Search

Silicon nitride (Si$_3$N$_4$) photonic integrated circuits (PICs) have shown low linear loss, negligible nonlinear loss, and high power handling over traditional silicon photonics. To achieve high-density photonic integration and high effective nonlinearity through tight optical confinement, thick stoichiometric Si$_3$N$_4$ films are indispensable. However, when using low-pressure chemical vapor deposition (LPCVD) to achieve high optical material transparency, Si$_3$N$_4$ films exhibit large tensile stress on the order of GPa. Methods for crack prevention are therefore essential. The photonic Damascene process has addressed this issue, attaining record low loss Si$_3$N$_4$ PICs, but it lacks control of the waveguide height. Conversely, precise waveguide dimension and ultra-low loss have been achieved with subtractive processing, but this method is not compatible with mass production due to the use of electron beam lithography. To date, an outstanding challenge is to attain both lithographic precision and ultra-low loss in high confinement Si$_3$N$_4$ PICs that are compatible with large-scale foundry manufacturing. Here, we present a single-step deposited, DUV-based subtractive method for producing wafer-scale ultra-low loss Si$_3$N$_4$ PICs that harmonize these necessities. By employing deep etching of densely distributed, interconnected trenches into the substrate, we effectively mitigate the tensile stress in the Si$_3$N$_4$ layer, enabling direct deposition of thick films without cracking and substantially prolonged storage duration. Lastly, we identify ultraviolet (UV) radiation-induced damage that can be remedied through rapid thermal annealing. Collectively, we develop ultra-low loss Si$_3$N$_4$ microresonators and 0.5 m-long spiral waveguides with losses down to 1.4 dB/m at 1550 nm with high production yield., Comment: 2023 Conference on Lasers and Electro-Optics (CLEO). IEEE, 2023

Data generation-based zero-shot learning, although effective in training Small Task-specific Models (STMs) via synthetic datasets generated by Pre-trained Language Models (PLMs), is often limited by the low quality of such synthetic datasets. Previous solutions have primarily focused on single PLM settings, where synthetic datasets are typically restricted to specific sub-spaces and often deviate from real-world distributions, leading to severe distribution bias. To mitigate such bias, we propose FuseGen, a novel data generation-based zero-shot learning framework that introduces a new criteria for subset selection from synthetic datasets via utilizing multiple PLMs and trained STMs. The chosen subset provides in-context feedback to each PLM, enhancing dataset quality through iterative data generation. Trained STMs are then used for sample re-weighting as well, further improving data quality. Extensive experiments across diverse tasks demonstrate that FuseGen substantially outperforms existing methods, highly effective in boosting STM performance in a PLM-agnostic way. Code is provided in https://github.com/LindaLydia/FuseGen., Comment: 17 pages, 8 figures, 12 tabels

The quantum dynamics of the intrinsic metric profoundly influence the neutral excitations in the fractional quantum Hall system, as established by Haldane in 2011 \cite{Haldane2011}, and further evidenced by a recent two-photon experiment \cite{Liang2024}. Despite these advancements, a comprehensive understanding of the dynamic properties of these excitations, especially at long wavelengths, continues to elude interest. In this study, we employ tensor-network methods to investigate the neutral excitations of the Laughlin and Moore-Read states on an infinite cylinder. This investigation deepens our understanding of the excitation spectrum in regions where traditional methods do not work effectively. The spectral functions for both states reveal the presence of $S=-2$ geometric excitations. For the first time, we unveil the complex spectra of both neutral fermion and bosonic Girvin-MacDonald-Platzman modes within the excitation continuum by calculating the three-particle density response function for the Moore-Read state. Our findings support the hypothesis of emergent supersymmetry and highlight the potential for detecting neutral fermions in future experiments.

The standard Reinforcement Learning from Human Feedback (RLHF) framework primarily focuses on optimizing the performance of large language models using pre-collected prompts. However, collecting prompts that provide comprehensive coverage is both tedious and challenging, and often fails to include scenarios that LLMs need to improve on the most. In this paper, we investigate alignment through the lens of two-agent games, involving iterative interactions between an adversarial and a defensive agent. The adversarial agent's task at each step is to generate prompts that expose the weakness of the defensive agent. In return, the defensive agent seeks to improve its responses to these newly identified prompts it struggled with, based on feedback from the reward model. We theoretically demonstrate that this iterative reinforcement learning optimization converges to a Nash Equilibrium for the game induced by the agents. Experimental results in safety scenarios demonstrate that learning in such a competitive environment not only fully trains agents but also leads to policies with enhanced generalization capabilities for both adversarial and defensive agents., Comment: Our code is released at https://github.com/ruizheng20/gpo

Spectroscopic data, particularly diffraction data, contain detailed crystal and microstructure information and thus are crucial for materials discovery. Powder X-ray diffraction (XRD) patterns are greatly effective in identifying crystals. Although machine learning (ML) has significantly advanced the analysis of powder XRD patterns, the progress is hindered by a lack of training data. To address this, we introduce SimXRD, the largest open-source simulated XRD pattern dataset so far, to accelerate the development of crystallographic informatics. SimXRD comprises 4,065,346 simulated powder X-ray diffraction patterns, representing 119,569 distinct crystal structures under 33 simulated conditions that mimic real-world variations. We find that the crystal symmetry inherently follows a long-tailed distribution and evaluate 21 sequence learning models on SimXRD. The results indicate that existing neural networks struggle with low-frequency crystal classifications. The present work highlights the academic significance and the engineering novelty of simulated XRD patterns in this interdisciplinary field.

The intersection of physics-based vision and deep learning presents an exciting frontier for advancing computer vision technologies. By leveraging the principles of physics to inform and enhance deep learning models, we can develop more robust and accurate vision systems. Physics-based vision aims to invert the processes to recover scene properties such as shape, reflectance, light distribution, and medium properties from images. In recent years, deep learning has shown promising improvements for various vision tasks, and when combined with physics-based vision, these approaches can enhance the robustness and accuracy of vision systems. This technical report summarizes the outcomes of the Physics-Based Vision Meets Deep Learning (PBDL) 2024 challenge, held in CVPR 2024 workshop. The challenge consisted of eight tracks, focusing on Low-Light Enhancement and Detection as well as High Dynamic Range (HDR) Imaging. This report details the objectives, methodologies, and results of each track, highlighting the top-performing solutions and their innovative approaches., Comment: CVPR 2024 PBDL Challenges: https://pbdl-ws.github.io/pbdl2024/challenge/index.html

Large language models (LLMs), pre-trained or fine-tuned on large code corpora, have shown effectiveness in generating code completions. However, in LLM-based code completion, LLMs may struggle to use correct and up-to-date Application Programming Interfaces (APIs) due to the rapid and continuous evolution of libraries. While existing studies have highlighted issues with predicting incorrect APIs, the specific problem of deprecated API usage in LLM-based code completion has not been thoroughly investigated. To address this gap, we conducted the first evaluation study on deprecated API usage in LLM-based code completion. This study involved seven advanced LLMs, 145 API mappings from eight popular Python libraries, and 28,125 completion prompts. The study results reveal the \textit{status quo} and \textit{root causes} of deprecated API usage in LLM-based code completion from the perspectives of \textit{model}, \textit{prompt}, and \textit{library}. Based on these findings, we propose two lightweight fixing approaches, \textsc{ReplaceAPI} and \textsc{InsertPrompt}, which can serve as baseline approaches for future research on mitigating deprecated API usage in LLM-based completion. Additionally, we provide implications for future research on integrating library evolution with LLM-driven software development.

