arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 667 records · Page 37Linked to original sources

arXiv2024

InternLM-XComposer2: Mastering Free-form Text-Image Composition and Comprehension in Vision-Language Large Model

We introduce InternLM-XComposer2, a cutting-edge vision-language model excelling in free-form text-image composition and comprehension. This model goes beyond conventional vision-language understanding, adeptly crafting interleaved text-image content from diverse inputs like outlines, detailed textual specifications, and reference images, enabling highly customizable content creation. InternLM-XComposer2 proposes a Partial LoRA (PLoRA) approach that applies additional LoRA parameters exclusively to image tokens to preserve the integrity of pre-trained language knowledge, striking a balance between precise vision understanding and text composition with literary talent. Experimental results demonstrate the superiority of InternLM-XComposer2 based on InternLM2-7B in producing high-quality long-text multi-modal content and its exceptional vision-language understanding performance across various benchmarks, where it not only significantly outperforms existing multimodal models but also matches or even surpasses GPT-4V and Gemini Pro in certain assessments. This highlights its remarkable proficiency in the realm of multimodal understanding. The InternLM-XComposer2 model series with 7B parameters are publicly available at https://github.com/InternLM/InternLM-XComposer.

cs.CV↗

arXiv2024

Supervised and unsupervised learning of (1+1)-dimensional even-offspring branching annihilating random walks

Machine learning (ML) of phase transitions (PTs) has gradually become an effective approach that enables us to explore the nature of various PTs more promptly in equilibrium and nonequilibrium systems. Unlike equilibrium systems, non-equilibrium systems display more complicated and diverse features because of the extra dimension of time, which is not readily tractable, both theoretically and numerically. The combination of ML and most renowned nonequilibrium model, directed percolation (DP), led to some significant findings. In this study, ML is applied to (1+1)-d, even offspring branching annihilating random walks (BAW), whose universality class is not DP-like. The supervised learning of (1+1)-d BAW via convolutional neural networks (CNN) results in a more accurate prediction of the critical point than the Monte Carlo (MC) simulation for the same system sizes. The dynamic exponent \;$z$\; and spatial correlation length correlation exponent \;$ν_{\perp}$\ were also measured and found to be consistent with their respective theoretical values. Furthermore, the unsupervised learning of (1+1)-d BAW via an autoencoder (AE) gives rise to a transition point, which is the same as the critical point. The latent layer of AE, through a single neuron, can be regarded as the order parameter of the system being properly re-scaled. Therefore, we believe that ML has exciting application prospects in reaction-diffusion systems such as BAW and DP.

nlin.CG↗

arXiv2024

The Geometry of the Modular Bootstrap

We explore the geometry behind the modular bootstrap and its image in the space of Taylor coefficients of the torus partition function. In the first part, we identify the geometry as an intersection of planes with the convex hull of moment curves on $R^+{\otimes}\mathbb{Z}$, with boundaries characterized by the total positivity of generalized Hankel matrices. We phrase the Hankel constraints as a semi-definite program, which has several advantages, such as constant computation time with increasing central charge. We derive bounds on the gap, twist-gap, and the space of Taylor coefficients themselves. We find that if the gap is above $Δ^*_{gap}$, where $\frac{c{-}1}{12}<Δ^*_{gap}< \frac{c}{12}$, all coefficients become bounded on both sides and kinks develop in the space. In the second part, we propose an analytic method of imposing the integrality condition for the degeneracy number in the spinless bootstrap, which leads to a non-convex geometry. We find that even at very low derivative order this condition rules out regions otherwise allowed by bootstraps at high derivative order.

hep-th↗

arXiv2024

Improving Natural Language Capability of Code Large Language Model

Code large language models (Code LLMs) have demonstrated remarkable performance in code generation. Nonetheless, most existing works focus on boosting code LLMs from the perspective of programming capabilities, while their natural language capabilities receive less attention. To fill this gap, we thus propose a novel framework, comprising two modules: AttentionExtractor, which is responsible for extracting key phrases from the user's natural language requirements, and AttentionCoder, which leverages these extracted phrases to generate target code to solve the requirement. This framework pioneers an innovative idea by seamlessly integrating code LLMs with traditional natural language processing tools. To validate the effectiveness of the framework, we craft a new code generation benchmark, called MultiNL-H, covering five natural languages. Extensive experimental results demonstrate the effectiveness of our proposed framework.

