Search

Biometrics authentication has become increasingly popular due to its security and convenience; however, traditional biometrics are becoming less desirable in scenarios such as new mobile devices, Virtual Reality, and Smart Vehicles. For example, while face authentication is widely used, it suffers from significant privacy concerns. The collection of complete facial data makes it less desirable for privacy-sensitive applications. Lip authentication, on the other hand, has emerged as a promising biometrics method. However, existing lip-based authentication methods heavily depend on static lip shape when the mouth is closed, which can be less robust due to lip shape dynamic motion and can barely work when the user is speaking. In this paper, we revisit the nature of lip biometrics and extract shape-independent features from the lips. We study the dynamic characteristics of lip biometrics based on articulator motion. Building on the knowledge, we propose a system for shape-independent continuous authentication via lip articulator dynamics. This system enables robust, shape-independent and continuous authentication, making it particularly suitable for scenarios with high security and privacy requirements. We conducted comprehensive experiments in different environments and attack scenarios and collected a dataset of 50 subjects. The results indicate that our system achieves an overall accuracy of 99.06% and demonstrates robustness under advanced mimic attacks and AI deepfake attacks, making it a viable solution for continuous biometric authentication in various applications.

Reasoning is fundamental to human intelligence, and critical for problem-solving, decision-making, and critical thinking. Reasoning refers to drawing new conclusions based on existing knowledge, which can support various applications like clinical diagnosis, basic education, and financial analysis. Though a good number of surveys have been proposed for reviewing reasoning-related methods, none of them has systematically investigated these methods from the viewpoint of their dependent knowledge base. Both the scenarios to which the knowledge bases are applied and their storage formats are significantly different. Hence, investigating reasoning methods from the knowledge base perspective helps us better understand the challenges and future directions. To fill this gap, this paper first classifies the knowledge base into symbolic and parametric ones. The former explicitly stores information in human-readable symbols, and the latter implicitly encodes knowledge within parameters. Then, we provide a comprehensive overview of reasoning methods using symbolic knowledge bases, parametric knowledge bases, and both of them. Finally, we identify the future direction toward enhancing reasoning capabilities to bridge the gap between human and machine intelligence.

Star formation quenching in galaxies is a critical process in galaxy formation. It is widely believed that the quenching process is dominated by the mass of galaxies and/or their environment. In Paper V, we addressed the challenge to disentangle the effects of mass and environment by employing the PAC method, which combines spectroscopic and deep photometric surveys. This approach enabled us to measure the excess surface density of blue and red galaxies around massive central galaxies down to $10^{9.0}M_{\odot}$. However, it is not straightforward to completely separate the two effects.To address this issue, in this paper, we derive the average quenched fraction of central (isolated) galaxies, $\bar{f}_{\mathrm{q}}^{\mathrm{cen}}(M_{*})$, by combining the 3D quenched fraction distribution $f^{\mathrm{sat}}_{\mathrm{q}}(r; M_{*,\mathrm{cen}}, M_{*,\mathrm{sat}})$, reconstructed from the $\bar{n}_2w_{\mathrm{p}}(r_{\mathrm{p}})$ measurements, with the stellar mass-halo mass relation in N-body simulations from Paper IV, and the observed total quenched fraction, $\bar{f}_{\mathrm{q}}^{\mathrm{all}}(M_{*})$. Using $f^{\mathrm{sat}}_{\mathrm{q}}(r;M_{*,\mathrm{cen}},M_{*,\mathrm{sat}})$, $\bar{f}_{\mathrm{q}}^{\mathrm{cen}}(M_{*})$, and the galaxy-halo connection, we assign a quenched probability to each (sub)halo in the simulation, enabling a comprehensive study of galaxy quenching. We find that the mass-quenched fraction increases from 0.3 to 0.87 across the stellar mass range $[10^{9.5}, 10^{11.0}]M_{\odot}$, while the environmental quenched fraction decreases from 0.17 to 0.03. The mass effect dominates galaxy quenching across the entire stellar mass range we studied. Moreover, more massive host halos are more effective at quenching their satellite galaxies, while satellite stellar mass has minimal influence on environmental quenching., Comment: Submitted to ApJ. Abstract abridged for arxiv. Comments welcome!

While many sophisticated exploration methods have been proposed, their lack of generality and high computational cost often lead researchers to favor simpler methods like $\epsilon$-greedy. Motivated by this, we introduce $\beta$-DQN, a simple and efficient exploration method that augments the standard DQN with a behavior function $\beta$. This function estimates the probability that each action has been taken at each state. By leveraging $\beta$, we generate a population of diverse policies that balance exploration between state-action coverage and overestimation bias correction. An adaptive meta-controller is designed to select an effective policy for each episode, enabling flexible and explainable exploration. $\beta$-DQN is straightforward to implement and adds minimal computational overhead to the standard DQN. Experiments on both simple and challenging exploration domains show that $\beta$-DQN outperforms existing baseline methods across a wide range of tasks, providing an effective solution for improving exploration in deep reinforcement learning.

A generic closed quantum many-body system will inevitably tend to thermalization, whose local information encoded in the initial state eventually scrambles into the full space, known as quantum ergodicity. A paradigmatic exception in closed quantum systems for strong ergodicity breaking is known as many-body localization, where strong disorder-induced localization prevents the occurrence of thermalization. It is generally recognized that a localized quantum system would be delocalized under dissipation induced by the environment. However, this consequence recently has received challenges where an exotic dissipation-induced localization mechanism is proposed, and transitions between localized and extended phases are found. In this Letter, we promote this mechanism to systems for weak ergodicity breaking hosting quantum many-body scars (QMBS). We find that the system relaxes to a steady state dominated by QMBS, and the dissipative dynamics exhibit dynamic revivals by suitably preparing an initial state. We point out an experimental realization of the controlled dissipation with a cold atomic setup. This makes the signature of ergodicity breaking visible over dissipative dynamics and offers potential possibilities for experimentally preparing stable QMBS with associated coherent dynamics., Comment: 7+1 pages, 4+1 figures. Comments are welcome