For a selection of 35 pulsars with large spin-up glitches ($\Delta{\nu}/\nu\geq10^{-6}$), which are monitored by the Jodrell Bank Observatory, we analyse 157 glitches and their recoveries. All parameters are measured consistently and we choose the best model to describe the post-glitch recovery based on Bayesian evidence. We present updated glitch epochs, sizes, changes of spin down rate, exponentially recovering components (amplitude and corresponding timescale) when present, as well as pulsars' second frequency derivatives and their glitch associated changes if detected. We discuss the different observed styles of post-glitch recovery as well as some particularly interesting sources. Several correlations are revealed between glitch parameters and pulsar spin parameters, including a very strong correlation between a pulsar's interglitch $|\ddot{\nu}|$ and $\dot{\nu}$, as well as between the glitch-induced spin-down rate change $\Delta\dot{\nu}_{\rm p}$ that does not relax exponentially and $\dot{\nu}$. We find that the ratio $\left|\Delta \dot{\nu}_{\mathrm{p}}/\ddot{\nu}\right|$ can be used as an estimate of glitch recurrence times, especially for those pulsars for which there are indications of a characteristic glitch size and interglitch waiting time. We calculate the interglitch braking index $n$ and find that pulsars with large glitches typically have $n$ greater than $3$, suggesting that internal torques dominate the rotational evolution between glitches. The external torque, e.g. from electromagnetic dipole radiation, could dominate the observed $\ddot{\nu}$ for the youngest pulsars ($\lesssim10^{4}\;\mathrm{yr}$), which may be expected to display $n\sim3$., Comment: 24 pages, 14 figures, 8 tables. Accepted for publication in MNRAS

Balancing influential covariates is crucial for valid treatment comparisons in clinical studies. While covariate-adaptive randomization is commonly used to achieve balance, its performance can be inadequate when the number of baseline covariates is large. It is therefore essential to identify the influential factors associated with the outcome and ensure balance among these critical covariates. In this article, we propose a novel adaptive randomization approach that integrates the patients' responses and covariates information to select sequentially significant covariates and maintain their balance. We establish theoretically the consistency of our covariate selection method and demonstrate that the improved covariate balancing, as evidenced by a faster convergence rate of the imbalance measure, leads to higher efficiency in estimating treatment effects. Furthermore, we provide extensive numerical and empirical studies to illustrate the benefits of our proposed method across various settings., Comment: 58 pages, 4 figures

Large language models (LLMs) have advanced to encompass extensive knowledge across diverse domains. Yet controlling what a large language model should not know is important for ensuring alignment and thus safe use. However, accurately and efficiently unlearning knowledge from an LLM remains challenging due to the potential collateral damage caused by the fuzzy boundary between retention and forgetting, and the large computational requirements for optimization across state-of-the-art models with hundreds of billions of parameters. In this work, we present \textbf{Embedding-COrrupted (ECO) Prompts}, a lightweight unlearning framework for large language models to address both the challenges of knowledge entanglement and unlearning efficiency. Instead of relying on the LLM itself to unlearn, we enforce an unlearned state during inference by employing a prompt classifier to identify and safeguard prompts to forget. We learn corruptions added to prompt embeddings via zeroth order optimization toward the unlearning objective offline and corrupt prompts flagged by the classifier during inference. We find that these embedding-corrupted prompts not only lead to desirable outputs that satisfy the unlearning objective but also closely approximate the output from a model that has never been trained on the data intended for forgetting. Through extensive experiments on unlearning, we demonstrate the superiority of our method in achieving promising unlearning at \textit{nearly zero side effects} in general domains and domains closely related to the unlearned ones. Additionally, we highlight the scalability of our method to 100 LLMs, ranging from 0.5B to 236B parameters, incurring no additional cost as the number of parameters increases. We have made our code publicly available at \url{https://github.com/chrisliu298/llm-unlearn-eco}., Comment: NeurIPS 2024 Poster

The objective of machine unlearning (MU) is to eliminate previously learned data from a model. However, it is challenging to strike a balance between computation cost and performance when using existing MU techniques. Taking inspiration from the influence of label smoothing on model confidence and differential privacy, we propose a simple gradient-based MU approach that uses an inverse process of label smoothing. This work introduces UGradSL, a simple, plug-and-play MU approach that uses smoothed labels. We provide theoretical analyses demonstrating why properly introducing label smoothing improves MU performance. We conducted extensive experiments on six datasets of various sizes and different modalities, demonstrating the effectiveness and robustness of our proposed method. The consistent improvement in MU performance is only at a marginal cost of additional computations. For instance, UGradSL improves over the gradient ascent MU baseline by 66% unlearning accuracy without sacrificing unlearning efficiency.

This paper focuses on the challenge of machine unlearning, aiming to remove the influence of specific training data on machine learning models. Traditionally, the development of unlearning algorithms runs parallel with that of membership inference attacks (MIA), a type of privacy threat to determine whether a data instance was used for training. However, the two strands are intimately connected: one can view machine unlearning through the lens of MIA success with respect to removed data. Recognizing this connection, we propose a game-theoretic framework that integrates MIAs into the design of unlearning algorithms. Specifically, we model the unlearning problem as a Stackelberg game in which an unlearner strives to unlearn specific training data from a model, while an auditor employs MIAs to detect the traces of the ostensibly removed data. Adopting this adversarial perspective allows the utilization of new attack advancements, facilitating the design of unlearning algorithms. Our framework stands out in two ways. First, it takes an adversarial approach and proactively incorporates the attacks into the design of unlearning algorithms. Secondly, it uses implicit differentiation to obtain the gradients that limit the attacker's success, thus benefiting the process of unlearning. We present empirical results to demonstrate the effectiveness of the proposed approach for machine unlearning.

