\underline{\boldsymbol{z}})_{M^{\rm sing}}$ of $M^{\rm sing}$ is a Frobenius algebra. We also show that $\mathfrak{B}_{m|n}(\underline{\boldsymbol{z}})_{M^{\rm sing}}$ is diagonalizable with a simple spectrum for a generic $\underline{\boldsymbol{z}}$ and give a description of an eigenbasis and its corresponding eigenvalues in terms of the Fuchsian differential operators with polynomial kernels. This may be interpreted as the completeness of a reformulation of the Bethe ansatz for $\mathfrak{B}_{m|n}(\underline{\boldsymbol{z}})_{M^{\rm sing}}$.

Knowledge Graphs (KGs) provide a structured representation of knowledge but often suffer from challenges of incompleteness. To address this, link prediction or knowledge graph completion (KGC) aims to infer missing new facts based on existing facts in KGs. Previous knowledge graph embedding models are limited in their ability to capture expressive features, especially when compared to deeper, multi-layer models. These approaches also assign a single static embedding to each entity and relation, disregarding the fact that entities and relations can exhibit different behaviors in varying graph contexts. Due to complex context over a fact triple of a KG, existing methods have to leverage complex non-linear context encoder, like transformer, to project entity and relation into low dimensional representations, resulting in high computation cost. To overcome these limitations, we propose Triple Receptance Perception (TRP) architecture to model sequential information, enabling the learning of dynamic context of entities and relations. Then we use tensor decomposition to calculate triple scores, providing robust relational decoding capabilities. This integration allows for more expressive representations. Experiments on benchmark datasets such as YAGO3-10, UMLS, FB15k, and FB13 in link prediction and triple classification tasks demonstrate that our method performs better than several state-of-the-art models, proving the effectiveness of the integration.

Built upon the previous work on the 4d anomalies and 5d cobordism invariants (namely, 5d invertible field theory [iTFT] or symmetry-protected topological state [SPTs]) of the Standard Model [SM] gauge theory with compatible (SU(3)$\times$SU(2)$\times$U(1))/$\mathbb{Z}_q$ gauge group for $q=1,2,3,6$, we further enumerate lower-dimensional iTFT / SPTs in 4d, 3d, 2d, and 1d. While the 4d SPTs are interesting gapped phases attachable to the SM, those integer classes of SPTs (either the torsion or the free cobordism class) are not universal and are difficult to detect. However, fractional SPTs are more universal and can be detected by a topological response similar to the Hall conductance under an appropriate symmetry twist background field. We find a nontrivial symmetry-extension from a magnetic 1-symmetry U(1)$_{[1]}^m$ by an electric 1-symmetry $\mathbb{Z}_{6/q,[1]}^e$ introducing the $k \in \mathbb{Z}_{6/q}$ fractionalization class. While we also introduce the background fields for Baryon minus Lepton number $({\bf B}-{\bf L})$ like 0-symmetries of U(1)$_{{\bf Q} - N_c {\bf L}}$, U(1)$_X$, and $\mathbb{Z}_{4,X}$, that introduces two more data $\alpha \in 1,3$ and $\beta \in 1,5$ for the SMs. We present their Hall conductance response data (denoted as $\sigma, \sigma'$, and $\sigma''$ respectively), between 0-symmetry and 1-symmetry background fields, and show their fractional dependence on the SM data $(q,\alpha,\beta, k)$., Comment: Sequel to: arXiv:1910.14668, arXiv:2112.14765, arXiv:2204.08393

Speech Emotion Recognition (SER) involves analyzing vocal expressions to determine the emotional state of speakers, where the comprehensive and thorough utilization of audio information is paramount. Therefore, we propose a novel approach on self-supervised learning (SSL) models that employs all available auxiliary information -- specifically metadata -- to enhance performance. Through a two-stage fine-tuning method in multi-task learning, we introduce the Augmented Residual Integration (ARI) module, which enhances transformer layers in encoder of SSL models. The module efficiently preserves acoustic features across all different levels, thereby significantly improving the performance of metadata-related auxiliary tasks that require various levels of features. Moreover, the Co-attention module is incorporated due to its complementary nature with ARI, enabling the model to effectively utilize multidimensional information and contextual relationships from metadata-related auxiliary tasks. Under pre-trained base models and speaker-independent setup, our approach consistently surpasses state-of-the-art (SOTA) models on multiple SSL encoders for the IEMOCAP dataset., Comment: accepted by ICASSP2025. \c{opyright}2025 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component

