arXiv subjects

Jun Liu

Publications and source records attributed to Jun Liu.

At least 433 records · Page 24Linked to original sources

arXiv2024

Learning Regions of Attraction in Unknown Dynamical Systems via Zubov-Koopman Lifting: Regularities and Convergence

The estimation for the region of attraction (ROA) of an asymptotically stable equilibrium point is crucial in the analysis of nonlinear systems. There has been a recent surge of interest in estimating the solution to Zubov's equation, whose non-trivial sub-level sets form the exact ROA. In this paper, we propose a lifting approach to map observable data into an infinite-dimensional function space, which generates a flow governed by the proposed `Zubov-Koopman' operators. By learning a Zubov-Koopman operator over a fixed time interval, we can indirectly approximate the solution to Zubov's equation through iterative application of the learned operator on certain functions. We also demonstrate that a transformation of such an approximator can be readily utilized as a near-maximal Lyapunov function. We approach our goal through a comprehensive investigation of the regularities of Zubov-Koopman operators and their associated quantities. Based on these findings, we present an algorithm for learning Zubov-Koopman operators that exhibit strong convergence to the true operator. We show that this approach reduces the amount of required data and can yield desirable estimation results, as demonstrated through numerical examples.

math.DS↗

arXiv2024

Poisoned LangChain: Jailbreak LLMs by LangChain

With the development of natural language processing (NLP), large language models (LLMs) are becoming increasingly popular. LLMs are integrating more into everyday life, raising public concerns about their security vulnerabilities. Consequently, the security of large language models is becoming critically important. Currently, the techniques for attacking and defending against LLMs are continuously evolving. One significant method type of attack is the jailbreak attack, which designed to evade model safety mechanisms and induce the generation of inappropriate content. Existing jailbreak attacks primarily rely on crafting inducement prompts for direct jailbreaks, which are less effective against large models with robust filtering and high comprehension abilities. Given the increasing demand for real-time capabilities in large language models, real-time updates and iterations of new knowledge have become essential. Retrieval-Augmented Generation (RAG), an advanced technique to compensate for the model's lack of new knowledge, is gradually becoming mainstream. As RAG enables the model to utilize external knowledge bases, it provides a new avenue for jailbreak attacks. In this paper, we conduct the first work to propose the concept of indirect jailbreak and achieve Retrieval-Augmented Generation via LangChain. Building on this, we further design a novel method of indirect jailbreak attack, termed Poisoned-LangChain (PLC), which leverages a poisoned external knowledge base to interact with large language models, thereby causing the large models to generate malicious non-compliant dialogues.We tested this method on six different large language models across three major categories of jailbreak issues. The experiments demonstrate that PLC successfully implemented indirect jailbreak attacks under three different scenarios, achieving success rates of 88.56%, 79.04%, and 82.69% respectively.

cs.CL↗

arXiv2024

Interactive Evolution: A Neural-Symbolic Self-Training Framework For Large Language Models

One of the primary driving forces contributing to the superior performance of Large Language Models (LLMs) is the extensive availability of human-annotated natural language data, which is used for alignment fine-tuning. This inspired researchers to investigate self-training methods to mitigate the extensive reliance on human annotations. However, the current success of self-training has been primarily observed in natural language scenarios, rather than in the increasingly important neural-symbolic scenarios. To this end, we propose an environment-guided neural-symbolic self-training framework named ENVISIONS. It aims to overcome two main challenges: (1) the scarcity of symbolic data, and (2) the limited proficiency of LLMs in processing symbolic language. Extensive evaluations conducted on three distinct domains demonstrate the effectiveness of our approach. Additionally, we have conducted a comprehensive analysis to uncover the factors contributing to ENVISIONS's success, thereby offering valuable insights for future research in this area. Code will be available at \url{https://github.com/xufangzhi/ENVISIONS}.

cs.CL↗

arXiv2024

Instability of Self-Driving Satellite Mega-Constellation: From Theory to Practical Impacts on Network Lifetime and Capacity

Low Earth Orbit (LEO) satellite mega-constellations aim to enable high-speed Internet for numerous users anywhere on Earth. To safeguard their network infrastructure in congested outer space, they perform automatic orbital maneuvers to avoid collisions with external debris and satellites. However, our control-theoretic analysis and empirical validation using Starlink's space situational awareness datasets discover that, these safety-oriented maneuvers themselves can threaten safety and networking via cascaded collision avoidance inside the mega-constellation. This domino effect forces a dilemma between long-term LEO network lifetime and short-term LEO network capacity. Its root cause is that, the decades-old local pairwise maneuver paradigm for standalone satellites is inherently unstable if scaled out to recent mega-constellation networks. We thus propose an alternative bilateral maneuver control that stabilizes self-driving mega-constellations for concurrent network lifetime and capacity boosts. Our operational trace-driven emulation shows a 8$\times$ network lifetime extension in Starlink without limiting its network capacity.