Efficiently modeling sequences with infinite context length has long been a challenging problem. Previous approaches have either suffered from quadratic computational complexity or limited extrapolation ability in length generalization. In this work, we present Samba, a simple hybrid architecture that layer-wise combines Mamba, a selective State Space Model (SSM), with Sliding Window Attention (SWA). Samba selectively compresses a given sequence into recurrent hidden states while still maintaining the ability to precisely recall recent memories with the attention mechanism. We scale Samba up to 3.8B parameters with 3.2T training tokens and demonstrate that it significantly outperforms state-of-the-art models across a variety of benchmarks. Pretrained on sequences of 4K length, Samba shows improved perplexity in context lengths of up to 1M in zero-shot. When finetuned on 4K-length sequences, Samba efficiently extrapolates to a 256K context length with perfect memory recall on the Passkey Retrieval task, and exhibits superior retrieval extrapolation on the challenging Phonebook task compared to full-attention models. As a linear-time sequence model, Samba achieves a 3.73x higher throughput compared to Transformers with grouped-query attention for user prompts of 128K length, and a 3.64x speedup when generating 64K tokens with unlimited streaming. Our code for training on open source data is publicly available at https://github.com/microsoft/Samba.

The widespread integration of Machine Learning systems in daily life, particularly in high-stakes domains, has raised concerns about the fairness implications. While prior works have investigated static fairness measures, recent studies reveal that automated decision-making has long-term implications and that off-the-shelf fairness approaches may not serve the purpose of achieving long-term fairness. Additionally, the existence of feedback loops and the interaction between models and the environment introduces additional complexities that may deviate from the initial fairness goals. In this survey, we review existing literature on long-term fairness from different perspectives and present a taxonomy for long-term fairness studies. We highlight key challenges and consider future research directions, analyzing both current issues and potential further explorations.

Universal adversarial perturbation (UAP), also known as image-agnostic perturbation, is a fixed perturbation map that can fool the classifier with high probabilities on arbitrary images, making it more practical for attacking deep models in the real world. Previous UAP methods generate a scale-fixed and texture-fixed perturbation map for all images, which ignores the multi-scale objects in images and usually results in a low fooling ratio. Since the widely used convolution neural networks tend to classify objects according to semantic information stored in local textures, it seems a reasonable and intuitive way to improve the UAP from the perspective of utilizing local contents effectively. In this work, we find that the fooling ratios significantly increase when we add a constraint to encourage a small-scale UAP map and repeat it vertically and horizontally to fill the whole image domain. To this end, we propose texture scale-constrained UAP (TSC-UAP), a simple yet effective UAP enhancement method that automatically generates UAPs with category-specific local textures that can fool deep models more easily. Through a low-cost operation that restricts the texture scale, TSC-UAP achieves a considerable improvement in the fooling ratio and attack transferability for both data-dependent and data-free UAP methods. Experiments conducted on two state-of-the-art UAP methods, eight popular CNN models and four classical datasets show the remarkable performance of TSC-UAP., Comment: 14 pages (accepted by TIFS2024)

Text-guided 3D visual grounding (T-3DVG), which aims to locate a specific object that semantically corresponds to a language query from a complicated 3D scene, has drawn increasing attention in the 3D research community over the past few years. Compared to 2D visual grounding, this task presents great potential and challenges due to its closer proximity to the real world and the complexity of data collection and 3D point cloud source processing. In this survey, we attempt to provide a comprehensive overview of the T-3DVG progress, including its fundamental elements, recent research advances, and future research directions. To the best of our knowledge, this is the first systematic survey on the T-3DVG task. Specifically, we first provide a general structure of the T-3DVG pipeline with detailed components in a tutorial style, presenting a complete background overview. Then, we summarize the existing T-3DVG approaches into different categories and analyze their strengths and weaknesses. We also present the benchmark datasets and evaluation metrics to assess their performances. Finally, we discuss the potential limitations of existing T-3DVG and share some insights on several promising research directions. The latest papers are continually collected at https://github.com/liudaizong/Awesome-3D-Visual-Grounding.

2D irregular packing is a classic combinatorial optimization problem with various applications, such as material utilization and texture atlas generation. This NP-hard problem requires efficient algorithms to optimize space utilization. Conventional numerical methods suffer from slow convergence and high computational cost. Existing learning-based methods, such as the score-based diffusion model, also have limitations, such as no rotation support, frequent collisions, and poor adaptability to arbitrary boundaries, and slow inferring. The difficulty of learning from teacher packing is to capture the complex geometric relationships among packing examples, which include the spatial (position, orientation) relationships of objects, their geometric features, and container boundary conditions. Representing these relationships in latent space is challenging. We propose GFPack++, an attention-based gradient field learning approach that addresses this challenge. It consists of two pivotal strategies: \emph{attention-based geometry encoding} for effective feature encoding and \emph{attention-based relation encoding} for learning complex relationships. We investigate the utilization distribution between the teacher and inference data and design a weighting function to prioritize tighter teacher data during training, enhancing learning effectiveness. Our diffusion model supports continuous rotation and outperforms existing methods on various datasets. We achieve higher space utilization over several widely used baselines, one-order faster than the previous diffusion-based method, and promising generalization for arbitrary boundaries. We plan to release our source code and datasets to support further research in this direction.

Jailbreaking is an emerging adversarial attack that bypasses the safety alignment deployed in off-the-shelf large language models (LLMs) and has evolved into multiple categories: human-based, optimization-based, generation-based, and the recent indirect and multilingual jailbreaks. However, delivering a practical jailbreak defense is challenging because it needs to not only handle all the above jailbreak attacks but also incur negligible delays to user prompts, as well as be compatible with both open-source and closed-source LLMs. Inspired by how the traditional security concept of shadow stacks defends against memory overflow attacks, this paper introduces a generic LLM jailbreak defense framework called SelfDefend, which establishes a shadow LLM as a defense instance to concurrently protect the target LLM instance in the normal stack and collaborate with it for checkpoint-based access control. The effectiveness of SelfDefend builds upon our observation that existing LLMs (both target and defense LLMs) have the capability to identify harmful prompts or intentions in user queries, which we empirically validate using the commonly used GPT-3.5/4 models across all major jailbreak attacks. To further improve the defense's robustness and minimize costs, we employ a data distillation approach to tune dedicated open-source defense models. These models outperform six state-of-the-art defenses and match the performance of GPT-4-based SelfDefend, with significantly lower extra delays. We also empirically show that the tuned models are robust to adaptive jailbreaks and prompt injections., Comment: This paper completes its earlier vision paper, available at arXiv:2402.15727. Updated to the latest analysis and results