Generating high-quality whole-body human object interaction motion sequences is becoming increasingly important in various fields such as animation, VR/AR, and robotics. The main challenge of this task lies in determining the level of involvement of each hand given the complex shapes of objects in different sizes and their different motion trajectories, while ensuring strong grasping realism and guaranteeing the coordination of movement in all body parts. Contrasting with existing work, which either generates human interaction motion sequences without detailed hand grasping poses or only models a static grasping pose, we propose a simple yet effective framework that jointly models the relationship between the body, hands, and the given object motion sequences within a single diffusion model. To guide our network in perceiving the object's spatial position and learning more natural grasping poses, we introduce novel contact-aware losses and incorporate a data-driven, carefully designed guidance. Experimental results demonstrate that our approach outperforms the state-of-the-art method and generates plausible whole-body motion sequences., Comment: Accepted by AAAI 2025

The twist phase of random light represents a nontrivial two-point phase, endowing the field with orbital angular momentum. Although the mutual transition of the spin and orbit angular momenta of coherent light has been revealed, the relationship between spin-orbital angular momentum interaction (SOI) and the twist phase has remained unexplored. This is because of the stochastic nature of random light, making it challenging to explore the properties of angular momenta that rely on well-defined spatial and polarization structures. This study addresses this gap from the view of the asymmetry coherent-mode decomposition for twisted random light to gain insight into the intricate interplay between the twist phase and the SOI within a tight focusing system. Our findings reveal that spin and orbit angular momentum transitions occur in the tightly focused twisted random light beam, yielding the transverse spin density controlled by the twist phase. This effect becomes more pronounced when the spin of random light and the chirality of the twist phase are the same. Our work may find significant applications in optical sensing, metrology, and quantum optics.

Using $e^+e^-$ collision data collected with the BESIII detector at the BEPCII collider at thirty-two center-of-mass energies from 3.50 to 4.95 GeV, corresponding to an integrated luminosity of 25 $\rm{fb^{-1}}$, we measure the Born cross section of the $e^+e^-\to\Sigma^0\bar{\Sigma}^0$ reaction and the effective form factor. No significant charmonium(-like) state, i.e., $\psi(3770)$, $\psi(4040)$, $\psi(4160)$, $\psi(4230)$, $\psi(4360)$, $\psi(4415)$, or $\psi(4660)$, decaying into the $\Sigma^0\bar{\Sigma}^0$ final state is observed by fitting the $e^+e^- \to \Sigma^0\bar{\Sigma}^0$ dressed cross section. The upper limits for the product of the branching fraction and the electronic partial width at the 90% confidence level are provided for each assumed charmonium(-like) state. In addition, the ratios of the Born cross section and the effective form factor between the $e^+e^-\to\Sigma^0\bar{\Sigma}^0$ and the $e^+e^-\to\Sigma^+\bar{\Sigma}^-$ reactions are provided, which can be used to validate the prediction of the vector meson dominance model., Comment: 9 pages, 3 figures, 1 Supplemental Material

Kerr stability for small angular momentum has been proved in the series of works by Klainerman-Szeftel, Giorgi-Klainerman-Szeftel and Shen. Some of the most basic conclusions of the result, concerning various physical quantities on the future null infinity are derived in the work of Klainerman-Szeftel. Further important conclusions were later derived in An-He-Shen and Chen-Klainerman. In this paper, based on the existence and uniqueness results for GCM spheres by Klainerman-Szeftel, we establish the existence of a canonical foliation on the future null infinity for which the null energy, linear momentum, center of mass and angular momentum are well defined and satisfy the expected physical laws of gravitational radiation. The rigid character of this foliation eliminates the usual ambiguities related to these quantities in the physics literature. We also show that under the initial assumption of Klainerman-Szeftel, the center of mass of the black hole has a large deformation (recoil) after the perturbation., Comment: 82 pages, 5 figures, 2 tables. Comments are welcome

Large Language Models (LLMs) have significantly transformed our daily life and established a new paradigm in natural language processing (NLP). However, the predominant pretraining of LLMs on extensive web-based texts remains insufficient for advanced scientific discovery, particularly in chemistry. The scarcity of specialized chemistry data, coupled with the complexity of multi-modal data such as 2D graph, 3D structure and spectrum, present distinct challenges. Although several studies have reviewed Pretrained Language Models (PLMs) in chemistry, there is a conspicuous absence of a systematic survey specifically focused on chemistry-oriented LLMs. In this paper, we outline methodologies for incorporating domain-specific chemistry knowledge and multi-modal information into LLMs, we also conceptualize chemistry LLMs as agents using chemistry tools and investigate their potential to accelerate scientific research. Additionally, we conclude the existing benchmarks to evaluate chemistry ability of LLMs. Finally, we critically examine the current challenges and identify promising directions for future research. Through this comprehensive survey, we aim to assist researchers in staying at the forefront of developments in chemistry LLMs and to inspire innovative applications in the field., Comment: COLING2025,We maintain an up-to-date Github repository at: https://github.com/OpenDFM/LLM4Chemistry

