arXiv subjects

Wenxin Ding

Publications and source records attributed to Wenxin Ding.

At least 19 recordsLinked to original sources

arXiv2026

An Empirical Analysis of AI Slop in Music Streaming

Generative AI models lower the bar for content creation, making it easy for any user to create professional-looking images, text and music with minimal effort. This has enabled a new cottage industry around creation of "AI slop" mass quantities of mediocre content produced to generate revenue, often through misrepresentation as human-authored content, or scams involving automated scripts and fake consumption. While there are obvious parallels between the AI-slop industry and "traditional" email spam networks, it might be too early to determine if AI slop generation can grow into a similar self-sustaining industry. In this paper, we look specifically at the music industry, and explore the question: Can we prevent AI music slop from growing into a self-sustaining shadow industry? To answer this question, we characterize the current state of AI slop in music, and its pipeline from generation, distribution, and consumption by users on streaming platforms. By examining growth and engagement on Spotify, we confirm that AI music exhibits AI slop characteristics: the overwhelming majority (93%) of AI music receive few, if any listener plays, and are rarely recommended. AI musicians "spray and pray," releasing large volumes of music across multiple genres in hopes of generating a hit. We also explore the AI slop pipeline by generating and publishing our own AI tracks onto streaming through 11 indie music distributors. We find distributors have inconsistent and largely unenforced policies on AI music, making it surprisingly easy to publish mass produced AI songs. Finally, we consider AI music detection, and find that current methods lack accuracy or robustness. As generation costs decrease, we believe slop generation in music will become self-sustainable, unless concrete steps are taken by the music industry. We consider and discuss potential mitigation methods based on our findings.

cs.CR↗

arXiv2025

Diffusion Model-Based Size Variable Virtual Try-On Technology and Evaluation Method

With the rapid development of e-commerce, virtual try-on technology has become an essential tool to satisfy consumers' personalized clothing preferences. Diffusion-based virtual try-on systems aim to naturally align garments with target individuals, generating realistic and detailed try-on images. However, existing methods overlook the importance of garment size variations in meeting personalized consumer needs. To address this, we propose a novel virtual try-on method named SV-VTON, which introduces garment sizing concepts into virtual try-on tasks. The SV-VTON method first generates refined masks for multiple garment sizes, then integrates these masks with garment images at varying proportions, enabling virtual try-on simulations across different sizes. In addition, we developed a specialized size evaluation module to quantitatively assess the accuracy of size variations. This module calculates differences between generated size increments and international sizing standards, providing objective measurements of size accuracy. To further validate SV-VTON's generalization capability across different models, we conducted experiments on multiple SOTA Diffusion models. The results demonstrate that SV-VTON consistently achieves precise multi-size virtual try-on across various SOTA models, and validates the effectiveness and rationality of the proposed method, significantly fulfilling users' personalized multi-size virtual try-on requirements.

cs.MM↗

arXiv2024

Understanding Implosion in Text-to-Image Generative Models

Recent works show that text-to-image generative models are surprisingly vulnerable to a variety of poisoning attacks. Empirical results find that these models can be corrupted by altering associations between individual text prompts and associated visual features. Furthermore, a number of concurrent poisoning attacks can induce "model implosion," where the model becomes unable to produce meaningful images for unpoisoned prompts. These intriguing findings highlight the absence of an intuitive framework to understand poisoning attacks on these models. In this work, we establish the first analytical framework on robustness of image generative models to poisoning attacks, by modeling and analyzing the behavior of the cross-attention mechanism in latent diffusion models. We model cross-attention training as an abstract problem of "supervised graph alignment" and formally quantify the impact of training data by the hardness of alignment, measured by an Alignment Difficulty (AD) metric. The higher the AD, the harder the alignment. We prove that AD increases with the number of individual prompts (or concepts) poisoned. As AD grows, the alignment task becomes increasingly difficult, yielding highly distorted outcomes that frequently map meaningful text prompts to undefined or meaningless visual representations. As a result, the generative model implodes and outputs random, incoherent images at large. We validate our analytical framework through extensive experiments, and we confirm and explain the unexpected (and unexplained) effect of model implosion while producing new, unforeseen insights. Our work provides a useful tool for studying poisoning attacks against diffusion models and their defenses.

cs.CR↗

arXiv2024

Local Density of States induced near Impurities in Mott Insulators

The local density of states near dopants or impurities has recently been probed by scanning tunneling microscopy in both the parent and very lightly doped compounds of the high-$T_c$ cuprate superconductors. Our calculations based on a slave-rotor description account for all the following key features of the observed local density of states: i) positions and amplitudes of the in-gap spectral weights of a single impurity; ii) the spectral weight transfer from the upper Hubbard band to the lower Hubbard band; iii) the difference between the cases of single and multiple impurities. For multiple impurities, our study explains the complete suppression of spectral weight observed at precisely the Fermi energy and links this property to zeros of the underlying bulk Green's function of the Mott insulating phase.

cond-mat.str-el↗

arXiv2024

Rich unconventional Hall effects in a single quasi-kagome Kondo Weyl semimetal candidate Ce$_3$TiSb$_5$

It is generally believed that electronic correlation, geometric frustration, and topology, \textit{individually}, can facilitate the emergence of various intriguing properties that have attracted a broad audience for both fundamental research and potential applications. Here, we report a series of unconventional Hall effects observed in a \textit{single} compound - quasi-kagome Kondo Weyl semimetal candidate Ce$_3$TiSb$_5$. In the paramagnetic phase, signature of dynamic $c$-$f$ hybridization is revealed by a reduction of anomalous Hall effect and is connected to frustration-promoted incoherent Kondo scattering. A large topological Hall effect exceeding 0.2 $\mu\Omega\cdot$cm is found at low temperatures, which should be ascribed to the non-collinear magnetic texture. In addition, a peculiar loop-shaped Hall effect with switching chirality is also seen, which is inferred to be associated with magnetic domain walls that pin history-dependent spin chirality and / or Fermi-arc surface states projected from the in-gap Weyl nodes. These exotic results place Ce$_3$TiSb$_5$ in a regime of highly-frustrated antiferromagnetic dense Kondo lattice with a nontrivial topology on an ``extended" global phase diagram, and highlight the interplay among electronic correlation, geometric frustration and topology.