In large-scale recommendation systems, the vast array of items makes it infeasible to obtain accurate user preferences for each product, resulting in a common issue of missing labels. Typically, only items previously recommended to users have associated ground truth data. Although there is extensive research on fairness concerning fully observed user-item interactions, the challenge of fairness in scenarios with missing labels remains underexplored. Previous methods often treat these samples missing labels as negative, which can significantly deviate from the ground truth fairness metrics. Our study addresses this gap by proposing a novel method employing a small randomized traffic to estimate fairness metrics accurately. We present theoretical bounds for the estimation error of our fairness metric and support our findings with empirical evidence on real data. Our numerical experiments on synthetic and TikTok's real-world data validate our theory and show the efficiency and effectiveness of our novel methods. To the best of our knowledge, we are the first to emphasize the necessity of random traffic in dataset collection for recommendation fairness, the first to publish a fairness-related dataset from TikTok and to provide reliable estimates of fairness metrics in the context of large-scale recommendation systems with missing labels.

Federated learning has enabled multiple parties to collaboratively train large language models without directly sharing their data (FedLLM). Following this training paradigm, the community has put massive efforts from diverse aspects including framework, performance, and privacy. However, an unpleasant fact is that there are currently no realistic datasets and benchmarks for FedLLM and previous works all rely on artificially constructed datasets, failing to capture properties in real-world scenarios. Addressing this, we propose FedLLM-Bench, which involves 8 training methods, 4 training datasets, and 6 evaluation metrics, to offer a comprehensive testbed for the FedLLM community. FedLLM-Bench encompasses three datasets (e.g., user-annotated multilingual dataset) for federated instruction tuning and one dataset (e.g., user-annotated preference dataset) for federated preference alignment, whose scale of client number ranges from 38 to 747. Our datasets incorporate several representative diversities: language, quality, quantity, instruction, length, embedding, and preference, capturing properties in real-world scenarios. Based on FedLLM-Bench, we conduct experiments on all datasets to benchmark existing FL methods and provide empirical insights (e.g., multilingual collaboration). We believe that our FedLLM-Bench can benefit the FedLLM community by reducing required efforts, providing a practical testbed, and promoting fair comparisons. Code and datasets are available at https://github.com/rui-ye/FedLLM-Bench., Comment: 22 pages

Due to the impressive capabilities of multimodal large language models (MLLMs), recent works have focused on employing MLLM-based agents for autonomous driving in large-scale and dynamic environments. However, prevalent approaches often directly translate high-level instructions into low-level vehicle control signals, which deviates from the inherent language generation paradigm of MLLMs and fails to fully harness their emergent powers. As a result, the generalizability of these methods is highly restricted by autonomous driving datasets used during fine-tuning. To tackle this challenge, we propose to connect high-level instructions and low-level control signals with mid-level language-driven commands, which are more fine-grained than high-level instructions but more universal and explainable than control signals, and thus can effectively bridge the gap in between. We implement this idea through a hierarchical multi-agent driving system named AD-H, including a MLLM planner for high-level reasoning and a lightweight controller for low-level execution. The hierarchical design liberates the MLLM from low-level control signal decoding and therefore fully releases their emergent capability in high-level perception, reasoning, and planning. We build a new dataset with action hierarchy annotations. Comprehensive closed-loop evaluations demonstrate several key advantages of our proposed AD-H system. First, AD-H can notably outperform state-of-the-art methods in achieving exceptional driving performance, even exhibiting self-correction capabilities during vehicle operation, a scenario not encountered in the training dataset. Second, AD-H demonstrates superior generalization under long-horizon instructions and novel environmental conditions, significantly surpassing current state-of-the-art methods. We will make our data and code publicly accessible at https://github.com/zhangzaibin/AD-H

The fundamental concept underlying topological phenomena posits the geometric phase associated with eigenstates. In contrast to this prevailing notion, theoretical studies on time-varying Hamiltonians allow for a new type of topological phenomenon, known as topological dynamics, where the evolution process allows a hidden topological invariant associated with continuous flows. To validate this conjecture, we study topological chiral and nonreciprocal dynamics by encircling the exceptional points (EPs) of non-Hermitian Hamiltonians in a trapped ion system. These dynamics are topologically robust against external perturbations even in the presence dissipation-induced nonadiabatic processes. Our findings indicate that they are protected by dynamical vorticity -- an emerging topological invariant associated with the energy dispersion of non-Hermitian band structures in a parallel transported eigenbasis. The symmetry breaking and other key features of topological dynamics are directly observed through quantum state tomography. Our results mark a significant step towards exploring topological properties of open quantum systems.

In this paper, we introduce a class of twisted matrix algebras of $M_2(E)$ and twisted direct products of $E\times E$ for an algebra $E$. Let $A$ be a noetherian Koszul Artin-Schelter regular algebra, $z\in A_2$ be a regular central element of $A$ and $B=A_P[y_1,y_2;\sigma]$ be a graded double Ore extension of $A$. We use the Clifford deformation $C_{A^!}(z)$ of Koszul dual $A^!$ to study the noncommutative quadric hypersurface $B/(z+y_1^2+y_2^2)$. We prove that the stable category of graded maximal Cohen-Macaulay modules over $B/(z+y_1^2+y_2^2)$ is equivalent to certain bounded derived categories, which involve a twisted matrix algebra of $M_2(C_{A^!}(z))$ or a twisted direct product of $C_{A^!}(z)\times C_{A^!}(z)$ depending on the values of $P$. These results are presented as skew versions of Kn\"orrer's periodicity theorem. Moreover, we show $B/(z+y_1^2+y_2^2)$ may not be a noncommutative graded isolated singularity even if $A/(z)$ is.