This paper presents a combined approach to enhancing the effectiveness of Jacobian-Free Newton-Krylov (JFNK) method by deep reinforcement learning (DRL) in identifying fixed points within the 2D Kuramoto-Sivashinsky Equation (KSE). JFNK approach entails a good initial guess for improved convergence when searching for fixed points. With a properly defined reward function, we utilise DRL as a preliminary step to enhance the initial guess in the converging process. We report new results of fixed points in the 2D KSE which have not been reported in the literature. Additionally, we explored control optimization for the 2D KSE to navigate the system trajectories between known fixed points, based on parallel reinforcement learning techniques. This combined method underscores the improved JFNK approach to finding new fixed-point solutions within the context of 2D KSE, which may be instructive for other high-dimensional dynamical systems.

Using a data sample of $(10087 \pm 44) \times {10^{6}}$ $J/{\psi}$ events collected with the BESIII detector, we search for the decays $\eta/\eta'\to e^+e^-\mu^+\mu^-$ and $\eta' \to \mu^+\mu^-\mu^+\mu^-$ via the radiative decays $J/{\psi}\to\gamma\eta$/$\gamma\eta'$. No excess of events over expected background is observed for any of the decays of interest. At 90% confidence level, we report the first upper limits on the branching fractions of $\eta' \to e^{+}e^{-}\mu^{+}\mu^{-}$ and $\eta' \to \mu^{+}\mu^{-}\mu^{+}\mu^{-}$ to be $ 1.75 \times {10^{-6}}$ and $5.28 \times {10^{-7}}$, respectively. In addition, we set an upper limit on the branching fraction of $\eta \to e^{+}e^{-}\mu^{+}\mu^{-}$ to be $6.88 \times {10^{-6}}$, which improves the previous result by about two orders of magnitude., Comment: 11 pages

Drug-target interaction is fundamental in understanding how drugs affect biological systems, and accurately predicting drug-target affinity (DTA) is vital for drug discovery. Recently, deep learning methods have emerged as a significant approach for estimating the binding strength between drugs and target proteins. However, existing methods simply utilize the drug's local information from molecular topology rather than global information. Additionally, the features of drugs and proteins are usually fused with a simple concatenation operation, limiting their effectiveness. To address these challenges, we proposed ViDTA, an enhanced DTA prediction framework. We introduce virtual nodes into the Graph Neural Network (GNN)-based drug feature extraction network, which acts as a global memory to exchange messages more efficiently. By incorporating virtual graph nodes, we seamlessly integrate local and global features of drug molecular structures, expanding the GNN's receptive field. Additionally, we propose an attention-based linear feature fusion network for better capturing the interaction information between drugs and proteins. Experimental results evaluated on various benchmarks including Davis, Metz, and KIBA demonstrate that our proposed ViDTA outperforms the state-of-the-art baselines., Comment: Accepted by International Conference on Bioinformatics and Biomedicine (BIBM 24)

Cross-view geo-localization (CVGL) has been widely applied in fields such as robotic navigation and augmented reality. Existing approaches primarily use single images or fixed-view image sequences as queries, which limits perspective diversity. In contrast, when humans determine their location visually, they typically move around to gather multiple perspectives. This behavior suggests that integrating diverse visual cues can improve geo-localization reliability. Therefore, we propose a novel task: Cross-View Image Set Geo-Localization (Set-CVGL), which gathers multiple images with diverse perspectives as a query set for localization. To support this task, we introduce SetVL-480K, a benchmark comprising 480,000 ground images captured worldwide and their corresponding satellite images, with each satellite image corresponds to an average of 40 ground images from varied perspectives and locations. Furthermore, we propose FlexGeo, a flexible method designed for Set-CVGL that can also adapt to single-image and image-sequence inputs. FlexGeo includes two key modules: the Similarity-guided Feature Fuser (SFF), which adaptively fuses image features without prior content dependency, and the Individual-level Attributes Learner (IAL), leveraging geo-attributes of each image for comprehensive scene perception. FlexGeo consistently outperforms existing methods on SetVL-480K and two public datasets, SeqGeo and KITTI-CVL, achieving a localization accuracy improvement of over 22% on SetVL-480K.