Semantic communication (SemCom) is regarded as a promising and revolutionary technology in 6G, aiming to transcend the constraints of ``Shannon's trap" by filtering out redundant information and extracting the core of effective data. Compared to traditional communication paradigms, SemCom offers several notable advantages, such as reducing the burden on data transmission, enhancing network management efficiency, and optimizing resource allocation. Numerous researchers have extensively explored SemCom from various perspectives, including network architecture, theoretical analysis, potential technologies, and future applications. However, as SemCom continues to evolve, a multitude of security and privacy concerns have arisen, posing threats to the confidentiality, integrity, and availability of SemCom systems. This paper presents a comprehensive survey of the technologies that can be utilized to secure SemCom. Firstly, we elaborate on the entire life cycle of SemCom, which includes the model training, model transfer, and semantic information transmission phases. Then, we identify the security and privacy issues that emerge during these three stages. Furthermore, we summarize the techniques available to mitigate these security and privacy threats, including data cleaning, robust learning, defensive strategies against backdoor attacks, adversarial training, differential privacy, cryptography, blockchain technology, model compression, and physical-layer security. Lastly, this paper outlines future research directions to guide researchers in related fields., Comment: 123 pages, 27 figures

Cryptocurrency investment is inherently difficult due to its shorter history compared to traditional assets, the need to integrate vast amounts of data from various modalities, and the requirement for complex reasoning. While deep learning approaches have been applied to address these challenges, their black-box nature raises concerns about trust and explainability. Recently, large language models (LLMs) have shown promise in financial applications due to their ability to understand multi-modal data and generate explainable decisions. However, single LLM faces limitations in complex, comprehensive tasks such as asset investment. These limitations are even more pronounced in cryptocurrency investment, where LLMs have less domain-specific knowledge in their training corpora. To overcome these challenges, we propose an explainable, multi-modal, multi-agent framework for cryptocurrency investment. Our framework uses specialized agents that collaborate within and across teams to handle subtasks such as data analysis, literature integration, and investment decision-making for the top 30 cryptocurrencies by market capitalization. The expert training module fine-tunes agents using multi-modal historical data and professional investment literature, while the multi-agent investment module employs real-time data to make informed cryptocurrency investment decisions. Unique intrateam and interteam collaboration mechanisms enhance prediction accuracy by adjusting final predictions based on confidence levels within agent teams and facilitating information sharing between teams. Empirical evaluation using data from November 2023 to September 2024 demonstrates that our framework outperforms single-agent models and market benchmarks in classification, asset pricing, portfolio, and explainability performance.

Analyzing long-term behaviors in high-dimensional nonlinear dynamical systems remains a significant challenge. The Koopman operator framework has emerged as a powerful tool to address this issue by providing a globally linear perspective on nonlinear dynamics. However, existing methods for approximating the Koopman operator and its spectral components, particularly in large-scale systems, often lack robust theoretical guarantees. Residual Dynamic Mode Decomposition (ResDMD) introduces a spectral residual measure to assess the convergence of the estimated Koopman spectrum, which helps filter out spurious spectral components. Nevertheless, it depends on pre-computed spectra, thereby inheriting their inaccuracies. To overcome its limitations, we introduce the Neural Network-ResDMD (NN-ResDMD), a method that directly estimates Koopman spectral components by minimizing the spectral residual. By leveraging neural networks, NN-ResDMD automatically identifies the optimal basis functions of the Koopman invariant subspace, eliminating the need for manual selection and improving the reliability of the analysis. Experiments on physical and biological systems demonstrate that NN-ResDMD significantly improves both accuracy and scalability, making it an effective tool for analyzing complex dynamical systems.

Despite very fast progress, efficiently training large language models (LLMs) in very long contexts remains challenging. Existing methods fall back to training LLMs with short contexts (a maximum of a few thousands tokens in training) and use inference time techniques when evaluating on long contexts (above 1M tokens context window at inference). As opposed to long-context-inference, training on very long context input prompts is quickly limited by GPU memory availability and by the prohibitively long training times it requires on state-of-the-art hardware. Meanwhile, many real-life applications require not only inference but also training/fine-tuning with long context on specific tasks. Such applications include, for example, augmenting the context with various sources of raw reference information for fact extraction, fact summarization, or fact reconciliation tasks. We propose adjoint sharding, a novel technique that comprises sharding gradient calculation during training to reduce memory requirements by orders of magnitude, making training on very long context computationally tractable. Adjoint sharding is based on the adjoint method and computes equivalent gradients to backpropagation. We also propose truncated adjoint sharding to speed up the algorithm while maintaining performance. We provide a distributed version, and a paralleled version of adjoint sharding to further speed up training. Empirical results show the proposed adjoint sharding algorithm reduces memory usage by up to 3X with a 1.27B parameter large language model on 1M context length training. This allows to increase the maximum context length during training or fine-tuning of a 1.27B parameter model from 35K tokens to above 100K tokens on a training infrastructure composed of five AWS P4 instances.

The rise of End-Edge-Cloud Collaboration (EECC) offers a promising paradigm for Artificial Intelligence (AI) model training across end devices, edge servers, and cloud data centers, providing enhanced reliability and reduced latency. Hierarchical Federated Learning (HFL) can benefit from this paradigm by enabling multi-tier model aggregation across distributed computing nodes. However, the potential of HFL is significantly constrained by the inherent heterogeneity and dynamic characteristics of EECC environments. Specifically, the uniform model structure bounded by the least powerful end device across all computing nodes imposes a performance bottleneck. Meanwhile, coupled heterogeneity in data distributions and resource capabilities across tiers disrupts hierarchical knowledge transfer, leading to biased updates and degraded performance. Furthermore, the mobility and fluctuating connectivity of computing nodes in EECC environments introduce complexities in dynamic node migration, further compromising the robustness of the training process. To address multiple challenges within a unified framework, we propose End-Edge-Cloud Federated Learning with Self-Rectified Knowledge Agglomeration (FedEEC), which is a novel EECC-empowered FL framework that allows the trained models from end, edge, to cloud to grow larger in size and stronger in generalization ability. FedEEC introduces two key innovations: (1) Bridge Sample Based Online Distillation Protocol (BSBODP), which enables knowledge transfer between neighboring nodes through generated bridge samples, and (2) Self-Knowledge Rectification (SKR), which refines the transferred knowledge to prevent suboptimal cloud model optimization. The proposed framework effectively handles both cross-tier resource heterogeneity and effective knowledge transfer between neighboring nodes, while satisfying the migration-resilient requirements of EECC., Comment: 16 pages, 7 tables, 6 figures. arXiv admin note: text overlap with arXiv:2312.11489