cs.CL↗

arXiv2024

Nonreciprocal charge transport in the titanium sesquioxide heterointerface superconductor

Nonreciprocal charge transport in heterostructural superconductors exhibits appealing quantum physical phenomena and holds the promising potential for superconducting circuits applications. Realizing a nonreciprocity is, however, fundamentally and technologically challenging, as it requires a material structure without a centre of inversion, which is scarce among superconducting materials. Here, we report an evidence of helical superconductivity, in which the Rashba spin-orbit coupling induces momentum-dependent superconducting gap in the inversion symmetry breaking heterointerface superconductor consisting of Mott insulating Ti$_2$O$_3$ and polar semiconducting GaN. Remarkably, the nonlinear responses emerge in the superconducting transition regime, when the magnetic field is precisely aligned in-plane orientations perpendicular to the applied current. In particular, the observed nonreciprocal supercurrent is extremely sensitive to the direction of the magnetic field for 0.5 degree, suggestive of a crossover from a symmetry breaking state to a symmetric one. Our finding not only unveils the underlying rich physical properties in heterointerface superconductors, but also provides an exciting opportunity for the development of novel mesoscopic superconducting devices.

cond-mat.supr-con↗

arXiv2024

Bilayer $t$-$J$-$J_\perp$ Model and Magnetically Mediated Pairing in the Pressurized Nickelate La$_3$Ni$_2$O$_7$

The recently discovered nickelate superconductor La$_3$Ni$_2$O$_7$ has a high transition temperature near 80 K under pressure, which offers additional avenues of unconventional superconductivity. Here with state-of-the-art tensor-network methods, we study a bilayer $t$-$J$-$J_\perp$ model for La$_3$Ni$_2$O$_7$ and find a robust $s$-wave superconductive (SC) order mediated by interlayer magnetic couplings. Large-scale density matrix renormalization group calculations find algebraic pairing correlations with Luttinger parameter of $K_{\rm SC} \simeq 1$. Infinite projected entangled-pair state method obtains a nonzero SC order directly in the thermodynamic limit, and estimates a strong pairing strength $\barΔ_z \sim \mathcal{O}(0.1)$. Tangent-space tensor renormalization group simulations further determine a high SC temperature $T_c^*/J \sim \mathcal{O}(0.1)$ and clarify the temperature evolution of SC order. Because of the intriguing orbital selective behaviors and strong Hund's rule coupling in the compound, $t$-$J$-$J_\perp$ model has strong interlayer spin exchange (while negligible interlayer hopping), which greatly enhances the SC pairing in the bilayer system. Such a magnetically mediated strong pairing has also been observed recently in the optical lattice of ultracold atoms. Our accurate and comprehensive tensor-network calculations reveal robust SC order in the bilayer $t$-$J$-$J_\perp$ model and shed light on the high-$T_c$ superconductivity in the pressurized nickelate La$_3$Ni$_2$O$_7$.

cond-mat.str-el↗

arXiv2024

The Study of Mode Switching behavior of PSR J0614+2229 Using the Parkes Ultra-wideband Receiver Observations

In this paper, we presented a detailed single pulse and polarization study of PSR J0614+2229 based on the archived data observed on 2019 August 15 (MJD 58710) and September 12 (MJD 58738) using the Ultra-wideband Low-frequency Receiver on the Parkes radio telescope. The single-pulse sequences show that this pulsar switches between two emission states, in which the emission of state A occurs earlier than that of state B in pulse longitude. We found that the variation in relative brightness between the two states is related to time and both states follow a simple power law very well. Based on the phase-aligned multi-frequency profiles, we found that there is a significant difference in the distributions of spectral index across the emission regions of the two states. Furthermore, we obtained the emission height evolution for the two emission states and found that, at a fixed frequency, the emission height of state A is higher than that of state B. What is even more interesting is that the emission heights of both states A and B have not changed with frequency. Our results suggest that the mode switching of this pulsar is possibly caused by changes in the emission heights that alter the distributions of spectral index across the emission regions of states A and B resulting in the frequency-dependent behaviors, i.e., intensity and pulse width.