Diffusion-based generative models have become the standard for image generation. ODE-based samplers and flow matching models improve efficiency, in comparison to diffusion models, by reducing sampling steps through learned vector fields. However, the theoretical foundations of flow matching models remain limited, particularly regarding the convergence of individual sample trajectories at terminal time - a critical property that impacts sample quality and being critical assumption for models like the consistency model. In this paper, we advance the theory of flow matching models through a comprehensive analysis of sample trajectories, centered on the denoiser that drives ODE dynamics. We establish the existence, uniqueness and convergence of ODE trajectories at terminal time, ensuring stable sampling outcomes under minimal assumptions. Our analysis reveals how trajectories evolve from capturing global data features to local structures, providing the geometric characterization of per-sample behavior in flow matching models. We also explain the memorization phenomenon in diffusion-based training through our terminal time analysis. These findings bridge critical gaps in understanding flow matching models, with practical implications for sampling stability and model design., Comment: Fixed typos and improved organization for clarity

We investigate the topological properties of multiple exceptional points in non-Hermitian two-level systems, emphasizing vorticity as a topological invariant arising from complex energy structures. We categorize EP pairs as fundamental building blocks of larger EP assemblies, distinguishing two types: type-I pairs with opposite vorticities and type-II pairs with identical vorticities. By analyzing the branch cut formation in a two-dimensional parameter space, we reveal the distinct topological features of each EP pair type. Furthermore, we extend our analysis to configurations with multiple EPs, demonstrating the cumulative vorticity and topological implications. To illustrate these theoretical structures, we model complex energy bands within a two-dimensional photonic crystal composed of lossy materials, identifying various EP pairs and their branch cuts. These findings contribute to the understanding of topological characteristics in non-Hermitian systems., Comment: 8 pages, 6 figures

With advances in foundational and vision-language models, and effective fine-tuning techniques, a large number of both general and special-purpose models have been developed for a variety of visual tasks. Despite the flexibility and accessibility of these models, no single model is able to handle all tasks and/or applications that may be envisioned by potential users. Recent approaches, such as visual programming and multimodal LLMs with integrated tools aim to tackle complex visual tasks, by way of program synthesis. However, such approaches overlook user constraints (e.g., performance / computational needs), produce test-time sample-specific solutions that are difficult to deploy, and, sometimes, require low-level instructions that maybe beyond the abilities of a naive user. To address these limitations, we introduce MMFactory, a universal framework that includes model and metrics routing components, acting like a solution search engine across various available models. Based on a task description and few sample input-output pairs and (optionally) resource and/or performance constraints, MMFactory can suggest a diverse pool of programmatic solutions by instantiating and combining visio-lingual tools from its model repository. In addition to synthesizing these solutions, MMFactory also proposes metrics and benchmarks performance / resource characteristics, allowing users to pick a solution that meets their unique design constraints. From the technical perspective, we also introduced a committee-based solution proposer that leverages multi-agent LLM conversation to generate executable, diverse, universal, and robust solutions for the user. Experimental results show that MMFactory outperforms existing methods by delivering state-of-the-art solutions tailored to user problem specifications. Project page is available at https://davidhalladay.github.io/mmfactory_demo.

Tensor Attention extends traditional attention mechanisms by capturing high-order correlations across multiple modalities, addressing the limitations of classical matrix-based attention. Meanwhile, Rotary Position Embedding ($\mathsf{RoPE}$) has shown superior performance in encoding positional information in long-context scenarios, significantly enhancing transformer models' expressiveness. Despite these empirical successes, the theoretical limitations of these technologies remain underexplored. In this study, we analyze the circuit complexity of Tensor Attention and $\mathsf{RoPE}$-based Tensor Attention, showing that with polynomial precision, constant-depth layers, and linear or sublinear hidden dimension, they cannot solve fixed membership problems or $(A_{F,r})^*$ closure problems, under the assumption that $\mathsf{TC}^0 \neq \mathsf{NC}^1$. These findings highlight a gap between the empirical performance and theoretical constraints of Tensor Attention and $\mathsf{RoPE}$-based Tensor Attention Transformers, offering insights that could guide the development of more theoretically grounded approaches to Transformer model design and scaling.

