arXiv subjects

Bin Wu

Publications and source records attributed to Bin Wu.

At least 109 records · Page 6Linked to original sources

arXiv2025

Carleman estimate for full-discrete approximations of the complex Ginzburg-Landau equation with dynamic boundary conditions and applications to controllability

In this paper, we investigate Carleman estimate and controllability result for the fully-discrete approximations of a one-dimensional Ginzburg-Landau equation with dynamic boundary conditions. We first establish a new discrete Carleman estimate for the corresponding adjoint system. Based on this Carleman estimate, we obtain a relaxed observability inequality for the adjoint system, and then a controllability result for the fully-discrete Ginzburg-Landau equation with dynamic boundary conditions.

math.AP↗

arXiv2025

Synthetic Data is an Elegant GIFT for Continual Vision-Language Models

Pre-trained Vision-Language Models (VLMs) require Continual Learning (CL) to efficiently update their knowledge and adapt to various downstream tasks without retraining from scratch. However, for VLMs, in addition to the loss of knowledge previously learned from downstream tasks, pre-training knowledge is also corrupted during continual fine-tuning. This issue is exacerbated by the unavailability of original pre-training data, leaving VLM's generalization ability degrading. In this paper, we propose GIFT, a novel continual fine-tuning approach that utilizes synthetic data to overcome catastrophic forgetting in VLMs. Taking advantage of recent advances in text-to-image synthesis, we employ a pre-trained diffusion model to recreate both pre-training and learned downstream task data. In this way, the VLM can revisit previous knowledge through distillation on matching diffusion-generated images and corresponding text prompts. Leveraging the broad distribution and high alignment between synthetic image-text pairs in VLM's feature space, we propose a contrastive distillation loss along with an image-text alignment constraint. To further combat in-distribution overfitting and enhance distillation performance with limited amount of generated data, we incorporate adaptive weight consolidation, utilizing Fisher information from these synthetic image-text pairs and achieving a better stability-plasticity balance. Extensive experiments demonstrate that our method consistently outperforms previous state-of-the-art approaches across various settings.

cs.CV↗

arXiv2025

Entropy-Based Decoding for Retrieval-Augmented Large Language Models

Augmenting Large Language Models (LLMs) with retrieved external knowledge has proven effective for improving the factual accuracy of generated responses. Despite their success, retrieval-augmented LLMs still face the distractibility issue, where the generated responses are negatively influenced by noise from both external and internal knowledge sources. In this paper, we introduce a novel, training-free decoding method guided by entropy considerations to mitigate this issue. Our approach utilizes entropy-based document-parallel ensemble decoding to prioritize low-entropy distributions from retrieved documents, thereby enhancing the extraction of relevant information of context. Additionally, it incorporates a contrastive decoding mechanism that contrasts the obtained low-entropy ensemble distribution with the high-entropy distribution derived from the model's internal knowledge across layers, which ensures a greater emphasis on reliable external information. Extensive experiments on open-domain question answering datasets demonstrate the superiority of our method.

cs.CL↗

arXiv2025

The Kramers escape rate of phase transitions for the 6-dimensional Gauss-Bonnet AdS black hole with triple phases

In this study, we obtain specific picture of the phase transitions for the 6-dimensional Gauss-Bonnet Anti-de Sitter (AdS) black hole with triple phases, using the generalized free energy we constructed and Kramers escape rate in stochastic motion. There are six possible phase transition processes between the three different stable states (small, medium, and large black hole states). During these phase transitions, there are two key temperatures. One is the temperature at which the medium black hole state emerges, and the other is the temperature at which the small black hole state annihilates. Meanwhile, two dynamic equilibrium processes are formed. One is a dynamic equilibrium of the transition from the medium black hole state to the large black hole state and the transition from the small black hole state to the medium black hole state. The other is a dynamic equilibrium of the transition from the small black hole state to the medium black hole state and the transition from the medium black hole state to the small black hole state.

gr-qc↗

arXiv2025

Automatic Database Configuration Debugging using Retrieval-Augmented Language Models