astro-ph.HE↗

arXiv2024

A GAN-based data poisoning framework against anomaly detection in vertical federated learning

In vertical federated learning (VFL), commercial entities collaboratively train a model while preserving data privacy. However, a malicious participant's poisoning attack may degrade the performance of this collaborative model. The main challenge in achieving the poisoning attack is the absence of access to the server-side top model, leaving the malicious participant without a clear target model. To address this challenge, we introduce an innovative end-to-end poisoning framework P-GAN. Specifically, the malicious participant initially employs semi-supervised learning to train a surrogate target model. Subsequently, this participant employs a GAN-based method to produce adversarial perturbations to degrade the surrogate target model's performance. Finally, the generator is obtained and tailored for VFL poisoning. Besides, we develop an anomaly detection algorithm based on a deep auto-encoder (DAE), offering a robust defense mechanism to VFL scenarios. Through extensive experiments, we evaluate the efficacy of P-GAN and DAE, and further analyze the factors that influence their performance.

cs.LG↗

arXiv2024

Space Networking Kit: A Novel Simulation Platform for Emerging LEO Mega-constellations

This paper presents SNK, a novel simulation platform designed to evaluate the network performance of constellation systems for global Internet services. SNK offers realtime communication visualization and supports the simulation of routing between edge node of network. The platform enables the evaluation of routing and network performance metrics such as latency, stretch, network capacity, and throughput under different network structures and density. The effectiveness of SNK is demonstrated through various simulation cases, including the routing between fixed edge stations or mobile edge stations and analysis of space network structures.

cs.NI↗

arXiv2024

Multi-Protocol Location Forwarding (MPLF) for Space Routing

The structure and routing architecture design is critical for achieving low latency and high capacity in future LEO space networks (SNs). Existing studies mainly focus on topologies of space networks, but there is a lack of analysis on constellation structures, which can greatly affect network performance. In addition, some routing architectures are designed for networks with a small number of network nodes such as Iridium while they introduce significant network overhead for high-density networks (i.e., mega-constellation networks containing thousands of satellites). In this paper, we conduct the quantitatively study on the design of network structure and routing architecture in space. The high density, high dynamics, and large scale nature of emerging Space Networks (SNs) pose significant challenges, such as unstable routing paths, low network reachability, high latency, and large jitter. To alleviate the above challenges, we design the structure of space network to maximum the connectivity through wisely adjusting the inter-plane inter satellite link. We further propose Multi-Protocol Location Forwarding (MPLF), a distributed routing architecture, targeting at minimizing the propagation latency with a distributed, convergence-free routing paradigm, while keeping routing stable and maximum the path diversity. Comprehensive experiments are conducted on a customized platform \textit{Space Networking Kits} (SNK) which demonstrate that our solution can outperform existing related schemes by about 14\% reduction of propagation latency and 66\% reduction of hops-count on average, while sustaining a high path diversity with only $O(1)$ time complexity.

cs.NI↗

arXiv2024

Universal scalefree non-Hermitian skin effect near the Bloch point

The scalefree non-Hermitian skin effect (NHSE) refers to the phenomenon that the localization length of skin modes scales proportionally with system size in non-Hermitian systems. Authors of recent studies have demonstrated that the scalefree NHSE can be induced through various mechanisms, including the critical NHSE, local non-Hermiticity, and the boundary impurity effect. Nevertheless, these methods require careful modeling and precise parameter tuning. In contrast, in this paper, we suggest that the scalefree NHSE is a universal phenomenon, observable in extensive systems if these systems can be described by non-Bloch band theory and host Bloch points on the energy spectrum in the thermodynamic limit. Crucially, we discover that the geometry of the generalized Brillouin zone determines the scaling rule of the localization length, which can scale either linearly or quadratically with the system size. In this paper, we enriches the phenomenon of the scalefree NHSE.