Contraction Hierarchies (CH) (Geisberger et al., 2008) is one of the most widely used algorithms for shortest-path queries on road networks. Compared to Dijkstra's algorithm, CH enables orders of magnitude faster query performance through a preprocessing phase, which iteratively categorizes vertices into hierarchies and adds shortcuts. However, constructing a CH is an expensive task. Existing solutions, including parallel ones, may suffer from long construction time. Especially in our experiments, we observe that existing parallel solutions demonstrate unsatisfactory scalability and have close performance to sequential algorithms. In this paper, we present \textsf{SPoCH} (\textbf{S}calable \textbf{P}arallelization \textbf{o}f \textbf{C}ontraction \textbf{H}ierarchies), an efficient and scalable CH construction algorithm in parallel. To address the challenges in previous work, our improvements focus on both redesigning the algorithm and leveraging parallel data structures. %to maintain the original and shortcut edges dynamically. We implement our algorithm and compare it with the state-of-the-art sequential and parallel implementations on 13 graphs, including road networks, synthetic graphs, and $k$-NN graphs. Our experiments show that \textsf{SPoCH} achieves $17$--$131\times$ speedups in CH construction over the best baseline, while maintaining competitive query performance and CH graph size.

We implement a network-based approach to study expertise in a complex real-world task: operating particle accelerators. Most real-world tasks we learn and perform (e.g., driving cars, operating complex machines, solving mathematical problems) are difficult to learn because they are complex, and the best strategies are difficult to find from many possibilities. However, how we learn such complex tasks remains a partially solved mystery, as we cannot explain how the strategies evolve with practice due to the difficulties of collecting and modeling complex behavioral data. As complex tasks are generally networks of many elementary subtasks, we model task performance as networks or graphs of subtasks and investigate how the networks change with expertise. We develop the networks by processing the text in a large archive of operator logs from 14 years of operations using natural language processing and machine learning. The network changes are examined using a set of measures at four levels of granularity - individual subtasks, interconnections among subtasks, groups of subtasks, and the whole complex task. We find that the operators consistently change with expertise at the subtask, the interconnection, and the whole-task levels, but they show remarkable similarity in how subtasks are grouped. These results indicate that the operators of all stages of expertise adopt a common divide-and-conquer approach by breaking the complex task into parts of manageable complexity, but they differ in the frequency and structure of nested subtasks. Operational logs are common data sources from real-world settings where people collaborate with hardware and software environments to execute complex tasks, and the network models investigated in this study can be expanded to accommodate multi-modal data. Therefore, our network-based approach provides a practical way to investigate expertise in the real world.

As deep neural networks (DNNs) are widely applied in the physical world, many researches are focusing on physical-world adversarial examples (PAEs), which introduce perturbations to inputs and cause the model's incorrect outputs. However, existing PAEs face two challenges: unsatisfactory attack performance (i.e., poor transferability and insufficient robustness to environment conditions), and difficulty in balancing attack effectiveness with stealthiness, where better attack effectiveness often makes PAEs more perceptible. In this paper, we explore a novel perturbation-based method to overcome the challenges. For the first challenge, we introduce a strategy Deceptive RF injection based on robust features (RFs) that are predictive, robust to perturbations, and consistent across different models. Specifically, it improves the transferability and robustness of PAEs by covering RFs of other classes onto the predictive features in clean images. For the second challenge, we introduce another strategy Adversarial Semantic Pattern Minimization, which removes most perturbations and retains only essential adversarial patterns in AEsBased on the two strategies, we design our method Robust Feature Coverage Attack (RFCoA), comprising Robust Feature Disentanglement and Adversarial Feature Fusion. In the first stage, we extract target class RFs in feature space. In the second stage, we use attention-based feature fusion to overlay these RFs onto predictive features of clean images and remove unnecessary perturbations. Experiments show our method's superior transferability, robustness, and stealthiness compared to existing state-of-the-art methods. Additionally, our method's effectiveness can extend to Large Vision-Language Models (LVLMs), indicating its potential applicability to more complex tasks., Comment: Accepted by AAAI2025

With the advancement of deep learning, object detectors (ODs) with various architectures have achieved significant success in complex scenarios like autonomous driving. Previous adversarial attacks against ODs have been focused on designing customized attacks targeting their specific structures (e.g., NMS and RPN), yielding some results but simultaneously constraining their scalability. Moreover, most efforts against ODs stem from image-level attacks originally designed for classification tasks, resulting in redundant computations and disturbances in object-irrelevant areas (e.g., background). Consequently, how to design a model-agnostic efficient attack to comprehensively evaluate the vulnerabilities of ODs remains challenging and unresolved. In this paper, we propose NumbOD, a brand-new spatial-frequency fusion attack against various ODs, aimed at disrupting object detection within images. We directly leverage the features output by the OD without relying on its internal structures to craft adversarial examples. Specifically, we first design a dual-track attack target selection strategy to select high-quality bounding boxes from OD outputs for targeting. Subsequently, we employ directional perturbations to shift and compress predicted boxes and change classification results to deceive ODs. Additionally, we focus on manipulating the high-frequency components of images to confuse ODs' attention on critical objects, thereby enhancing the attack efficiency. Our extensive experiments on nine ODs and two datasets show that NumbOD achieves powerful attack performance and high stealthiness., Comment: Accepted by AAAI 2025