In recent years, there has been notable interest in investigating combinatorial optimization (CO) problems by neural-based framework. An emerging strategy to tackle these challenging problems involves the adoption of graph neural networks (GNNs) as an alternative to traditional algorithms, a subject that has attracted considerable attention. Despite the growing popularity of GNNs and traditional algorithm solvers in the realm of CO, there is limited research on their integrated use and the correlation between them within an end-to-end framework. The primary focus of our work is to formulate a more efficient and precise framework for CO by employing decision-focused learning on graphs. Additionally, we introduce a decision-focused framework that utilizes GNNs to address CO problems with auxiliary support. To realize an end-to-end approach, we have designed two cascaded modules: (a) an unsupervised trained graph predictive model, and (b) a solver for quadratic binary unconstrained optimization. Empirical evaluations are conducted on various classical tasks, including maximum cut, maximum independent set, and minimum vertex cover. The experimental results on classical CO problems (i.e. MaxCut, MIS, and MVC) demonstrate the superiority of our method over both the standalone GNN approach and classical methods., Comment: 9 pages

We establish a cosmological-model-independent method to extract the apparent magnitude and its derivative at different redshifts from the Pantheon+ type Ia supernova sample, and find that the obtained values deviate clearly from the prediction of the $\Lambda$CDM model at the lowest redshift. This deviation can be explained as a result of a transition of the absolute magnitude $M$ in the low redshift region. The observations seem to favor this transition since the minimum values of $\chi^2$ for two ansatzes of a varying $M$ are less than that of a constant $M$. The Hubble constant tension is alleviated from larger than $5\sigma$ to be about $1$ to $2\sigma$ for a varying $M$, and the growth tension can be resolved after attributing the variation of $M$ to a modification of the effective Newton's constant., Comment: 20 pages, 5 figures, 4 tables. Published in Physical Review D (Letter)

Recently, large-scale language-image generative models have gained widespread attention and many works have utilized generated data from these models to further enhance the performance of perception tasks. However, not all generated data can positively impact downstream models, and these methods do not thoroughly explore how to better select and utilize generated data. On the other hand, there is still a lack of research oriented towards active learning on generated data. In this paper, we explore how to perform active learning specifically for generated data in the long-tailed instance segmentation task. Subsequently, we propose BSGAL, a new algorithm that online estimates the contribution of the generated data based on gradient cache. BSGAL can handle unlimited generated data and complex downstream segmentation tasks effectively. Experiments show that BSGAL outperforms the baseline approach and effectually improves the performance of long-tailed segmentation. Our code can be found at https://github.com/aim-uofa/DiverGen., Comment: Accepted by ICML 2024

Linear complementary dual (LCD) codes can be used to against side-channel attacks and fault noninvasive attacks. Let $d_{a}(n,6)$ and $d_{l}(n,6)$ be the minimum weights of all binary optimal linear codes and LCD codes with length $n$ and dimension 6, respectively.In this article, we aim to obtain the values of $d_{l}(n,6)$ for $n\geq 51$ by investigating the nonexistence and constructions of LCD codes with given parameters. Suppose that $s \ge 0$ and $0\leq t\leq 62$ are two integers and $n=63s+t$. Using the theories of defining vectors, generalized anti-codes, reduced codes and nested codes, we exactly determine $d_{l}(n,6)$ for $t \notin\{21,22,25,26,33,34,37,38,45,46\}$, while we show that $d_{l}(n,6)\in$$\{d_{a}(n,6)$ $-1,d_{a}(n,6)\}$ for $t\in\{21,22,26,34,37,38,46\}$ and $ d_{l}(n,6)\in$$ \{d_{a}(n,6)-2,$ $d_{a}(n,6)-1\}$ for$t\in{25,33,45\}$., Comment: optimal linear code, LCD code,generalized anti-code, defining vector, reduced code

In recent years, graph neural networks (GNNs) have become increasingly popular for solving NP-hard combinatorial optimization (CO) problems, such as maximum cut and maximum independent set. The core idea behind these methods is to represent a CO problem as a graph and then use GNNs to learn the node/graph embedding with combinatorial information. Although these methods have achieved promising results, given a specific CO problem, the design of GNN architectures still requires heavy manual work with domain knowledge. Existing automated GNNs are mostly focused on traditional graph learning problems, which is inapplicable to solving NP-hard CO problems. To this end, we present a new class of \textbf{AUTO}mated \textbf{G}NNs for solving \textbf{NP}-hard problems, namely \textbf{AutoGNP}. We represent CO problems by GNNs and focus on two specific problems, i.e., mixed integer linear programming and quadratic unconstrained binary optimization. The idea of AutoGNP is to use graph neural architecture search algorithms to automatically find the best GNNs for a given NP-hard combinatorial optimization problem. Compared with existing graph neural architecture search algorithms, AutoGNP utilizes two-hop operators in the architecture search space. Moreover, AutoGNP utilizes simulated annealing and a strict early stopping policy to avoid local optimal solutions. Empirical results on benchmark combinatorial problems demonstrate the superiority of our proposed model., Comment: 9 pages

We introduce Xmodel-LM, a compact and efficient 1.1B language model pre-trained on around 2 trillion tokens. Trained on our self-built dataset (Xdata), which balances Chinese and English corpora based on downstream task optimization, Xmodel-LM exhibits remarkable performance despite its smaller size. It notably surpasses existing open-source language models of similar scale. Our model checkpoints and code are publicly accessible on GitHub at https://github.com/XiaoduoAILab/XmodelLM.

The increasing difficulty in accurately detecting forged images generated by AIGC(Artificial Intelligence Generative Content) poses many risks, necessitating the development of effective methods to identify and further locate forged areas. In this paper, to facilitate research efforts, we construct a DA-HFNet forged image dataset guided by text or image-assisted GAN and Diffusion model. Our goal is to utilize a hierarchical progressive network to capture forged artifacts at different scales for detection and localization. Specifically, it relies on a dual-attention mechanism to adaptively fuse multi-modal image features in depth, followed by a multi-branch interaction network to thoroughly interact image features at different scales and improve detector performance by leveraging dependencies between layers. Additionally, we extract more sensitive noise fingerprints to obtain more prominent forged artifact features in the forged areas. Extensive experiments validate the effectiveness of our approach, demonstrating significant performance improvements compared to state-of-the-art methods for forged image detection and localization.The code and dataset will be released in the future.