quant-ph↗

arXiv2024

The decay property of the $X(3842)$ as the $ψ_{_3}(1^3D_{_3})$ state

In this paper, the new particle $X(3842)$ discovered by the LHCb Collaboration is identified to be the $ψ_{_3}(1^3D_{_3})$ state. We study its strong decays with the combination of the Bethe-Salpeter method and the $^3P_{_0}$ model. Its electromagnetic (EM) decay is also calculated by the Bethe-Salpeter method within Mandelstam formalism. The strong decay widths are {$Γ[X(3842)\rightarrow D^{0}\bar{D}^{0}]=1.28$ MeV}, $Γ[X(3823)\rightarrow D^{+}D^{-}]=1.08$ MeV, and the ratio ${\cal B}[X(3842)\rightarrow D^{+}D^{-}]/{\cal B}[X(3823)\rightarrow D^{0}\bar{D}^{0}]=0.84$. The EM decay width is $Γ[X(3842)\rightarrowχ_{_{c2}}γ]=0.29$ MeV. We also estimate the total width to be 2.87 MeV, which is in good agreement with the experimental data $2.79^{+0.86}_{-0.86}$ MeV. Since the used relativistic wave functions include different partial waves, we also study the contributions of different partial waves in electromagnetic decay.

hep-ph↗

arXiv2024

Towards Code Watermarking with Dual-Channel Transformations

The expansion of the open source community and the rise of large language models have raised ethical and security concerns on the distribution of source code, such as misconduct on copyrighted code, distributions without proper licenses, or misuse of the code for malicious purposes. Hence it is important to track the ownership of source code, in which watermarking is a major technique. Yet, drastically different from natural languages, source code watermarking requires far stricter and more complicated rules to ensure the readability as well as the functionality of the source code. Hence we introduce SrcMarker, a watermarking system to unobtrusively encode ID bitstrings into source code, without affecting the usage and semantics of the code. To this end, SrcMarker performs transformations on an AST-based intermediate representation that enables unified transformations across different programming languages. The core of the system utilizes learning-based embedding and extraction modules to select rule-based transformations for watermarking. In addition, a novel feature-approximation technique is designed to tackle the inherent non-differentiability of rule selection, thus seamlessly integrating the rule-based transformations and learning-based networks into an interconnected system to enable end-to-end training. Extensive experiments demonstrate the superiority of SrcMarker over existing methods in various watermarking requirements.

cs.CR↗

arXiv2024

Evolutionary Alternating Direction Method of Multipliers for Constrained Multi-Objective Optimization with Unknown Constraints

Constrained multi-objective optimization problems (CMOPs) pervade real-world applications in science, engineering, and design. Constraint violation has been a building block in designing evolutionary multi-objective optimization algorithms for solving constrained multi-objective optimization problems. However, in certain scenarios, constraint functions might be unknown or inadequately defined, making constraint violation unattainable and potentially misleading for conventional constrained evolutionary multi-objective optimization algorithms. To address this issue, we present the first of its kind evolutionary optimization framework, inspired by the principles of the alternating direction method of multipliers that decouples objective and constraint functions. This framework tackles CMOPs with unknown constraints by reformulating the original problem into an additive form of two subproblems, each of which is allotted a dedicated evolutionary population. Notably, these two populations operate towards complementary evolutionary directions during their optimization processes. In order to minimize discrepancy, their evolutionary directions alternate, aiding the discovery of feasible solutions. Comparative experiments conducted against five state-of-the-art constrained evolutionary multi-objective optimization algorithms, on 120 benchmark test problem instances with varying properties, as well as two real-world engineering optimization problems, demonstrate the effectiveness and superiority of our proposed framework. Its salient features include faster convergence and enhanced resilience to various Pareto front shapes.