With the rapid advancement of deep learning, the model robustness has become a significant research hotspot, \ie, adversarial attacks on deep neural networks. Existing works primarily focus on image classification tasks, aiming to alter the model's predicted labels. Due to the output complexity and deeper network architectures, research on adversarial examples for segmentation models is still limited, particularly for universal adversarial perturbations. In this paper, we propose a novel universal adversarial attack method designed for segmentation models, which includes dual feature separation and low-frequency scattering modules. The two modules guide the training of adversarial examples in the pixel and frequency space, respectively. Experiments demonstrate that our method achieves high attack success rates surpassing the state-of-the-art methods, and exhibits strong transferability across different models., Comment: Accepted by ICASSP 2025

Reasoning future unknowable facts on temporal knowledge graphs (TKGs) is a challenging task, holding significant academic and practical values for various fields. Existing studies exploring explainable reasoning concentrate on modeling comprehensible temporal paths relevant to the query. Yet, these path-based methods primarily focus on local temporal paths appearing in recent times, failing to capture the complex temporal paths in TKG and resulting in the loss of longer historical relations related to the query. Motivated by the Dual Process Theory in cognitive science, we propose a \textbf{Cogn}itive \textbf{T}emporal \textbf{K}nowledge \textbf{E}xtrapolation framework (CognTKE), which introduces a novel temporal cognitive relation directed graph (TCR-Digraph) and performs interpretable global shallow reasoning and local deep reasoning over the TCR-Digraph. Specifically, the proposed TCR-Digraph is constituted by retrieving significant local and global historical temporal relation paths associated with the query. In addition, CognTKE presents the global shallow reasoner and the local deep reasoner to perform global one-hop temporal relation reasoning (System 1) and local complex multi-hop path reasoning (System 2) over the TCR-Digraph, respectively. The experimental results on four benchmark datasets demonstrate that CognTKE achieves significant improvement in accuracy compared to the state-of-the-art baselines and delivers excellent zero-shot reasoning ability. \textit{The code is available at https://github.com/WeiChen3690/CognTKE}., Comment: AAAI2025 Accept, 12 pages, 9 figures

The takeaway recommendation system aims to recommend users' future takeaway purchases based on their historical purchase behaviors, thereby improving user satisfaction and boosting merchant sales. Existing methods focus on incorporating auxiliary information or leveraging knowledge graphs to alleviate the sparsity issue of user purchase sequences. However, two main challenges limit the performance of these approaches: (1) capturing dynamic user preferences on complex geospatial information and (2) efficiently integrating spatial-temporal knowledge from both graphs and sequence data with low computational costs. In this paper, we propose a novel spatial-temporal knowledge distillation model for takeaway recommendation (STKDRec) based on the two-stage training process. Specifically, during the first pre-training stage, a spatial-temporal knowledge graph (STKG) encoder is trained to extract high-order spatial-temporal dependencies and collaborative associations from the STKG. During the second spatial-temporal knowledge distillation (STKD) stage, a spatial-temporal Transformer (ST-Transformer) is employed to comprehensively model dynamic user preferences on various types of fine-grained geospatial information from a sequential perspective. Furthermore, the STKD strategy is introduced to transfer graph-based spatial-temporal knowledge to the ST-Transformer, facilitating the adaptive fusion of rich knowledge derived from both the STKG and sequence data while reducing computational overhead. Extensive experiments on three real-world datasets show that STKDRec significantly outperforms the state-of-the-art baselines., Comment: Accepted by AAAI2025

Lung diseases represent a significant global health challenge, with Chest X-Ray (CXR) being a key diagnostic tool due to their accessibility and affordability. Nonetheless, the detection of pulmonary lesions is often hindered by overlapping bone structures in CXR images, leading to potential misdiagnoses. To address this issue, we developed an end-to-end framework called BS-LDM, designed to effectively suppress bone in high-resolution CXR images. This framework is based on conditional latent diffusion models and incorporates a multi-level hybrid loss-constrained vector-quantized generative adversarial network which is crafted for perceptual compression, ensuring the preservation of details. To further enhance the framework's performance, we introduce offset noise and a temporal adaptive thresholding strategy. These additions help minimize discrepancies in generating low-frequency information, thereby improving the clarity of the generated soft tissue images. Additionally, we have compiled a high-quality bone suppression dataset named SZCH-X-Rays. This dataset includes 818 pairs of high-resolution CXR and dual-energy subtraction soft tissue images collected from a partner hospital. Moreover, we processed 241 data pairs from the JSRT dataset into negative images, which are more commonly used in clinical practice. Our comprehensive experimental and clinical evaluations reveal that BS-LDM excels in bone suppression, underscoring its significant clinical value., Comment: 10 pages, 7 figures