cs.NI↗

arXiv2024

Correlation Matching Transformation Transformers for UHD Image Restoration

This paper proposes UHDformer, a general Transformer for Ultra-High-Definition (UHD) image restoration. UHDformer contains two learning spaces: (a) learning in high-resolution space and (b) learning in low-resolution space. The former learns multi-level high-resolution features and fuses low-high features and reconstructs the residual images, while the latter explores more representative features learning from the high-resolution ones to facilitate better restoration. To better improve feature representation in low-resolution space, we propose to build feature transformation from the high-resolution space to the low-resolution one. To that end, we propose two new modules: Dual-path Correlation Matching Transformation module (DualCMT) and Adaptive Channel Modulator (ACM). The DualCMT selects top C/r (r is greater or equal to 1 which controls the squeezing level) correlation channels from the max-pooling/mean-pooling high-resolution features to replace low-resolution ones in Transformers, which can effectively squeeze useless content to improve the feature representation in low-resolution space to facilitate better recovery. The ACM is exploited to adaptively modulate multi-level high-resolution features, enabling to provide more useful features to low-resolution space for better learning. Experimental results show that our UHDformer reduces about ninety-seven percent model sizes compared with most state-of-the-art methods while significantly improving performance under different training sets on 3 UHD image restoration tasks, including low-light image enhancement, image dehazing, and image deblurring. The source codes will be made available at https://github.com/supersupercong/UHDformer.

cs.CV↗

arXiv2024

Reach-Avoid Control Synthesis for a Quadrotor UAV with Formal Safety Guarantees

Reach-avoid specifications are one of the most common tasks in autonomous aerial vehicle (UAV) applications. Despite the intensive research and development associated with control of aerial vehicles, generating feasible trajectories though complex environments and tracking them with formal safety guarantees remain challenging. In this paper, we propose a control framework for a quadrotor UAV that enables accomplishing reach-avoid tasks with formal safety guarantees. In this proposed framework, we integrate geometric control theory for tracking and polynomial trajectory generation using Bezier curves, where tracking errors are accounted for in the trajectory synthesis process. To estimate the tracking errors, we revisit the stability analysis of the closed-loop quadrotor system, when geometric control is implemented. We show that the tracking error dynamics exhibit local exponential stability when geometric control is implemented with any positive control gains, and we derive tight uniform bounds of the tracking error. We also introduce sufficient conditions to be imposed on the desired trajectory utilizing the derived uniform bounds to ensure the well-definedness of the closed-loop system. For the trajectory synthesis, we present an efficient algorithm that enables constructing a safe tube by means of sampling-based planning and safe hyper-rectangular set computations. Then, we compute the trajectory, given as a piecewise continuous Bezier curve, through the safe tube, where a heuristic efficient approach that utilizes iterative linear programming is employed. We present extensive numerical simulations with a cluttered environment to illustrate the effectiveness of the proposed framework in reach-avoid planning scenarios.

eess.SY↗

arXiv2024

PathReasoner: Modeling Reasoning Path with Equivalent Extension for Logical Question Answering

Logical reasoning task has attracted great interest since it was proposed. Faced with such a task, current competitive models, even large language models (e.g., ChatGPT and PaLM 2), still perform badly. Previous promising LMs struggle in logical consistency modeling and logical structure perception. To this end, we model the logical reasoning task by transforming each logical sample into reasoning paths and propose an architecture \textbf{PathReasoner}. It addresses the task from the views of both data and model. To expand the diversity of the logical samples, we propose an atom extension strategy supported by equivalent logical formulas, to form new reasoning paths. From the model perspective, we design a stack of transformer-style blocks. In particular, we propose a path-attention module to joint model in-atom and cross-atom relations with the high-order diffusion strategy. Experiments show that PathReasoner achieves competitive performances on two logical reasoning benchmarks and great generalization abilities.

cs.CL↗

arXiv2024

LTGC: Long-tail Recognition via Leveraging LLMs-driven Generated Content

Long-tail recognition is challenging because it requires the model to learn good representations from tail categories and address imbalances across all categories. In this paper, we propose a novel generative and fine-tuning framework, LTGC, to handle long-tail recognition via leveraging generated content. Firstly, inspired by the rich implicit knowledge in large-scale models (e.g., large language models, LLMs), LTGC leverages the power of these models to parse and reason over the original tail data to produce diverse tail-class content. We then propose several novel designs for LTGC to ensure the quality of the generated data and to efficiently fine-tune the model using both the generated and original data. The visualization demonstrates the effectiveness of the generation module in LTGC, which produces accurate and diverse tail data. Additionally, the experimental results demonstrate that our LTGC outperforms existing state-of-the-art methods on popular long-tailed benchmarks.