Significance: Changes in the structure of tissue occur in many disease processes, such as the boundaries of cancerous tumors and burn injuries. Spectroscopic and polarimetric alterations of terahertz light caused by Mie scattering patterns has the potential to be a diagnostic marker. Aim: We present an analysis of Monte Carlo simulation of Mie scattering of polarized terahertz light from cancerous tumor budding, compare the simulation to experimental results obtained in phantom models, and present an analysis of a polarization-sensitive terahertz scan of an ex vivo porcine burn injury. Approach: Using a Monte Carlo simulation, we modeled the changes in diffuse intensity and degree of polarization of broadband off-specular terahertz light due to scattering particles in highly attenuating tissue. We extracted the Mueller matrix of the tissue using this model and analyzed the Lu-Chipman product decomposition matrices. We compared this model to experimental data from four phantoms consisting of polypropylene particles of varying sizes embedded in gelatin. Finally, we induced a full-thickness burn injury in ex vivo porcine skin samples and compared experimental data from burned and healthy regions of the tissue. Results: Simulation revealed contrast in the Stokes vectors and Mueller Matrix elements for varying scattering particle sizes. Experimental phantom results showed contrast between different sizes of scattering particles in degree of polarization and diffuse intensity in agreement with Monte Carlo simulation results. Finally, we demonstrated a similar diffused imaging signal contrast between burned and healthy regions of ex vivo porcine skin. Conclusion: Polarimetric terahertz imaging has the potential to detect structural changes due to biological disease processes.

Recently released open-source pre-trained foundational image segmentation and object detection models (SAM2+GroundingDINO) allow for geometrically consistent segmentation of objects of interest in multi-view 2D images. Users can use text-based or click-based prompts to segment objects of interest without requiring labeled training datasets. Gaussian Splatting allows for the learning of the 3D representation of a scene's geometry and radiance based on 2D images. Combining Google Earth Studio, SAM2+GroundingDINO, 2D Gaussian Splatting, and our improvements in mask refinement based on morphological operations and contour simplification, we created a pipeline to extract the 3D mesh of any building based on its name, address, or geographic coordinates.

Diffusion models have shown exceptional performance in visual generation tasks. Recently, these models have shifted from traditional U-Shaped CNN-Attention hybrid structures to fully transformer-based isotropic architectures. While these transformers exhibit strong scalability and performance, their reliance on complicated self-attention operation results in slow inference speeds. Contrary to these works, we rethink one of the simplest yet fastest module in deep learning, 3x3 Convolution, to construct a scaled-up purely convolutional diffusion model. We first discover that an Encoder-Decoder Hourglass design outperforms scalable isotropic architectures for Conv3x3, but still under-performing our expectation. Further improving the architecture, we introduce sparse skip connections to reduce redundancy and improve scalability. Based on the architecture, we introduce conditioning improvements including stage-specific embeddings, mid-block condition injection, and conditional gating. These improvements lead to our proposed Diffusion CNN (DiC), which serves as a swift yet competitive diffusion architecture baseline. Experiments on various scales and settings show that DiC surpasses existing diffusion transformers by considerable margins in terms of performance while keeping a good speed advantage. Project page: https://github.com/YuchuanTian/DiC, Comment: 11 pages, 6 figures

We present STORM, a spatio-temporal reconstruction model designed for reconstructing dynamic outdoor scenes from sparse observations. Existing dynamic reconstruction methods often rely on per-scene optimization, dense observations across space and time, and strong motion supervision, resulting in lengthy optimization times, limited generalization to novel views or scenes, and degenerated quality caused by noisy pseudo-labels for dynamics. To address these challenges, STORM leverages a data-driven Transformer architecture that directly infers dynamic 3D scene representations--parameterized by 3D Gaussians and their velocities--in a single forward pass. Our key design is to aggregate 3D Gaussians from all frames using self-supervised scene flows, transforming them to the target timestep to enable complete (i.e., "amodal") reconstructions from arbitrary viewpoints at any moment in time. As an emergent property, STORM automatically captures dynamic instances and generates high-quality masks using only reconstruction losses. Extensive experiments on public datasets show that STORM achieves precise dynamic scene reconstruction, surpassing state-of-the-art per-scene optimization methods (+4.3 to 6.6 PSNR) and existing feed-forward approaches (+2.1 to 4.7 PSNR) in dynamic regions. STORM reconstructs large-scale outdoor scenes in 200ms, supports real-time rendering, and outperforms competitors in scene flow estimation, improving 3D EPE by 0.422m and Acc5 by 28.02%. Beyond reconstruction, we showcase four additional applications of our model, illustrating the potential of self-supervised learning for broader dynamic scene understanding., Comment: Project page at: https://jiawei-yang.github.io/STORM/

Synthesizing photo-realistic visual observations from an ego vehicle's driving trajectory is a critical step towards scalable training of self-driving models. Reconstruction-based methods create 3D scenes from driving logs and synthesize geometry-consistent driving videos through neural rendering, but their dependence on costly object annotations limits their ability to generalize to in-the-wild driving scenarios. On the other hand, generative models can synthesize action-conditioned driving videos in a more generalizable way but often struggle with maintaining 3D visual consistency. In this paper, we present DreamDrive, a 4D spatial-temporal scene generation approach that combines the merits of generation and reconstruction, to synthesize generalizable 4D driving scenes and dynamic driving videos with 3D consistency. Specifically, we leverage the generative power of video diffusion models to synthesize a sequence of visual references and further elevate them to 4D with a novel hybrid Gaussian representation. Given a driving trajectory, we then render 3D-consistent driving videos via Gaussian splatting. The use of generative priors allows our method to produce high-quality 4D scenes from in-the-wild driving data, while neural rendering ensures 3D-consistent video generation from the 4D scenes. Extensive experiments on nuScenes and street view images demonstrate that DreamDrive can generate controllable and generalizable 4D driving scenes, synthesize novel views of driving videos with high fidelity and 3D consistency, decompose static and dynamic elements in a self-supervised manner, and enhance perception and planning tasks for autonomous driving.