Compositional understanding allows visual language models to interpret complex relationships between objects, attributes, and relations in images and text. However, most existing methods often rely on hard negative examples and fine-tuning, which can overestimate improvements and are limited by the difficulty of obtaining hard negatives. In this work, we introduce Zero-Shot Compositional Understanding (ZS-CU), a novel task that enhances compositional understanding without requiring hard negative training data. We propose YUKINO (Yielded Compositional Understanding Knowledge via Textual Inversion with NO), which uses textual inversion to map unlabeled images to pseudo-tokens in a pre-trained CLIP model. We propose introducing "no" logical regularization to address the issue of token interaction in inversion. Additionally, we suggest using knowledge distillation to reduce the time complexity of textual inversion. Experimental results show that YUKINO outperforms the existing multi-modal SOTA models by over 8% on the SugarCREPE benchmark, and also achieves significant improvements in image retrieval tasks.

As Large Language Models become integral to decision-making, optimism about their power is tempered with concern over their errors. Users may over-rely on LLM advice that is confidently stated but wrong, or under-rely due to mistrust. Reliance interventions have been developed to help users of LLMs, but they lack rigorous evaluation for appropriate reliance. We benchmark the performance of three relevant interventions by conducting a randomized online experiment with 400 participants attempting two challenging tasks: LSAT logical reasoning and image-based numerical estimation. For each question, participants first answered independently, then received LLM advice modified by one of three reliance interventions and answered the question again. Our findings indicate that while interventions reduce over-reliance, they generally fail to improve appropriate reliance. Furthermore, people became more confident after making incorrect reliance decisions in certain contexts, demonstrating poor calibration. Based on our findings, we discuss implications for designing effective reliance interventions in human-LLM collaboration.

Natural organisms can convert environmental stimuli into sensory feedback to regulate their body and realize active adaptivity. However, realizing such a feedback-regulation mechanism in synthetic material systems remains a grand challenge. It is believed that achieving complex feedback mechanisms in responsive materials will pave the way toward autonomous, intelligent structure and actuation without complex electronics. Inspired by living systems, we report a general principle to design and construct such feedback loops in light-responsive materials. Specifically, we design a baffle-actuator mechanism to incorporate programmed feedback into the opto-mechanical responsiveness. By simply addressing the baffle position with respect to the incident light beam, positive and negative feedback are programmed. We demonstrate the transformation of a light-bending strip into a switcher, where the intensity of light determines the energy barrier under positive feedback, realizing multi-stable shape-morphing. By leveraging the negative feedback and associated homeostasis, we demonstrate two soft robots, i.e., a locomotor and a swimmer. Furthermore, we unveil the ubiquity of feedback in light-responsive materials, which provides new insight into self-regulated robotic matters.

Assembling increasingly larger-scale defect-free optical tweezer-trapped atom arrays is essential for quantum computation and quantum simulations based on atoms. Here, we propose an AI-enabled, rapid, constant-time-overhead rearrangement protocol, and we experimentally assemble defect-free 2D and 3D atom arrays with up to 2024 atoms with a constant time cost of 60 ms. The AI model calculates the holograms for real-time atom rearrangement. With precise controls over both position and phase, a high-speed spatial light modulator moves all the atoms simultaneously. This protocol can be readily used to generate defect-free arrays of tens of thousands of atoms with current technologies, and become a useful toolbox for quantum error correction.

Hateful meme detection aims to prevent the proliferation of hateful memes on various social media platforms. Considering its impact on social environments, this paper introduces a previously ignored but significant threat to hateful meme detection: backdoor attacks. By injecting specific triggers into meme samples, backdoor attackers can manipulate the detector to output their desired outcomes. To explore this, we propose the Meme Trojan framework to initiate backdoor attacks on hateful meme detection. Meme Trojan involves creating a novel Cross-Modal Trigger (CMT) and a learnable trigger augmentor to enhance the trigger pattern according to each input sample. Due to the cross-modal property, the proposed CMT can effectively initiate backdoor attacks on hateful meme detectors under an automatic application scenario. Additionally, the injection position and size of our triggers are adaptive to the texts contained in the meme, which ensures that the trigger is seamlessly integrated with the meme content. Our approach outperforms the state-of-the-art backdoor attack methods, showing significant improvements in effectiveness and stealthiness. We believe that this paper will draw more attention to the potential threat posed by backdoor attacks on hateful meme detection., Comment: Accepted by AAAI25