Database management system (DBMS) configuration debugging, e.g., diagnosing poorly configured DBMS knobs and generating troubleshooting recommendations, is crucial in optimizing DBMS performance. However, the configuration debugging process is tedious and, sometimes challenging, even for seasoned database administrators (DBAs) with sufficient experience in DBMS configurations and good understandings of the DBMS internals (e.g., MySQL or Oracle). To address this difficulty, we propose Andromeda, a framework that utilizes large language models (LLMs) to enable automatic DBMS configuration debugging. Andromeda serves as a natural surrogate of DBAs to answer a wide range of natural language (NL) questions on DBMS configuration issues, and to generate diagnostic suggestions to fix these issues. Nevertheless, directly prompting LLMs with these professional questions may result in overly generic and often unsatisfying answers. To this end, we propose a retrieval-augmented generation (RAG) strategy that effectively provides matched domain-specific contexts for the question from multiple sources. They come from related historical questions, troubleshooting manuals and DBMS telemetries, which significantly improve the performance of configuration debugging. To support the RAG strategy, we develop a document retrieval mechanism addressing heterogeneous documents and design an effective method for telemetry analysis. Extensive experiments on real-world DBMS configuration debugging datasets show that Andromeda significantly outperforms existing solutions.

cs.DB↗

arXiv2024

Thermodynamic bounce effect in quantum BTZ black hole

A novel thermodynamic phenomenon has been observed in the quantum Bañados-Teitelboim-Zanelli (qBTZ) black hole, utilizing generalized free energy and Kramer escape rate. This phenomenon also reveals the unique property of the quantum black hole. The stochastic thermal motion of various thermodynamic states within the black hole system induces phase transitions, under the influence of generalized free energy which obtained by extending Maxwell's construction. Through the analysis of Kramer escape rate, it is discovered that the qBTZ black hole thermodynamic system exhibits a bounce effect. It originates from the non-monotonicity of entropy in black hole thermodynamic systems. Furthermore, the overall thermodynamic picture of the qBTZ black hole has been obtained under different quantum backreactions.

gr-qc↗

arXiv2024

SciSafeEval: A Comprehensive Benchmark for Safety Alignment of Large Language Models in Scientific Tasks

Large language models (LLMs) have a transformative impact on a variety of scientific tasks across disciplines including biology, chemistry, medicine, and physics. However, ensuring the safety alignment of these models in scientific research remains an underexplored area, with existing benchmarks primarily focusing on textual content and overlooking key scientific representations such as molecular, protein, and genomic languages. Moreover, the safety mechanisms of LLMs in scientific tasks are insufficiently studied. To address these limitations, we introduce SciSafeEval, a comprehensive benchmark designed to evaluate the safety alignment of LLMs across a range of scientific tasks. SciSafeEval spans multiple scientific languages-including textual, molecular, protein, and genomic-and covers a wide range of scientific domains. We evaluate LLMs in zero-shot, few-shot and chain-of-thought settings, and introduce a "jailbreak" enhancement feature that challenges LLMs equipped with safety guardrails, rigorously testing their defenses against malicious intention. Our benchmark surpasses existing safety datasets in both scale and scope, providing a robust platform for assessing the safety and performance of LLMs in scientific contexts. This work aims to facilitate the responsible development and deployment of LLMs, promoting alignment with safety and ethical standards in scientific research.

cs.CL↗

arXiv2024

Analysis of the form factors of $B_c\rightarrow D^{()}$, $D_{s}^{()}$ and their nonleptonic decays

This article is devoted to calculating the form factors of $B_c \to D^{*}$, $B_c \to D$, $B_c \to D_s^{*}$ and $B_c \to D_s$ transitions in the framework of three-point QCD sum rules. At the QCD side, the contributions of $\langle\overline{q}q\rangle$, $\langle\overline{q}g_{s}σGq\rangle$, $\langle g_{s}^{2}G^{2}\rangle$, $\langle f^{3}G^{3}\rangle$ and $\langle\overline{q}q\rangle \langle g_{s}^{2}G^{2}\rangle$ are taken into account. With the obtained form factors, the decay widths and branching ratios of several two-body nonleptonic decay processes $B_c \to η_c D^{*}$, $η_c D$, $ J/ψD^{*}$, $ J/ψD$, $η_c D_s^{*}$, $η_c D_s$, $J/ψD_s^{*}$ and $J/ψD_s$ are predicted. These results about the form factors and decay properties of $B_c$ meson provide useful information for us to study the heavy-quark dynamics.

hep-ph↗

arXiv2024

Quantum simulation of thermal field theories

We present our recent studies on thermal field theories using quantum algorithms. We first delve into the representation of quantum fields via qubits on general digital quantum computers alongside the quantum algorithms employed to evaluate thermal properties of generic quantum field theories. Then, we show our numerical results of thermal field theories in 1+1 dimensions using quantum simulators. Both fermion and scalar fields will be discussed. These studies aim to understand thermal fixed points for our forthcoming work on studying thermalisation in quantum field theories in real time quantum simulation.