Surgery video segmentation is an important topic in the surgical AI field. It allows the AI model to understand the spatial information of a surgical scene. Meanwhile, due to the lack of annotated surgical data, surgery segmentation models suffer from limited performance. With the emergence of SAM2 model, a large foundation model for video segmentation trained on natural videos, zero-shot surgical video segmentation became more realistic but meanwhile remains to be explored. In this paper, we systematically evaluate the performance of SAM2 model in zero-shot surgery video segmentation task. We conducted experiments under different configurations, including different prompting strategies, robustness, etc. Moreover, we conducted an empirical evaluation over the performance, including 9 datasets with 17 different types of surgeries.

The ability of perceiving fine-grained spatial and temporal information is crucial for video-language retrieval. However, the existing video retrieval benchmarks, such as MSRVTT and MSVD, fail to efficiently evaluate the fine-grained retrieval ability of video-language models (VLMs) due to a lack of detailed annotations. To address this problem, we present FIBER, a FIne-grained BEnchmark for text to video Retrieval, containing 1,000 videos sourced from the FineAction dataset. Uniquely, our FIBER benchmark provides detailed human-annotated spatial annotations and temporal annotations for each video, making it possible to independently evaluate the spatial and temporal bias of VLMs on video retrieval task. Besides, we employ a text embedding method to unlock the capability of fine-grained video-language understanding of Multimodal Large Language Models (MLLMs). Surprisingly, the experiment results show that our Video Large Language Encoder (VLLE) performs comparably to CLIP-based models on traditional benchmarks and has a stronger capability of fine-grained representation with lower spatial-temporal bias. Project page: https://fiber-bench.github.io.

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/.

In this paper, we will characterize the structure of the negative part of the Zariski decomposition of $K_{\cF}+D$ for a foliated surface $(X,\cF)$ associated with a $\bQ$-divisor $D$ whenever $K_{\cF}+D$ is pseudoeffective. As applications, we will study the effective behavior of the multiple linear system $|m(K_{\cF}+D)|$ for any sufficiently divisible integer $m>0$.

Loop I/North Polar Spur (NPS) is the giant arc structure above the Galactic plane observed in the radio sky. It is the most conspicuous feature in low frequency radio sky maps besides the galactic plane itself. There is a long-standing debate about its origin. While the majority consider it as a nearby supernova remnant (SNR), it has also been suggested to be a giant bubble close to the Galactic Center (GC), associated with the Fermi Bubble and eROSITA X-ray bubble. There is also the possibility that a nearby SNR and a bubble near the GC happens to overlay each other. At ultralong wavelength band (wavelength $\gtrsim 10$ m or frequency $\lesssim 30$ MHz), particularly below $\sim 10$ MHz, the free-free absorption of radio signal by the diffuse electrons in interstellar medium (ISM) becomes significant, resulting in sky morphology differs largely from higher frequencies. In this paper, we predict the Loop I/NPS morphology at ultralong wavelength band. We develop emissivity models for the two Loop I/NPS origin models. We find that, at ultralong wavelength band, for the SNR model, the full Loop I/NPS is still a bright arc even at frequency as low as $\sim 1$ MHz; however, in the GC model, the Loop I/NPS appears only at $b\gtrsim 30\degree$, at $b\lesssim 30 \degree$ the Loop I/NPS is invisible due to the absorption by ISM electrons between the GC and the Sun. Upcoming ultralong wavelentgh projects such as DSL and FARSIDE can potentially distinguish these two models and provide decisive information about the origin of Loop I/NPS., Comment: 11 pages,5 figures

In this study, the spin-orbit torque (SOT) in light metal oxide systems is investigated using an experimental approach based on harmonic Hall voltage techniques in out-of-plane (OOP) angular geometry for samples with in-plane magnetic anisotropy. In parallel, an analytical derivation of this alternative OOP harmonic Hall detection geometry has been developed, followed by experimental validation to extract SOT effective fields. In addition, to accurately quantifying SOT, this method allows complete characterization of thermoelectric effects, opening promising avenues for accurate SOT characterization in related systems. In particular, this study corroborates the critical role of naturally oxidized copper interfaced with metallic Cu in the generation of orbital current in Co(2)|Pt(4)|CuOx(3), demonstrating a two-fold increase in damping-like torques compared to a reference sample with an oxidized Al capping layer. These findings offer promising directions for future research on the application aspect of non-equilibrium orbital angular momentum., Comment: 15 pages, 4 figures, 44 references

In this work, we update the unsupervised machine learning (UML) step by proposing an algorithm based on ConvNeXt large model coding to improve the efficiency of unlabeled galaxy morphology classifications. The method can be summarized into three key aspects as follows: (1) a convolutional autoencoder is used for image denoising and reconstruction and the rotational invariance of the model is improved by polar coordinate extension; (2) utilizing a pre-trained convolutional neural network (CNN) named ConvNeXt for encoding the image data. The features were further compressed via a principal component analysis (PCA) dimensionality reduction; (3) adopting a bagging-based multi-model voting classification algorithm to enhance robustness. We applied this model to I-band images of a galaxy sample with $I_{\rm mag}< 25$ in the COSMOS field. Compared to the original unsupervised method, the number of clustering groups required by the new method is reduced from 100 to 20. Finally, we managed to classify about 53\% galaxies, significantly improving the classification efficiency. To verify the validity of the morphological classification, we selected massive galaxies with $M(*)>10^{10}(M(sun))$ for morphological parameter tests. The corresponding rules between the classification results and the physical properties of galaxies on multiple parameter surfaces are consistent with the existing evolution model. Our method has demonstrated the feasibility of using large model encoding to classify galaxy morphology, which not only improves the efficiency of galaxy morphology classification, but also saves time and manpower. Furthermore, in comparison to the original UML model, the enhanced classification performance is more evident in qualitative analysis and has successfully surpassed a greater number of parameter tests., Comment: Accepted by A&A; 12 pages, 12 figures