cs.CV↗

arXiv2024

UPAM: Unified Prompt Attack in Text-to-Image Generation Models Against Both Textual Filters and Visual Checkers

Text-to-Image (T2I) models have raised security concerns due to their potential to generate inappropriate or harmful images. In this paper, we propose UPAM, a novel framework that investigates the robustness of T2I models from the attack perspective. Unlike most existing attack methods that focus on deceiving textual defenses, UPAM aims to deceive both textual and visual defenses in T2I models. UPAM enables gradient-based optimization, offering greater effectiveness and efficiency than previous methods. Given that T2I models might not return results due to defense mechanisms, we introduce a Sphere-Probing Learning (SPL) scheme to support gradient optimization even when no results are returned. Additionally, we devise a Semantic-Enhancing Learning (SEL) scheme to finetune UPAM for generating target-aligned images. Our framework also ensures attack stealthiness. Extensive experiments demonstrate UPAM's effectiveness and efficiency.

cs.CV↗

arXiv2024

Off-the-shelf ChatGPT is a Good Few-shot Human Motion Predictor

To facilitate the application of motion prediction in practice, recently, the few-shot motion prediction task has attracted increasing research attention. Yet, in existing few-shot motion prediction works, a specific model that is dedicatedly trained over human motions is generally required. In this work, rather than tackling this task through training a specific human motion prediction model, we instead propose a novel FMP-OC framework. In FMP-OC, in a totally training-free manner, we enable Few-shot Motion Prediction, which is a non-language task, to be performed directly via utilizing the Off-the-shelf language model ChatGPT. Specifically, to lead ChatGPT as a language model to become an accurate motion predictor, in FMP-OC, we first introduce several novel designs to facilitate extracting implicit knowledge from ChatGPT. Moreover, we also incorporate our framework with a motion-in-context learning mechanism. Extensive experiments demonstrate the efficacy of our proposed framework.

cs.CV↗

arXiv2024

Large Language Models can Deliver Accurate and Interpretable Time Series Anomaly Detection

Time series anomaly detection (TSAD) plays a crucial role in various industries by identifying atypical patterns that deviate from standard trends, thereby maintaining system integrity and enabling prompt response measures. Traditional TSAD models, which often rely on deep learning, require extensive training data and operate as black boxes, lacking interpretability for detected anomalies. To address these challenges, we propose LLMAD, a novel TSAD method that employs Large Language Models (LLMs) to deliver accurate and interpretable TSAD results. LLMAD innovatively applies LLMs for in-context anomaly detection by retrieving both positive and negative similar time series segments, significantly enhancing LLMs' effectiveness. Furthermore, LLMAD employs the Anomaly Detection Chain-of-Thought (AnoCoT) approach to mimic expert logic for its decision-making process. This method further enhances its performance and enables LLMAD to provide explanations for their detections through versatile perspectives, which are particularly important for user decision-making. Experiments on three datasets indicate that our LLMAD achieves detection performance comparable to state-of-the-art deep learning methods while offering remarkable interpretability for detections. To the best of our knowledge, this is the first work that directly employs LLMs for TSAD.

cs.CL↗

arXiv2024

Talbot-like pattern evolution in complex structured light from unitary transformation

Astigmatic unitary transformations allow for the adiabatic connections of all feasible states of paraxial Gaussian beams on the same modal sphere, i.e., Hermite-Laguerre-Gaussian (HLG) modes. Here, we present a comprehensive investigation into the unitary modal evolution of complex structured Gaussian beams, comprised by HLG modes from disparate modal spheres, via astigmatic transformation. The non-synchronized higher-order geometric phases in cyclic transformations originates a Talbot-effect-like modal evolution in the superposition state of these HLG modes, resulting in pattern variations and revivals in transformations with specific geodesic loops. Using Ince-Gaussian modes as an illustrative example, we systematically analyze and experimentally corroborate the beamforming mechanism behind the pattern evolution. Our results outline a generic modal conversion theory of structured Gaussian beams via astigmatic unitary transformation, offering a new approach for shaping spatial modal structure. These findings may inspire a wide variety of applications based on structured light.

physics.optics↗

arXiv2024

Parallel-in-Time Iterative Methods for Pricing American Options

For pricing American options, %after suitable discretization in space and time, a sequence of discrete linear complementarity problems (LCPs) or equivalently Hamilton-Jacobi-Bellman (HJB) equations need to be solved in a sequential time-stepping manner. In each time step, the policy iteration or its penalty variant is often applied due to their fast convergence rates. In this paper, we aim to solve for all time steps simultaneously, by applying the policy iteration to an ``all-at-once form" of the HJB equations, where two different parallel-in-time preconditioners are proposed to accelerate the solution of the linear systems within the policy iteration. Our proposed methods are generally applicable for such all-at-once forms of the HJB equation, arising from option pricing problems with optimal stopping and nontrivial underlying asset models. Numerical examples are presented to show the feasibility and robust convergence behavior of the proposed methodology.