quant-ph↗

arXiv2024

Equilibrium convergence in large games

This paper presents a general closed graph property for (randomized strategy) Nash equilibrium correspondence in large games. In particular, we show that for any large game with a convergent sequence of fiinite-player games, the limit of any convergent sequence of Nash equilibria of the corresponding finite-player games can be induced by a Nash equilibrium of the large game. Such a result goes beyond earlier results on the closed graph property for pure strategy Nash equilibrium correspondence in large games in multiple aspects. An application on equilibrium selection in large games is also presented.

math.OC↗

arXiv2024

A Comprehensive Review of Propagation Models in Complex Networks: From Deterministic to Deep Learning Approaches

Understanding propagation mechanisms in complex networks is essential for fields like epidemiology and multi-robot networks. This paper reviews various propagation models, from traditional deterministic frameworks to advanced data-driven and deep learning approaches. We differentiate between static and dynamic networks, noting that static models provide foundational insights, while dynamic models capture real-world temporal changes. Deterministic models like the SIR framework offer clear mathematical insights but often lack adaptability to randomness, whereas stochastic models enhance realism at the cost of interpretability. Behavior-based models focus on individual decision-making, demanding more computational resources. Data-driven approaches improve accuracy in nonlinear scenarios by adapting to evolving networks, using either traditional models or model-free machine learning techniques. We explore supervised and unsupervised learning methods, as well as reinforcement learning, which operates without predefined datasets. The application of graph neural networks (GNNs) is also discussed, highlighting their effectiveness in modeling propagation in complex networks. The paper underscores key applications and challenges associated with each model type, emphasizing the increasing importance of hybrid and machine learning-based solutions in contemporary network propagation issues.

cs.SI↗

arXiv2024

Instruction Tuning With Loss Over Instructions

Instruction tuning plays a crucial role in shaping the outputs of language models (LMs) to desired styles. In this work, we propose a simple yet effective method, Instruction Modelling (IM), which trains LMs by applying a loss function to the instruction and prompt part rather than solely to the output part. Through experiments across 21 diverse benchmarks, we show that, in many scenarios, IM can effectively improve the LM performance on both NLP tasks (e.g., MMLU, TruthfulQA, and HumanEval) and open-ended generation benchmarks (e.g., MT-Bench and AlpacaEval). Remarkably, in the most advantageous case, IM boosts model performance on AlpacaEval 1.0 by over 100%. We identify two key factors influencing the effectiveness of IM: (1) The ratio between instruction length and output length in the training data; and (2) The number of training examples. We observe that IM is especially beneficial when trained on datasets with lengthy instructions paired with brief outputs, or under the Superficial Alignment Hypothesis (SAH) where a small amount of training examples are used for instruction tuning. Further analysis substantiates our hypothesis that our improvement can be attributed to reduced overfitting to instruction tuning datasets. It is worth noting that we are not proposing \ours as a replacement for current fine-tuning processes. Instead, our work aims to provide practical guidance for instruction tuning LMs, especially in low-resource scenarios.

cs.CL↗

arXiv2024

Experimental demonstration of frequency downconverted arm length stabilization for a future upgraded gravitational wave detector

Ground-based laser interferometric gravitational wave detectors consist of complex multiple optical cavity systems. An arm-length stabilization (ALS) system has played an important role in bringing such complex detector into operational state and enhance the duty cycle. The sensitivity of these detectors can be improved if the thermal noise of their test mass mirror coatings is reduced. Crystalline AlGaAs coatings are a promising candidate for this. However, traditional ALS system with frequency-doubled 532 nm light is no longer an option with AlGaAs coatings due to the narrow bandgap of GaAs, thus alternative locking schemes must be developed. In this letter, we describe an experimental demonstration of a novel ALS scheme which is compatible with AlGaAs coatings. This ALS scheme will enable the use of AlGaAs coatings and contribute to improved sensitivity of future detectors.

physics.optics↗

arXiv2024

Deep Reinforcement Learning for Decentralized Multi-Robot Control: A DQN Approach to Robustness and Information Integration