Privacy computing receives increasing attention but writing privacy computing code remains challenging for developers due to limited library functions that necessitate extensive function implementation from scratch as well as the data-oblivious requirement which contradicts intuitive thinking and usual practices of programmers. Large language models (LLMs) have demonstrated surprising capabilities in coding tasks and achieved state-of-the-art performance across many benchmarks. However, even with extensive prompting, existing LLMs struggle with code translation task for privacy computing, such as translating Python to MP-SPDZ, due to the scarcity of MP-SPDZ data required for effective pre-training or fine-tuning. To address the limitation, this paper proposes SPDZCoder, a rule-based framework to teach LLMs to synthesize privacy computing code without asking experts to write tons of code and by leveraging the instruction-following and in-context learning ability of LLMs. Specifically, SPDZCoder decouples the translation task into the refactoring stage and the generation stage, which can mitigate the semantic-expressing differences at different levels. In addition, SPDZCoder can further improve its performance by a feedback stage. SPDZCoder does not require fine-tuning since it adopts an in-context learning paradigm of LLMs. To evaluate SPDZCoder, we manually created a benchmark dataset, named SPDZEval, containing six classes of difficult tasks to implement in MP-SPDZ. We conduct experiments on SPDZEval and the experimental results shows that SPDZCoder achieves the state-of-the-art performance in pass@1 and pass@2 across six data splits. Specifically, SPDZCoder achieves an overall correctness of 85.94% and 92.01% in pass@1 and pass@2, respectively, significantly surpassing baselines (at most 30.35% and 49.84% in pass@1 and pass@2, respectively) by a large margin.

Large Language Models (LLMs) are becoming essential tools for various natural language processing tasks but often suffer from generating outdated or incorrect information. Retrieval-Augmented Generation (RAG) addresses this issue by incorporating external, real-time information retrieval to ground LLM responses. However, the existing RAG systems frequently struggle with the quality of retrieval documents, as irrelevant or noisy documents degrade performance, increase computational overhead, and undermine response reliability. To tackle this problem, we propose Multi-Agent Filtering Retrieval-Augmented Generation (MAIN-RAG), a training-free RAG framework that leverages multiple LLM agents to collaboratively filter and score retrieved documents. Specifically, MAIN-RAG introduces an adaptive filtering mechanism that dynamically adjusts the relevance filtering threshold based on score distributions, effectively minimizing noise while maintaining high recall of relevant documents. The proposed approach leverages inter-agent consensus to ensure robust document selection without requiring additional training data or fine-tuning. Experimental results across four QA benchmarks demonstrate that MAIN-RAG consistently outperforms traditional RAG approaches, achieving a 2-11% improvement in answer accuracy while reducing the number of irrelevant retrieved documents. Quantitative analysis further reveals that our approach achieves superior response consistency and answer accuracy over baseline methods, offering a competitive and practical alternative to training-based solutions.

The China Space Station (CSS) is currently in orbit and carries the high-precision optical atomic clock with stability of approximately $2.0 \times 10^{-15} / \sqrt{\tau}$ in its experiment module. We have developed a model to determine the gravity potential (GP) based on the gravity frequency shift equation and have created both one-way and dual-frequency transfer models up to $c^{-4}$. These models consider effects from the troposphere, ionosphere, and solid Earth tides. The proposed model is suitable for measurements at the magnitude of $10^{-19}$. Based on the CSS mission, we conducted the simulation experiments. The results indicate that when processing the simulation frequency signal using the proposed model, we can obtain the GP with the accuracies of $ (1.13\pm0.71)\,\mathrm{m^2/s^2}$, $ (0.09\pm0.89)\,\mathrm{m^2/s^2}$, and $(0.66\pm1.18)\,\mathrm{m^2/s^2}$ for cutoff elevation angles of $5^{\circ}$, $10^{\circ}$ and $15^{\circ}$, respectively. With the high-precision optical atomic clock onboard the CSS, the proposed model enables us to measure the GP differences in the magnitude of centimeter-level accuracy.

We present a three-stage framework for training deep learning models specializing in antibody sequence-structure co-design. We first pre-train a language model using millions of antibody sequence data. Then, we employ the learned representations to guide the training of a diffusion model for joint optimization over both sequence and structure of antibodies. During the final alignment stage, we optimize the model to favor antibodies with low repulsion and high attraction to the antigen binding site, enhancing the rationality and functionality of the designs. To mitigate conflicting energy preferences, we extend AbDPO (Antibody Direct Preference Optimization) to guide the model towards Pareto optimality under multiple energy-based alignment objectives. Furthermore, we adopt an iterative learning paradigm with temperature scaling, enabling the model to benefit from diverse online datasets without requiring additional data. In practice, our proposed methods achieve high stability and efficiency in producing a better Pareto front of antibody designs compared to top samples generated by baselines and previous alignment techniques. Through extensive experiments, we showcase the superior performance of our methods in generating nature-like antibodies with high binding affinity consistently., Comment: 30 pages

Data-centric methods have shown great potential in understanding and predicting spatiotemporal dynamics, enabling better design and control of the object system. However, deep learning models often lack interpretability, fail to obey intrinsic physics, and struggle to cope with the various domains. While geometry-based methods, e.g., graph neural networks (GNNs), have been proposed to further tackle these challenges, they still need to find the implicit physical laws from large datasets and rely excessively on rich labeled data. In this paper, we herein introduce the conservation-informed GNN (CiGNN), an end-to-end explainable learning framework, to learn spatiotemporal dynamics based on limited training data. The network is designed to conform to the general conservation law via symmetry, where conservative and non-conservative information passes over a multiscale space enhanced by a latent temporal marching strategy. The efficacy of our model has been verified in various spatiotemporal systems based on synthetic and real-world datasets, showing superiority over baseline models. Results demonstrate that CiGNN exhibits remarkable accuracy and generalizability, and is readily applicable to learning for prediction of various spatiotemporal dynamics in a spatial domain with complex geometry.

Customized processors are attractive solutions for vast domain-specific applications due to their high energy efficiency. However, designing a processor in traditional flows is time-consuming and expensive. To address this, researchers have explored methods including the use of agile development tools like Chisel or SpinalHDL, high-level synthesis (HLS) from programming languages like C or SystemC, and more recently, leveraging large language models (LLMs) to generate hardware description language (HDL) code from natural language descriptions. However, each method has limitations in terms of expressiveness, correctness, and performance, leading to a persistent contradiction between the level of automation and the effectiveness of the design. Overall, how to automatically design highly efficient and practical processors with minimal human effort remains a challenge. In this paper, we propose AGON, a novel framework designed to leverage LLMs for the efficient design of out-of-order (OoO) customized processors with minimal human effort. Central to AGON is the nano-operator function (nOP function) based Intermediate Representation (IR), which bridges high-level descriptions and hardware implementations while decoupling functionality from performance optimization, thereby providing an automatic design framework that is expressive and efficient, has correctness guarantees, and enables PPA (Power, Performance, and Area) optimization. Experimental results show that superior to previous LLM-assisted automatic design flows, AGON facilitates designing a series of customized OoO processors that achieve on average 2.35 $\times$ speedup compared with BOOM, a general-purpose CPU designed by experts, with minimal design effort.