cs.NE↗

arXiv2024

Magnon Damping Minimum and Logarithmic Scaling in a Kondo-Heisenberg Model

Recently, an anomalous temperature evolution of spin wave excitations has been observed in a van der Waals metallic ferromagnet Fe$_3$GeTe$_2$ (FGT) [S. Bao, et al., Phys. Rev. X 12, 011022 (2022)], whose theoretical understanding yet remains elusive. Here we study the spin dynamics of a ferromagnetic Kondo-Heisenberg lattice model at finite temperature, and propose a mechanism of magnon damping that explains the intriguing experimental results. In particular, we find the magnon damping rate $γ(T)$ firstly decreases as temperature lowers, due to the reduced magnon-magnon scatterings. It then reaches a minimum at $T_{\rm d}^*$, and rises up again following a logarithmic scaling $γ(T) \sim \ln{(T_0/T)}$ (with $T_0$ a constant) for $T < T_{\rm d}^*$, which can be attributed to electron-magnon scatterings of spin-flip type. Moreover, we obtain the phase diagram containing the ferromagnetic and Kondo insulator phases by varying the Kondo coupling, which may be relevant for experiments on pressured FGT. The presence of a magnon damping minimum and logarithmic scaling at low temperature indicates the emergence of the Kondo effect reflected in the collective excitations of local moments in a Kondo lattice system.

cond-mat.str-el↗

arXiv2023

A Generalist FaceX via Learning Unified Facial Representation

This work presents FaceX framework, a novel facial generalist model capable of handling diverse facial tasks simultaneously. To achieve this goal, we initially formulate a unified facial representation for a broad spectrum of facial editing tasks, which macroscopically decomposes a face into fundamental identity, intra-personal variation, and environmental factors. Based on this, we introduce Facial Omni-Representation Decomposing (FORD) for seamless manipulation of various facial components, microscopically decomposing the core aspects of most facial editing tasks. Furthermore, by leveraging the prior of a pretrained StableDiffusion (SD) to enhance generation quality and accelerate training, we design Facial Omni-Representation Steering (FORS) to first assemble unified facial representations and then effectively steer the SD-aware generation process by the efficient Facial Representation Controller (FRC). %Without any additional features, Our versatile FaceX achieves competitive performance compared to elaborate task-specific models on popular facial editing tasks. Full codes and models will be available at https://github.com/diffusion-facex/FaceX.

cs.CV↗

arXiv2023

Investigating Inter-Satellite Link Spanning Patterns on Networking Performance in Mega-constellations

Low Earth orbit (LEO) mega-constellations rely on inter-satellite links (ISLs) to provide global connectivity. We note that in addition to the general constellation parameters, the ISL spanning patterns are also greatly influence the final network structure and thus the network performance. In this work, we formulate the ISL spanning patterns, apply different patterns to mega-constellation and generate multiple structures. Then, we delve into the performance estimation of these networks, specifically evaluating network capacity, throughput, latency, and routing path stretch. The experimental findings provide insights into the optimal network structure under diverse conditions, showcasing superior performance when compared to alternative network configurations.

cs.NI↗

arXiv2023

Observation of a 1/3 Magnetisation Plateau Phase as Evidence for the Kitaev Interaction in a Honeycomb-Lattice Antiferromagnet

Fractional magnetisation plateaus, in which the magnetisation is pinned at a fraction of its saturated value within a range of external magnetic field, are spectacular macroscopic manifestations of the collective quantum behaviours. One prominent example of the plateau phase is found in spin-1/2 triangular-lattice antiferromagnets featuring strong geometrical frustration, and is often interpreted as quantum-fluctuation-stabilised state in magnetic field via the "order-by-disorder" mechanism. Here, we observe an unprecedented 1/3 magnetisation plateau between 5.2 and 7.4 T at 2 K in a spin-1 antiferromagnet Na$_3$Ni$_2$BiO$_6$ with a honeycomb lattice, where conventionally no geometrical frustration is anticipated. By carrying out elastic neutron scattering measurements, we propose the spin structure of the plateau phase to be an unusual partial spin-flop ferrimagnetic order, transitioning from the zigzag antiferromagnetic order in zero field. Our theoretical calculations show that the plateau phase is stabilised by the bond-anisotropic Kitaev interaction. These results provide a new paradigm for the exploration of rich quantum phases in frustrated magnets and exotic Kitaev physics in high-spin systems.

cond-mat.str-el↗