Contrastive Language-Image Pretraining (CLIP) has achieved remarkable success, leading to rapid advancements in multimodal studies. However, CLIP faces a notable challenge in terms of inefficient data utilization. It relies on a single contrastive supervision for each image-text pair during representation learning, disregarding a substantial amount of valuable information that could offer richer supervision. Additionally, the retention of non-informative tokens leads to increased computational demands and time costs, particularly in CLIP's ViT image encoder. To address these issues, we propose Multi-Perspective Language-Image Pretraining (MLIP). In MLIP, we leverage the frequency transform's sensitivity to both high and low-frequency variations, which complements the spatial domain's sensitivity limited to low-frequency variations only. By incorporating frequency transforms and token-level alignment, we expand CILP's single supervision into multi-domain and multi-level supervision, enabling a more thorough exploration of informative image features. Additionally, we introduce a token merging method guided by comprehensive semantics from the frequency and spatial domains. This allows us to merge tokens to multi-granularity tokens with a controllable compression rate to accelerate CLIP. Extensive experiments validate the effectiveness of our design., Comment: ICML 2024

Ponzi schemes, a form of scam, have been discovered in Ethereum smart contracts in recent years, causing massive financial losses. Existing detection methods primarily focus on rule-based approaches and machine learning techniques that utilize static information as features. However, these methods have significant limitations. Rule-based approaches rely on pre-defined rules with limited capabilities and domain knowledge dependency. Using static information like opcodes for machine learning fails to effectively characterize Ponzi contracts, resulting in poor reliability and interpretability. Moreover, relying on static information like transactions for machine learning requires a certain number of transactions to achieve detection, which limits the scalability of detection and hinders the identification of 0-day Ponzi schemes. In this paper, we propose PonziGuard, an efficient Ponzi scheme detection approach based on contract runtime behavior. Inspired by the observation that a contract's runtime behavior is more effective in disguising Ponzi contracts from the innocent contracts, PonziGuard establishes a comprehensive graph representation called contract runtime behavior graph (CRBG), to accurately depict the behavior of Ponzi contracts. Furthermore, it formulates the detection process as a graph classification task on CRBG, enhancing its overall effectiveness. The experiment results show that PonziGuard surpasses the current state-of-the-art approaches in the ground-truth dataset. We applied PonziGuard to Ethereum Mainnet and demonstrated its effectiveness in real-world scenarios. Using PonziGuard, we identified 805 Ponzi contracts on Ethereum Mainnet, which have resulted in an estimated economic loss of 281,700 Ether or approximately $500 million USD. We also found 0-day Ponzi schemes in the recently deployed 10,000 smart contracts., Comment: Submitted to ACM Transactions on Software Engineering and Methodology

Detecting out-of-distribution (OOD) samples is essential when deploying machine learning models in open-world scenarios. Zero-shot OOD detection, requiring no training on in-distribution (ID) data, has been possible with the advent of vision-language models like CLIP. Existing methods build a text-based classifier with only closed-set labels. However, this largely restricts the inherent capability of CLIP to recognize samples from large and open label space. In this paper, we propose to tackle this constraint by leveraging the expert knowledge and reasoning capability of large language models (LLM) to Envision potential Outlier Exposure, termed EOE, without access to any actual OOD data. Owing to better adaptation to open-world scenarios, EOE can be generalized to different tasks, including far, near, and fine-grained OOD detection. Technically, we design (1) LLM prompts based on visual similarity to generate potential outlier class labels specialized for OOD detection, as well as (2) a new score function based on potential outlier penalty to distinguish hard OOD samples effectively. Empirically, EOE achieves state-of-the-art performance across different OOD tasks and can be effectively scaled to the ImageNet-1K dataset. The code is publicly available at: https://github.com/tmlr-group/EOE., Comment: ICML 2024

In hedge funds, convex compensation schemes are adopted to stimulate a high-profit performance for portfolio managers. In economics, non-monotone risk aversion is proposed to argue that individuals may not be risk-averse when the wealth level is low. Combining these two ingredients, we study the optimal control strategy of the manager in incomplete markets. Generally, we propose a wide family of utility functions, the piecewise symmetric asymptotic hyperbolic absolute risk aversion (PSAHARA) utility, to model the two ingredients, containing both non-concavity and non-differentiability as some abnormalities. Technically, we propose an additional assumption and prove concavification techniques of non--concave utility functions with a left unbounded domain in incomplete markets. Next, we derive an explicit optimal control for the family of PSAHARA utilities. This control is expressed into a unified four-term structure, featuring the asymptotic Merton term. Furthermore, we provide a detailed asymptotic analysis and numerical illustration of the optimal portfolio. We obtain several key insights, including that the convex compensation still induces a great risk-taking behavior in the case that the preference is modeled by SAHARA utility. Finally, we conduct a real-data analysis of the U.S. stock market under the above model and conclude that the PSAHARA portfolio is very risk-seeking and leads to a high return and a high volatility.

Recent advancements in Latent Diffusion Models (LDMs) have propelled them to the forefront of various generative tasks. However, their iterative sampling process poses a significant computational burden, resulting in slow generation speeds and limiting their application in text-to-audio generation deployment. In this work, we introduce AudioLCM, a novel consistency-based model tailored for efficient and high-quality text-to-audio generation. AudioLCM integrates Consistency Models into the generation process, facilitating rapid inference through a mapping from any point at any time step to the trajectory's initial point. To overcome the convergence issue inherent in LDMs with reduced sample iterations, we propose the Guided Latent Consistency Distillation with a multi-step Ordinary Differential Equation (ODE) solver. This innovation shortens the time schedule from thousands to dozens of steps while maintaining sample quality, thereby achieving fast convergence and high-quality generation. Furthermore, to optimize the performance of transformer-based neural network architectures, we integrate the advanced techniques pioneered by LLaMA into the foundational framework of transformers. This architecture supports stable and efficient training, ensuring robust performance in text-to-audio synthesis. Experimental results on text-to-sound generation and text-to-music synthesis tasks demonstrate that AudioLCM needs only 2 iterations to synthesize high-fidelity audios, while it maintains sample quality competitive with state-of-the-art models using hundreds of steps. AudioLCM enables a sampling speed of 333x faster than real-time on a single NVIDIA 4090Ti GPU, making generative models practically applicable to text-to-audio generation deployment. Our extensive preliminary analysis shows that each design in AudioLCM is effective.

Large language models (LLMs) are being rapidly developed, and a key component of their widespread deployment is their safety-related alignment. Many red-teaming efforts aim to jailbreak LLMs, where among these efforts, the Greedy Coordinate Gradient (GCG) attack's success has led to a growing interest in the study of optimization-based jailbreaking techniques. Although GCG is a significant milestone, its attacking efficiency remains unsatisfactory. In this paper, we present several improved (empirical) techniques for optimization-based jailbreaks like GCG. We first observe that the single target template of "Sure" largely limits the attacking performance of GCG; given this, we propose to apply diverse target templates containing harmful self-suggestion and/or guidance to mislead LLMs. Besides, from the optimization aspects, we propose an automatic multi-coordinate updating strategy in GCG (i.e., adaptively deciding how many tokens to replace in each step) to accelerate convergence, as well as tricks like easy-to-hard initialisation. Then, we combine these improved technologies to develop an efficient jailbreak method, dubbed I-GCG. In our experiments, we evaluate on a series of benchmarks (such as NeurIPS 2023 Red Teaming Track). The results demonstrate that our improved techniques can help GCG outperform state-of-the-art jailbreaking attacks and achieve nearly 100% attack success rate. The code is released at https://github.com/jiaxiaojunQAQ/I-GCG.