The spin-12 three-leg antiferromagnetic Heisenberg spin ladder is studied under open boundary condition (OBC) and cylinder boundary condition (CBC), using the density matrix renormalization group and matrix product state methods, respectively. Specifically, we calculate the energy density, entanglement entropy, and concurrence while discussing the effects of interleg interaction J2 and the alternating coupling parameter gamma on these quantities. It is found that the introduction of gamma can completely reverse the concurrence distribution between odd and even bonds. Under CBC, the generation of the interleg concurrence is inhibited when gamma=0, and the introduction of gamma can cause interleg concurrence between chains 1 and 3, in which the behavior is more complicated due to the competition between CBC and gamma. Additionally, we find that gamma induces two types of long-distance entanglement (LDE) in the system under OBC: intraleg LDE and inter-leg one. When the system size is sufficiently large, both types of LDE reach similar strength and stabilize at a constant value. The study indicates that the three-leg ladder makes it easier to generate LDE compared with the two-leg system. However, the generation of LDE is inhibited under CBC which the spin frustration exists. In addition, the calculated results of energy, entanglement entropy and concurrence all show that there are essential relations between these quantities and phase transitions of the system. Further, we predict a phase transition point near gamma=0.54 under OBC. The present study provides valuable insights into understanding the phase diagram of this class of systems., Comment: 24 pages,11 figures

Quantum process tomography (QPT), used for reconstruction of an unknown quantum process from measurement data, is a fundamental tool for the diagnostic and full characterization of quantum systems. It relies on querying a set of quantum states as input to the quantum process. Previous works commonly use a straightforward strategy to select a set of quantum states randomly, overlooking differences in informativeness among quantum states. Since querying the quantum system requires multiple experiments that can be prohibitively costly, it is always the case that there are not enough quantum states for high-quality reconstruction. In this paper, we propose a general framework for active learning (AL) to adaptively select a set of informative quantum states that improves the reconstruction most efficiently. In particular, we introduce a learning framework that leverages the widely-used variational quantum circuits (VQCs) to perform the QPT task and integrate our AL algorithms into the query step. We design and evaluate three various types of AL algorithms: committee-based, uncertainty-based, and diversity-based, each exhibiting distinct advantages in terms of performance and computational cost. Additionally, we provide a guideline for selecting algorithms suitable for different scenarios. Numerical results demonstrate that our algorithms achieve significantly improved reconstruction compared to the baseline method that selects a set of quantum states randomly. Moreover, these results suggest that active learning based approaches are applicable to other complicated learning tasks in large-scale quantum information processing.

Accurate multi-view 3D object detection is essential for applications such as autonomous driving. Researchers have consistently aimed to leverage LiDAR's precise spatial information to enhance camera-based detectors through methods like depth supervision and bird-eye-view (BEV) feature distillation. However, existing approaches often face challenges due to the inherent differences between LiDAR and camera data representations. In this paper, we introduce the TiGDistill-BEV, a novel approach that effectively bridges this gap by leveraging the strengths of both sensors. Our method distills knowledge from diverse modalities(e.g., LiDAR) as the teacher model to a camera-based student detector, utilizing the Target Inner-Geometry learning scheme to enhance camera-based BEV detectors through both depth and BEV features by leveraging diverse modalities. Specially, we propose two key modules: an inner-depth supervision module to learn the low-level relative depth relations within objects which equips detectors with a deeper understanding of object-level spatial structures, and an inner-feature BEV distillation module to transfer high-level semantics of different key points within foreground targets. To further alleviate the domain gap, we incorporate both inter-channel and inter-keypoint distillation to model feature similarity. Extensive experiments on the nuScenes benchmark demonstrate that TiGDistill-BEV significantly boosts camera-based only detectors achieving a state-of-the-art with 62.8% NDS and surpassing previous methods by a significant margin. The codes is available at: https://github.com/Public-BOTs/TiGDistill-BEV.git., Comment: 13 pages, 8 figures. arXiv admin note: substantial text overlap with arXiv:2212.13979

In the process of exploring quantum algorithms, researchers often need to conduct equivalence checking of quantum circuits with different structures or to reconstruct a circuit in a variational manner, aiming to reduce the depth of the target circuit. Whereas the exponential resource overhead for describing quantum systems classically makes the existing methods not amenable to serving large-scale quantum circuits. Grounded in the entangling quantum generative adversarial network (EQ-GAN), we present in this article a new architecture which enables one to redesign large-scale quantum circuits on quantum hardware. For concreteness, we apply this architecture to three crucial applications in circuit optimization, including the equivalence checking of (non-) parameterized circuits, as well as the variational reconstruction of quantum circuits. The feasibility of our approach is demonstrated by the excellent results of these applications, which are implemented both in classical computers and current NISQ hardware. We believe our work should facilitate the implementation and validation of the advantages of quantum algorithms., Comment: 9 pages,11 figures

In previous studies of relativistic thermodynamics, the temperature of a static system, as perceived by a moving observer, has traditionally been treated as a scalar. This assumption has also been extended to the research on the cosmic microwave background. However, the validity of this assumption is a consequence of the massless nature of photons. More generally, when an observer is in relative motion to a system, the thermal equilibrium state is characterized by a four-vector temperature. In this paper, we study the non-interacting massive Bosonic and Fermionic field systems. We derive the Lorentz transformation of the energy spectral density in the equilibrium state of these fields. In the massless limit for bosonic field, our results recover the transformation of black body radiation [G. W. Ford and R. F. O' Connell., Phys. Rev. E, 88, 044101(2013)], which corresponds to a scalar temperature with dipole anisotropy. For the massive fields, the moving equilibrium state cannot be characterized by a corresponding scalar temperature. This result shows the necessity of introducing four-vector temperature in relativistic thermodynamics.