Multi-page websites dominate modern web development. However, existing design-to-code methods rely on simplified assumptions, limiting to single-page, self-contained webpages without external resource connection. To address this gap, we introduce the Multi-Page Resource-Aware Webpage (MRWeb) generation task, which transforms UI designs into multi-page, functional web UIs with internal/external navigation, image loading, and backend routing. We propose a novel resource list data structure to track resources, links, and design components. Our study applies existing methods to the MRWeb problem using a newly curated dataset of 500 websites (300 synthetic, 200 real-world). Specifically, we identify the best metric to evaluate the similarity of the web UI, assess the impact of the resource list on MRWeb generation, analyze MLLM limitations, and evaluate the effectiveness of the MRWeb tool in real-world workflows. The results show that resource lists boost navigation functionality from 0% to 66%-80% while facilitating visual similarity. Our proposed metrics and evaluation framework provide new insights into MLLM performance on MRWeb tasks. We release the MRWeb tool, dataset, and evaluation framework to promote further research.

We present improvements to the pointing accuracy of the South Pole Telescope (SPT) using machine learning. The ability of the SPT to point accurately at the sky is limited by its structural imperfections, which are impacted by the extreme weather at the South Pole. Pointing accuracy is particularly important during SPT participation in observing campaigns with the Event Horizon Telescope (EHT), which requires stricter accuracy than typical observations with the SPT. We compile a training dataset of historical observations of astronomical sources made with the SPT-3G and EHT receivers on the SPT. We train two XGBoost models to learn a mapping from current weather conditions to two telescope drive control arguments -- one which corrects for errors in azimuth and the other for errors in elevation. Our trained models achieve root mean squared errors on withheld test data of $2.14''$ in cross-elevation and $3.57''$ in elevation, well below our goal of $5''$ along each axis. We deploy our models on the telescope control system and perform further in situ test observations during the EHT observing campaign in 2024 April. Our models result in significantly improved pointing accuracy: for sources within the range of input variables where the models are best trained, average combined pointing error improved 33%, from $15.9''$ to $10.6''$. These improvements, while significant, fall shy of our ultimate goal, but they serve as a proof of concept for the development of future models. Planned upgrades to the EHT receiver on the SPT will necessitate even stricter pointing accuracy which will be achievable with our methods., Comment: 23 pages, 9 figures, to be submitted to Journal of Astronomical Instrumentation (JAI)

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

In this report, we introduce Qwen2.5, a comprehensive series of large language models (LLMs) designed to meet diverse needs. Compared to previous iterations, Qwen 2.5 has been significantly improved during both the pre-training and post-training stages. In terms of pre-training, we have scaled the high-quality pre-training datasets from the previous 7 trillion tokens to 18 trillion tokens. This provides a strong foundation for common sense, expert knowledge, and reasoning capabilities. In terms of post-training, we implement intricate supervised finetuning with over 1 million samples, as well as multistage reinforcement learning. Post-training techniques enhance human preference, and notably improve long text generation, structural data analysis, and instruction following. To handle diverse and varied use cases effectively, we present Qwen2.5 LLM series in rich sizes. Open-weight offerings include base and instruction-tuned models, with quantized versions available. In addition, for hosted solutions, the proprietary models currently include two mixture-of-experts (MoE) variants: Qwen2.5-Turbo and Qwen2.5-Plus, both available from Alibaba Cloud Model Studio. Qwen2.5 has demonstrated top-tier performance on a wide range of benchmarks evaluating language understanding, reasoning, mathematics, coding, human preference alignment, etc. Specifically, the open-weight flagship Qwen2.5-72B-Instruct outperforms a number of open and proprietary models and demonstrates competitive performance to the state-of-the-art open-weight model, Llama-3-405B-Instruct, which is around 5 times larger. Qwen2.5-Turbo and Qwen2.5-Plus offer superior cost-effectiveness while performing competitively against GPT-4o-mini and GPT-4o respectively. Additionally, as the foundation, Qwen2.5 models have been instrumental in training specialized models such as Qwen2.5-Math, Qwen2.5-Coder, QwQ, and multimodal models.

In this paper, we solve unbalanced optimal transport (UOT) problem on surfaces represented by point clouds. Based on alternating direction method of multipliers algorithm, the original UOT problem can be solved by an iteration consists of three steps. The key ingredient is to solve a Poisson equation on point cloud which is solved by tangent radial basis function (TRBF) method. The proposed TRBF method requires only the point cloud and normal vectors to discretize the Poisson equation which simplify the computation significantly. Numerical experiments conducted on point clouds with varying geometry and topology demonstrate the effectiveness of the proposed method.