The widespread use of diffusion methods enables the creation of highly realistic images on demand, thereby posing significant risks to the integrity and safety of online information and highlighting the necessity of DeepFake detection. Our analysis of features extracted by traditional image encoders reveals that both low-level and high-level features offer distinct advantages in identifying DeepFake images produced by various diffusion methods. Inspired by this finding, we aim to develop an effective representation that captures both low-level and high-level features to detect diffusion-based DeepFakes. To address the problem, we propose a text modality-oriented feature extraction method, termed TOFE. Specifically, for a given target image, the representation we discovered is a corresponding text embedding that can guide the generation of the target image with a specific text-to-image model. Experiments conducted across ten diffusion types demonstrate the efficacy of our proposed method.

Dynamic Scene Graph Generation (DSGG) focuses on identifying visual relationships within the spatial-temporal domain of videos. Conventional approaches often employ multi-stage pipelines, which typically consist of object detection, temporal association, and multi-relation classification. However, these methods exhibit inherent limitations due to the separation of multiple stages, and independent optimization of these sub-problems may yield sub-optimal solutions. To remedy these limitations, we propose a one-stage end-to-end framework, termed OED, which streamlines the DSGG pipeline. This framework reformulates the task as a set prediction problem and leverages pair-wise features to represent each subject-object pair within the scene graph. Moreover, another challenge of DSGG is capturing temporal dependencies, we introduce a Progressively Refined Module (PRM) for aggregating temporal context without the constraints of additional trackers or handcrafted trajectories, enabling end-to-end optimization of the network. Extensive experiments conducted on the Action Genome benchmark demonstrate the effectiveness of our design. The code and models are available at \url{https://github.com/guanw-pku/OED}., Comment: Accepted by CVPR'24

Classical Chinese Understanding (CCU) holds significant value in preserving and exploration of the outstanding traditional Chinese culture. Recently, researchers have attempted to leverage the potential of Large Language Models (LLMs) for CCU by capitalizing on their remarkable comprehension and semantic capabilities. However, no comprehensive benchmark is available to assess the CCU capabilities of LLMs. To fill this gap, this paper introduces C$^{3}$bench, a Comprehensive Classical Chinese understanding benchmark, which comprises 50,000 text pairs for five primary CCU tasks, including classification, retrieval, named entity recognition, punctuation, and translation. Furthermore, the data in C$^{3}$bench originates from ten different domains, covering most of the categories in classical Chinese. Leveraging the proposed C$^{3}$bench, we extensively evaluate the quantitative performance of 15 representative LLMs on all five CCU tasks. Our results not only establish a public leaderboard of LLMs' CCU capabilities but also gain some findings. Specifically, existing LLMs are struggle with CCU tasks and still inferior to supervised models. Additionally, the results indicate that CCU is a task that requires special attention. We believe this study could provide a standard benchmark, comprehensive baselines, and valuable insights for the future advancement of LLM-based CCU research. The evaluation pipeline and dataset are available at \url{https://github.com/SCUT-DLVCLab/C3bench}.

Retrosynthesis is a fundamental but challenging task in organic chemistry, with broad applications in fields such as drug design and synthesis. Given a target molecule, the goal of retrosynthesis is to find out a series of reactions which could be assembled into a synthetic route which starts from purchasable molecules and ends at the target molecule. The uncertainty of reactions used in retrosynthetic planning, which is caused by hallucinations of backward models, has recently been noticed. In this paper we propose a succinct probabilistic model to describe such uncertainty. Based on the model, we propose a new retrosynthesis planning algorithm called retro-prob to maximize the successful synthesis probability of target molecules, which acquires high efficiency by utilizing the chain rule of derivatives. Experiments on the Paroutes benchmark show that retro-prob outperforms previous algorithms, retro* and retro-fallback, both in speed and in the quality of synthesis plans.

Standard common factor models, such as the linear normal factor model, rely on strict parametric assumptions, which require rigorous model-data fit assessment to prevent fallacious inferences. However, overall goodness-of-fit diagnostics conventionally used in factor analysis do not offer diagnostic information on where the misfit originates. In the current work, we propose a new fit assessment framework for common factor models by extending the theory of generalized residuals (Haberman & Sinharay, 2013). This framework allows for the flexible adaptation of test statistics to identify various sources of misfit. In addition, the resulting goodness-of-fit tests provide more informative diagnostics, as the evaluation is performed conditionally on latent variables. Several examples of test statistics suitable for assessing various model assumptions are presented within this framework, and their performance is evaluated by simulation studies and a real data example., Comment: 40 pages, 12 figures

With the rising popularity of Large Language Models (LLMs), assessing their trustworthiness through security tasks has gained critical importance. Regarding the new task of universal goal hijacking, previous efforts have concentrated solely on optimization algorithms, overlooking the crucial role of the prompt. To fill this gap, we propose a universal goal hijacking method called POUGH that incorporates semantic-guided prompt processing strategies. Specifically, the method starts with a sampling strategy to select representative prompts from a candidate pool, followed by a ranking strategy that prioritizes the prompts. Once the prompts are organized sequentially, the method employs an iterative optimization algorithm to generate the universal fixed suffix for the prompts. Experiments conducted on four popular LLMs and ten types of target responses verified the effectiveness of our method., Comment: 15 pages

To improve performance in contemporary deep learning, one is interested in scaling up the neural network in terms of both the number and the size of the layers. When ramping up the width of a single layer, graceful scaling of training has been linked to the need to normalize the weights and their updates in the "natural norm" particular to that layer. In this paper, we significantly generalize this idea by defining the modular norm, which is the natural norm on the full weight space of any neural network architecture. The modular norm is defined recursively in tandem with the network architecture itself. We show that the modular norm has several promising applications. On the practical side, the modular norm can be used to normalize the updates of any base optimizer so that the learning rate becomes transferable across width and depth. This means that the user does not need to compute optimizer-specific scale factors in order to scale training. On the theoretical side, we show that for any neural network built from "well-behaved" atomic modules, the gradient of the network is Lipschitz-continuous in the modular norm, with the Lipschitz constant admitting a simple recursive formula. This characterization opens the door to porting standard ideas in optimization theory over to deep learning. We have created a Python package called Modula that automatically normalizes weight updates in the modular norm of the architecture. The package is available via "pip install modula" with source code at https://github.com/jxbz/modula.