Time series forecasting plays a crucial role in data mining, driving rapid advancements across numerous industries. With the emergence of large models, time series foundation models (TSFMs) have exhibited remarkable generalization capabilities, such as zero-shot learning, through large-scale pre-training. Meanwhile, Retrieval-Augmented Generation (RAG) methods have been widely employed to enhance the performance of foundation models on unseen data, allowing models to access to external knowledge. In this paper, we introduce TimeRAF, a Retrieval-Augmented Forecasting model that enhance zero-shot time series forecasting through retrieval-augmented techniques. We develop customized time series knowledge bases that are tailored to the specific forecasting tasks. TimeRAF employs an end-to-end learnable retriever to extract valuable information from the knowledge base. Additionally, we propose Channel Prompting for knowledge integration, which effectively extracts relevant information from the retrieved knowledge along the channel dimension. Extensive experiments demonstrate the effectiveness of our model, showing significant improvement across various domains and datasets.

The domain gap between remote sensing imagery and natural images has recently received widespread attention and Vision-Language Models (VLMs) have demonstrated excellent generalization performance in remote sensing multimodal tasks. However, current research is still limited in exploring how remote sensing VLMs handle different types of visual inputs. To bridge this gap, we introduce \textbf{UniRS}, the first vision-language model \textbf{uni}fying multi-temporal \textbf{r}emote \textbf{s}ensing tasks across various types of visual input. UniRS supports single images, dual-time image pairs, and videos as input, enabling comprehensive remote sensing temporal analysis within a unified framework. We adopt a unified visual representation approach, enabling the model to accept various visual inputs. For dual-time image pair tasks, we customize a change extraction module to further enhance the extraction of spatiotemporal features. Additionally, we design a prompt augmentation mechanism tailored to the model's reasoning process, utilizing the prior knowledge of the general-purpose VLM to provide clues for UniRS. To promote multi-task knowledge sharing, the model is jointly fine-tuned on a mixed dataset. Experimental results show that UniRS achieves state-of-the-art performance across diverse tasks, including visual question answering, change captioning, and video scene classification, highlighting its versatility and effectiveness in unifying these multi-temporal remote sensing tasks. Our code and dataset will be released soon., Comment: 12 pages, 5 figures

Chart summarization, which focuses on extracting key information from charts and interpreting it in natural language, is crucial for generating and delivering insights through effective and accessible data analysis. Traditional methods for chart understanding and summarization often rely on multi-stage pipelines, which may produce suboptimal semantic alignment between visual and textual information. In comparison, recently developed LLM-based methods are more dependent on the capability of foundation images or languages, while ignoring the characteristics of chart data and its relevant challenges. To address these limitations, we propose ChartAdapter, a novel lightweight transformer module designed to bridge the gap between charts and textual summaries. ChartAdapter employs learnable query vectors to extract implicit semantics from chart data and incorporates a cross-modal alignment projector to enhance vision-to-language generative learning. By integrating ChartAdapter with an LLM, we enable end-to-end training and efficient chart summarization. To further enhance the training, we introduce a three-stage hierarchical training procedure and develop a large-scale dataset specifically curated for chart summarization, comprising 190,618 samples. Experimental results on the standard Chart-to-Text testing set demonstrate that our approach significantly outperforms existing methods, including state-of-the-art models, in generating high-quality chart summaries. Ablation studies further validate the effectiveness of key components in ChartAdapter. This work highlights the potential of tailored LLM-based approaches to advance chart understanding and sets a strong foundation for future research in this area.

With rapid advances in large language models (LLMs), there has been an increasing application of LLMs in creative content ideation and generation. A critical question emerges: can current LLMs provide ideas that are diverse enough to truly bolster the collective creativity? We examine two state-of-the-art LLMs, GPT-4 and LLaMA-3, on story generation and discover that LLM-generated stories often consist of plot elements that are echoed across a number of generations. To quantify this phenomenon, we introduce the Sui Generis score, which estimates how unlikely a plot element is to appear in alternative storylines generated by the same LLM. Evaluating on 100 short stories, we find that LLM-generated stories often contain combinations of idiosyncratic plot elements echoed frequently across generations, while the original human-written stories are rarely recreated or even echoed in pieces. Moreover, our human evaluation shows that the ranking of Sui Generis scores among story segments correlates moderately with human judgment of surprise level, even though score computation is completely automatic without relying on human judgment.

Large language models (LLMs) have demonstrated exceptional performance in text generation within current NLP research. However, the lack of factual accuracy is still a dark cloud hanging over the LLM skyscraper. Structural knowledge prompting (SKP) is a prominent paradigm to integrate external knowledge into LLMs by incorporating structural representations, achieving state-of-the-art results in many knowledge-intensive tasks. However, existing methods often focus on specific problems, lacking a comprehensive exploration of the generalization and capability boundaries of SKP. This paper aims to evaluate and rethink the generalization capability of the SKP paradigm from four perspectives including Granularity, Transferability, Scalability, and Universality. To provide a thorough evaluation, we introduce a novel multi-granular, multi-level benchmark called SUBARU, consisting of 9 different tasks with varying levels of granularity and difficulty., Comment: Work in progress

Image content safety has become a significant challenge with the rise of visual media on online platforms. Meanwhile, in the age of AI-generated content (AIGC), many image generation models are capable of producing harmful content, such as images containing sexual or violent material. Thus, it becomes crucial to identify such unsafe images based on established safety rules. Pre-trained Multimodal Large Language Models (MLLMs) offer potential in this regard, given their strong pattern recognition abilities. Existing approaches typically fine-tune MLLMs with human-labeled datasets, which however brings a series of drawbacks. First, relying on human annotators to label data following intricate and detailed guidelines is both expensive and labor-intensive. Furthermore, users of safety judgment systems may need to frequently update safety rules, making fine-tuning on human-based annotation more challenging. This raises the research question: Can we detect unsafe images by querying MLLMs in a zero-shot setting using a predefined safety constitution (a set of safety rules)? Our research showed that simply querying pre-trained MLLMs does not yield satisfactory results. This lack of effectiveness stems from factors such as the subjectivity of safety rules, the complexity of lengthy constitutions, and the inherent biases in the models. To address these challenges, we propose a MLLM-based method includes objectifying safety rules, assessing the relevance between rules and images, making quick judgments based on debiased token probabilities with logically complete yet simplified precondition chains for safety rules, and conducting more in-depth reasoning with cascaded chain-of-thought processes if necessary. Experiment results demonstrate that our method is highly effective for zero-shot image safety judgment tasks.