The superiority of Multi-Robot Systems (MRS) in various complex environments is unquestionable. However, in complex situations such as search and rescue, environmental monitoring, and automated production, robots are often required to work collaboratively without a central control unit. This necessitates an efficient and robust decentralized control mechanism to process local information and guide the robots' behavior. In this work, we propose a new decentralized controller design method that utilizes the Deep Q-Network (DQN) algorithm from deep reinforcement learning, aimed at improving the integration of local information and robustness of multi-robot systems. The designed controller allows each robot to make decisions independently based on its local observations while enhancing the overall system's collaborative efficiency and adaptability to dynamic environments through a shared learning mechanism. Through testing in simulated environments, we have demonstrated the effectiveness of this controller in improving task execution efficiency, strengthening system fault tolerance, and enhancing adaptability to the environment. Furthermore, we explored the impact of DQN parameter tuning on system performance, providing insights for further optimization of the controller design. Our research not only showcases the potential application of the DQN algorithm in the decentralized control of multi-robot systems but also offers a new perspective on how to enhance the overall performance and robustness of the system through the integration of local information.

cs.RO↗

arXiv2024

State-of-the-art in Robot Learning for Multi-Robot Collaboration: A Comprehensive Survey

With the continuous breakthroughs in core technology, the dawn of large-scale integration of robotic systems into daily human life is on the horizon. Multi-robot systems (MRS) built on this foundation are undergoing drastic evolution. The fusion of artificial intelligence technology with robot hardware is seeing broad application possibilities for MRS. This article surveys the state-of-the-art of robot learning in the context of Multi-Robot Cooperation (MRC) of recent. Commonly adopted robot learning methods (or frameworks) that are inspired by humans and animals are reviewed and their advantages and disadvantages are discussed along with the associated technical challenges. The potential trends of robot learning and MRS integration exploiting the merging of these methods with real-world applications is also discussed at length. Specifically statistical methods are used to quantitatively corroborate the ideas elaborated in the article.

cs.RO↗

arXiv2024

Cold nuclear matter effects on azimuthal decorrelation in heavy-ion collisions

The assumption of factorization lies at the core of calculations of medium effects on observables computable in perturbative Quantum Chromodynamics. In this work we examine this assumption, for which we propose a setup to study hard processes and bulk nuclear matter in heavy-ion collisions on the same footing using the Glauber modelling of heavy nuclei. To exemplify this approach, we calculate the leading-order corrections to azimuthal decorrelation in Drell-Yan and boson-jet processes due to cold nuclear matter effects, not considering radiation. At leading order in both the hard momentum scale and the nuclear size, the impact-parameter dependent cross section is found to factorize for both processes. The factorization formula involves a convolution of the hard cross section with the medium-modified parton distributions, and, for boson-jet production, the medium-modified jet function.

hep-ph↗

arXiv2024

Quantum computation in fermionic thermal field theories

Thermal properties of quantum fields at finite temperature are crucial to understanding strongly interacting matter and recent development in quantum computing has provided an alternative and promising avenue of study. In this work, we study thermal field theories involving only fermions using quantum algorithms. We first delve into the presentations of fermion fields via qubits on digital quantum computers alongside the quantum algorithms such as quantum imaginary time evolutions employed to evaluate thermal properties of generic quantum field theories. Specifically, we show numerical results such as the thermal distribution and the energy density of thermal field theories for Majorana fermions in 1+1 dimensions using quantum simulators. In addition to free field theory, we also study the effects of interactions resulting from coupling with a spatially homogeneous Majorana field. In both cases, we show analytically that thermal properties of the system can be described using phase-space distributions, and the quantum simulation results agree with analytical and semiclassical expectations. Our work is an important step to understand thermal fixed points, preparing for quantum simulation of thermalization in real time.

hep-ph↗

arXiv2024

Understanding the Role of User Profile in the Personalization of Large Language Models

Utilizing user profiles to personalize Large Language Models (LLMs) has been shown to enhance the performance on a wide range of tasks. However, the precise role of user profiles and their effect mechanism on LLMs remains unclear. This study first confirms that the effectiveness of user profiles is primarily due to personalization information rather than semantic information. Furthermore, we investigate how user profiles affect the personalization of LLMs. Within the user profile, we reveal that it is the historical personalized response produced or approved by users that plays a pivotal role in personalizing LLMs. This discovery unlocks the potential of LLMs to incorporate a greater number of user profiles within the constraints of limited input length. As for the position of user profiles, we observe that user profiles integrated into different positions of the input context do not contribute equally to personalization. Instead, where the user profile that is closer to the beginning affects more on the personalization of LLMs. Our findings reveal the role of user profiles for the personalization of LLMs, and showcase how incorporating user profiles impacts performance providing insight to leverage user profiles effectively.

cs.CL↗