math.NA↗

arXiv2024

A regularized eigenmatrix method for unstructured sparse recovery

The recently developed data-driven eigenmatrix method shows very promising reconstruction accuracy in sparse recovery for a wide range of kernel functions and random sample locations. However, its current implementation can lead to numerical instability if the threshold tolerance is not appropriately chosen. To incorporate regularization techniques, we propose to regularize the eigenmatrix method by replacing the computation of an ill-conditioned pseudo-inverse by the solution of an ill-conditioned least square system, which can be efficiently treated by Tikhonov regularization. Extensive numerical examples confirmed the improved effectiveness of our proposed method, especially when the noise levels are relatively high.

math.NA↗

arXiv2024

Untargeted Adversarial Attack on Knowledge Graph Embeddings

Knowledge graph embedding (KGE) methods have achieved great success in handling various knowledge graph (KG) downstream tasks. However, KGE methods may learn biased representations on low-quality KGs that are prevalent in the real world. Some recent studies propose adversarial attacks to investigate the vulnerabilities of KGE methods, but their attackers are target-oriented with the KGE method and the target triples to predict are given in advance, which lacks practicability. In this work, we explore untargeted attacks with the aim of reducing the global performances of KGE methods over a set of unknown test triples and conducting systematic analyses on KGE robustness. Considering logic rules can effectively summarize the global structure of a KG, we develop rule-based attack strategies to enhance the attack efficiency. In particular,we consider adversarial deletion which learns rules, applying the rules to score triple importance and delete important triples, and adversarial addition which corrupts the learned rules and applies them for negative triples as perturbations. Extensive experiments on two datasets over three representative classes of KGE methods demonstrate the effectiveness of our proposed untargeted attacks in diminishing the link prediction results. And we also find that different KGE methods exhibit different robustness to untargeted attacks. For example, the robustness of methods engaged with graph neural networks and logic rules depends on the density of the graph. But rule-based methods like NCRL are easily affected by adversarial addition attacks to capture negative rules

cs.LG↗

arXiv2024

A characterization of wavelet sets on Vilenkin groups with its application to construction of MRA wavelets

Let $G$ be a Vilenkin group. In 2008, Y. A. Farkov constructed wavelets on $G$ via the multiresolution analysis method. In this article, a characterization of wavelet sets on $G$ is established, which provides another method for the construction of wavelets. As an application, the relation between multiresolution analyses and wavelets determined from wavelet sets is also presented. To some extent, these results positively answer a question mentioned by P. Mahapatra and D. Singh in [Bull. Sci. Math. 167 (2021), Paper No. 102945, 20 pp].

math.CA↗

arXiv2024

Towards enhancing quantum expectation estimation of matrices through partial Pauli decomposition techniques and post-processing

We introduce an approach for estimating the expectation values of arbitrary $n$-qubit matrices $M \in \mathbb{C}^{2^n\times 2^n}$ on a quantum computer. In contrast to conventional methods like the Pauli decomposition that utilize $4^n$ distinct quantum circuits for this task, our technique employs at most $2^n$ unique circuits, with even fewer required for matrices with limited bandwidth. Termed the \textit{partial Pauli decomposition}, our method involves observables formed as the Kronecker product of a single-qubit Pauli operator and orthogonal projections onto the computational basis. By measuring each such observable, one can simultaneously glean information about $2^n$ distinct entries of $M$ through post-processing of the measurement counts. This reduction in quantum resources is especially crucial in the current noisy intermediate-scale quantum era, offering the potential to accelerate quantum algorithms that rely heavily on expectation estimation, such as the variational quantum eigensolver and the quantum approximate optimization algorithm.

quant-ph↗

arXiv2024

Nyonic Technical Report

This report details the development and key achievements of our latest language model designed for custom large language models. The advancements introduced include a novel Online Data Scheduler that supports flexible training data adjustments and curriculum learning. The model's architecture is fortified with state-of-the-art techniques such as Rotary Positional Embeddings, QK-LayerNorm, and a specially crafted multilingual tokenizer to enhance stability and performance. Moreover, our robust training framework incorporates advanced monitoring and rapid recovery features to ensure optimal efficiency. Our Wonton 7B model has demonstrated competitive performance on a range of multilingual and English benchmarks. Future developments will prioritize narrowing the performance gap with more extensively trained models, thereby enhancing the model's real-world efficacy and adaptability.GitHub: \url{https://github.com/nyonicai/nyonic-public}

cs.CL↗