Text-to-image diffusion models have achieved tremendous success in the field of controllable image generation, while also coming along with issues of privacy leakage and data copyrights. Membership inference arises in these contexts as a potential auditing method for detecting unauthorized data usage. While some efforts have been made on diffusion models, they are not applicable to text-to-image diffusion models due to the high computation overhead and enhanced generalization capabilities. In this paper, we first identify a conditional overfitting phenomenon in text-to-image diffusion models, indicating that these models tend to overfit the conditional distribution of images given the corresponding text rather than the marginal distribution of images only. Based on this observation, we derive an analytical indicator, namely Conditional Likelihood Discrepancy (CLiD), to perform membership inference, which reduces the stochasticity in estimating memorization of individual samples. Experimental results demonstrate that our method significantly outperforms previous methods across various data distributions and dataset scales. Additionally, our method shows superior resistance to overfitting mitigation strategies, such as early stopping and data augmentation., Comment: 18 pages, 5 figures. NeurIPS 2024. Code will be released at: https://github.com/zhaisf/CLiD

Planetary rovers can use onboard data analysis to adapt their measurement plan on the fly, improving the science value of data collected between commands from Earth. This paper describes the implementation of an adaptive sampling algorithm used by PIXL, the X-ray fluorescence spectrometer of the Mars 2020 Perseverance rover. PIXL is deployed using the rover arm to measure X-ray spectra of rocks with a scan density of several thousand points over an area of typically 5 x 7 mm. The adaptive sampling algorithm is programmed to recognize points of interest and to increase the signal-to-noise ratio at those locations by performing longer integrations. Two approaches are used to formulate the sampling rules based on past quantification data: 1) Expressions that isolate particular regions within a ternary compositional diagram, and 2) Machine learning rules that threshold for a high weight percent of particular compounds. The design of the rulesets are outlined and the performance of the algorithm is quantified using measurements from the surface of Mars. To our knowledge, PIXL's adaptive sampling represents the first autonomous decision-making based on real-time compositional analysis by a spacecraft on the surface of another planet., Comment: 24 pages including 11 figures and 7 tables. Submitted for publication to the journal Icarus

The CODEX-$\beta$ apparatus is a demonstrator for the proposed future CODEX-b experiment, a long-lived-particle detector foreseen for operation at IP8 during HL-LHC data-taking. The demonstrator project, intended to collect data in 2025, is described, with a particular focus on the design, construction, and installation of the new apparatus.

Electrons residing in flat-band system can play a vital role in triggering spectacular phenomenology due to relatively large interactions and spontaneous breaking of different degeneracies. In this work we demonstrate chirally twisted triple bilayer graphene, a new moir\'e structure formed by three pieces of helically stacked Bernal bilayer graphene, as a highly tunable flat-band system. In addition to the correlated insulators showing at integer moir\'e fillings, commonly attributed to interaction induced symmetry broken isospin flavors in graphene, we observe abundant insulating states at half-integer moir\'e fillings, suggesting a longer-range interaction and the formation of charge density wave insulators which spontaneously break the moir\'e translation symmetry. With weak out-of-plane magnetic field applied, as observed half-integer filling states are enhanced and more quarter-integer filling states appear, pointing towards further quadrupling moir\'e unit cells. The insulating states at fractional fillings combined with Hartree-Fock calculations demonstrate the observation of a new type of correlated charge density wave insulators in graphene and points to a new accessible twist manner engineering correlated moir\'e electronics.

Recent prevailing works on graph machine learning typically follow a similar methodology that involves designing advanced variants of graph neural networks (GNNs) to maintain the superior performance of GNNs on different graphs. In this paper, we aim to streamline the GNN design process and leverage the advantages of Large Language Models (LLMs) to improve the performance of GNNs on downstream tasks. We formulate a new paradigm, coined "LLMs-as-Consultants," which integrates LLMs with GNNs in an interactive manner. A framework named LOGIN (LLM Consulted GNN training) is instantiated, empowering the interactive utilization of LLMs within the GNN training process. First, we attentively craft concise prompts for spotted nodes, carrying comprehensive semantic and topological information, and serving as input to LLMs. Second, we refine GNNs by devising a complementary coping mechanism that utilizes the responses from LLMs, depending on their correctness. We empirically evaluate the effectiveness of LOGIN on node classification tasks across both homophilic and heterophilic graphs. The results illustrate that even basic GNN architectures, when employed within the proposed LLMs-as-Consultants paradigm, can achieve comparable performance to advanced GNNs with intricate designs. Our codes are available at https://github.com/QiaoYRan/LOGIN.

In this paper, we study an under-explored but important factor of diffusion generative models, i.e., the combinatorial complexity. Data samples are generally high-dimensional, and for various structured generation tasks, there are additional attributes which are combined to associate with data samples. We show that the space spanned by the combination of dimensions and attributes is insufficiently sampled by existing training scheme of diffusion generative models, causing degraded test time performance. We present a simple fix to this problem by constructing stochastic processes that fully exploit the combinatorial structures, hence the name ComboStoc. Using this simple strategy, we show that network training is significantly accelerated across diverse data modalities, including images and 3D structured shapes. Moreover, ComboStoc enables a new way of test time generation which uses insynchronized time steps for different dimensions and attributes, thus allowing for varying degrees of control over them.

Recently, a new kind of collinear magnetism, dubbed altermagnetism, has attracted considerable interests. A key characteristic of altermagnet is the momentum-dependent band and spin splitting without net magnetization. However, finding altermagnetic materials with large splitting near the Fermi level, which necessarily requires three-dimensional k-space mapping and is crucial for spintronic applications and emergent phenomena, remains challenging. Here by employing synchrotron-based angle-resolved photoemission spectroscopy (ARPES), spin-resolved ARPES and model calculations, we uncover a large altermagnetic splitting, up to ~1.0 eV, near the Fermi level in CrSb. We verify its bulk-type g-wave altermagnetism through systematic three-dimensional k-space mapping, which unambiguously reveals the altermagnetic symmetry and associated nodal planes. The ARPES results are well captured by density functional theory calculations. Spin-resolved ARPES measurements further verify the spin polarizations of the split bands near Fermi level. In addition, tight-binding model analysis indicates that the large altermagnetic splitting arises from strong third-nearest-neighbor hopping mediated by Sb ions, which breaks both the space-time reversal symmetry and the translational spin-rotation symmetry. The large band/spin splitting near Fermi level in metallic CrSb, together with its high TN (up to 705 K) and simple spin configuration, paves the way for exploring emergent phenomena and spintronic applications based on altermagnets., Comment: The manuscript contains 19 pages, 5 figures and 1 table. The supplementary information contains 15 pages, 13 figures and 5 tables