In this paper, we present a systematic study of the Chern--Simons theory with gauge group \(\mathrm{SL}(2,\mathbb{R})\times\mathrm{SL}(2,\mathbb{R})\) restricted to a wedge-identified manifold in the hyperbolic upper-half-space. The wedge geometry is created by imposing an angular cutoff in the \((x,y)\) plane and identifying two boundary lines, which introduces a single noncontractible loop in the manifold. By imposing the flat-connection condition of the Chern--Simons gauge fields, the path integral reduces to a finite-dimensional matrix integral in \(\mathrm{SL}(2,\mathbb{R})\times\mathrm{SL}(2,\mathbb{R})\) . Although Chern-Simons theory is a topological theory, the resulting matrix integral remains nontrivial due to noncompact directions and boundary constraints. The large-\(k\) expansion of the matrix integral is carried out by selecting a classical configuration in the space of holonomies and expanding around it in inverse powers of \(k\). The resulting coefficients of the asymptotic series exhibit factorial growth, enabling us to apply the Borel resummation techniques. Summation over these subleading sectors removes potential ambiguities in the Borel integral and clarifies the emergence of a resurgent transseries structure. In the Borel-resurgent analysis, we show that, despite the apparent simplicity of the reduced action, the wedge geometry yields a rich interplay of perturbative and non-perturbative phenomena. This work presents an explicit example of how a finite-dimensional matrix integral in its expansions is physically meaningful through Borel resummation.

The remarkable performance of models like the OpenAI o1 can be attributed to their ability to emulate human-like long-time thinking during inference. These models employ extended chain-of-thought (CoT) processes, exploring multiple strategies to enhance problem-solving capabilities. However, a critical question remains: How to intelligently and efficiently scale computational resources during testing. This paper presents the first comprehensive study on the prevalent issue of overthinking in these models, where excessive computational resources are allocated for simple problems with minimal benefit. We introduce novel efficiency metrics from both outcome and process perspectives to evaluate the rational use of computational resources by o1-like models. Using a self-training paradigm, we propose strategies to mitigate overthinking, streamlining reasoning processes without compromising accuracy. Experimental results show that our approach successfully reduces computational overhead while preserving model performance across a range of testsets with varying difficulty levels, such as GSM8K, MATH500, GPQA, and AIME., Comment: Work in progress

The evidence for the existence of dark matter (DM) is compelling, yet its nature remains elusive. A particularly interesting and minimal scenario involves DM with pure gravitational interactions. In the early Universe, such DM can be unavoidably generated via annihilation of particles in the standard model (SM) thermal plasma. It is known that the SM thermal plasma also produces gravitational waves (GWs). In this study, we point out a simple and tight connection between the amplitude of the thermal GWs and the properties of pure gravitational DM. Notably, future GW experiments in the ultra-high frequency regime have the potential to shed light on the mass and spin of pure gravitational DM., Comment: v1: two columns, 4 pages, 2 figures

An appropriate choice of batch sizes in large-scale model training is crucial, yet it involves an intrinsic yet inevitable dilemma: large-batch training improves training efficiency in terms of memory utilization, while generalization performance often deteriorates due to small amounts of gradient noise. Despite this dilemma, the common practice of choosing batch sizes in language model training often prioritizes training efficiency -- employing either constant large sizes with data parallelism or implementing batch size warmup schedules. However, such batch size schedule designs remain heuristic and often fail to adapt to training dynamics, presenting the challenge of designing adaptive batch size schedules. Given the abundance of available datasets and the data-hungry nature of language models, data parallelism has become an indispensable distributed training paradigm, enabling the use of larger batch sizes for gradient computation. However, vanilla data parallelism requires replicas of model parameters, gradients, and optimizer states at each worker, which prohibits training larger models with billions of parameters. To optimize memory usage, more advanced parallelism strategies must be employed. In this work, we propose general-purpose and theoretically principled adaptive batch size schedules compatible with data parallelism and model parallelism. We develop a practical implementation with PyTorch Fully Sharded Data Parallel, facilitating the pretraining of language models of different sizes. We empirically demonstrate that our proposed approaches outperform constant batch sizes and heuristic batch size warmup schedules in the pretraining of models in the Llama family, with particular focus on smaller models with up to 3 billion parameters. We also establish theoretical convergence guarantees for such adaptive batch size schedules with Adam for general smooth nonconvex objectives.

We introduce Vinci, a real-time embodied smart assistant built upon an egocentric vision-language model. Designed for deployment on portable devices such as smartphones and wearable cameras, Vinci operates in an "always on" mode, continuously observing the environment to deliver seamless interaction and assistance. Users can wake up the system and engage in natural conversations to ask questions or seek assistance, with responses delivered through audio for hands-free convenience. With its ability to process long video streams in real-time, Vinci can answer user queries about current observations and historical context while also providing task planning based on past interactions. To further enhance usability, Vinci integrates a video generation module that creates step-by-step visual demonstrations for tasks that require detailed guidance. We hope that Vinci can establish a robust framework for portable, real-time egocentric AI systems, empowering users with contextual and actionable insights. We release the complete implementation for the development of the device in conjunction with a demo web platform to test uploaded videos at https://github.com/OpenGVLab/vinci.

As a verified need, consistent editing across in-the-wild images remains a technical challenge arising from various unmanageable factors, like object poses, lighting conditions, and photography environments. Edicho steps in with a training-free solution based on diffusion models, featuring a fundamental design principle of using explicit image correspondence to direct editing. Specifically, the key components include an attention manipulation module and a carefully refined classifier-free guidance (CFG) denoising strategy, both of which take into account the pre-estimated correspondence. Such an inference-time algorithm enjoys a plug-and-play nature and is compatible to most diffusion-based editing methods, such as ControlNet and BrushNet. Extensive results demonstrate the efficacy of Edicho in consistent cross-image editing under diverse settings. We will release the code to facilitate future studies., Comment: Project page: https://github.com/EzioBy/edicho