cond-mat.str-el↗

arXiv2024

Anomalous Hall effect and quantum criticality in geometrically frustrated heavy fermion metals

Studies on the heavy-fermion pyrochlore iridate (Pr$_2$Ir$_2$O$_7$) point to the role of time-reversal-symmetry breaking in geometrically frustrated Kondo lattices. With this motivation, here we study the effect of Kondo coupling and chiral spin liquids in a frustrated $J_1-J_2$ model on a square lattice. We treat the Kondo effect within a slave-fermion approach, and discuss our results in the context of a proposed global phase diagram for heavy fermion metals. We calculate the anomalous Hall response for the chiral states of both the Kondo destroyed and Kondo screened phases. Across the quantum critical point, a reconstruction of the Fermi surface leads to a sudden change of the Berry curvature distribution and, consequently, a jump of the anomalous Hall conductance. We discuss the implications of our results for the heavy-fermion pyrochlore iridate and propose an interface structure based on Kondo insulators to further explore such effects.

cond-mat.str-el↗

arXiv2024

Nightshade: Prompt-Specific Poisoning Attacks on Text-to-Image Generative Models

Data poisoning attacks manipulate training data to introduce unexpected behaviors into machine learning models at training time. For text-to-image generative models with massive training datasets, current understanding of poisoning attacks suggests that a successful attack would require injecting millions of poison samples into their training pipeline. In this paper, we show that poisoning attacks can be successful on generative models. We observe that training data per concept can be quite limited in these models, making them vulnerable to prompt-specific poisoning attacks, which target a model's ability to respond to individual prompts. We introduce Nightshade, an optimized prompt-specific poisoning attack where poison samples look visually identical to benign images with matching text prompts. Nightshade poison samples are also optimized for potency and can corrupt an Stable Diffusion SDXL prompt in <100 poison samples. Nightshade poison effects "bleed through" to related concepts, and multiple attacks can composed together in a single prompt. Surprisingly, we show that a moderate number of Nightshade attacks can destabilize general features in a text-to-image generative model, effectively disabling its ability to generate meaningful images. Finally, we propose the use of Nightshade and similar tools as a last defense for content creators against web scrapers that ignore opt-out/do-not-crawl directives, and discuss possible implications for model trainers and content creators.

cs.CR↗

arXiv2024

Towards Scalable and Robust Model Versioning

As the deployment of deep learning models continues to expand across industries, the threat of malicious incursions aimed at gaining access to these deployed models is on the rise. Should an attacker gain access to a deployed model, whether through server breaches, insider attacks, or model inversion techniques, they can then construct white-box adversarial attacks to manipulate the model's classification outcomes, thereby posing significant risks to organizations that rely on these models for critical tasks. Model owners need mechanisms to protect themselves against such losses without the necessity of acquiring fresh training data - a process that typically demands substantial investments in time and capital. In this paper, we explore the feasibility of generating multiple versions of a model that possess different attack properties, without acquiring new training data or changing model architecture. The model owner can deploy one version at a time and replace a leaked version immediately with a new version. The newly deployed model version can resist adversarial attacks generated leveraging white-box access to one or all previously leaked versions. We show theoretically that this can be accomplished by incorporating parameterized hidden distributions into the model training data, forcing the model to learn task-irrelevant features uniquely defined by the chosen data. Additionally, optimal choices of hidden distributions can produce a sequence of model versions capable of resisting compound transferability attacks over time. Leveraging our analytical insights, we design and implement a practical model versioning method for DNN classifiers, which leads to significant robustness improvements over existing methods. We believe our work presents a promising direction for safeguarding DNN services beyond their initial deployment.

cs.LG↗

arXiv2023

A Two-stage Personalized Virtual Try-on Framework with Shape Control and Texture Guidance

The Diffusion model has a strong ability to generate wild images. However, the model can just generate inaccurate images with the guidance of text, which makes it very challenging to directly apply the text-guided generative model for virtual try-on scenarios. Taking images as guiding conditions of the diffusion model, this paper proposes a brand new personalized virtual try-on model (PE-VITON), which uses the two stages (shape control and texture guidance) to decouple the clothing attributes. Specifically, the proposed model adaptively matches the clothing to human body parts through the Shape Control Module (SCM) to mitigate the misalignment of the clothing and the human body parts. The semantic information of the input clothing is parsed by the Texture Guided Module (TGM), and the corresponding texture is generated by directional guidance. Therefore, this model can effectively solve the problems of weak reduction of clothing folds, poor generation effect under complex human posture, blurred edges of clothing, and unclear texture styles in traditional try-on methods. Meanwhile, the model can automatically enhance the generated clothing folds and textures according to the human posture, and improve the authenticity of virtual try-on. In this paper, qualitative and quantitative experiments are carried out on high-resolution paired and unpaired datasets, the results show that the proposed model outperforms the state-of-the-art model.

cs.CV↗

arXiv2023

Characterizing the Optimal 0-1 Loss for Multi-class Classification with a Test-time Attacker

Finding classifiers robust to adversarial examples is critical for their safe deployment. Determining the robustness of the best possible classifier under a given threat model for a given data distribution and comparing it to that achieved by state-of-the-art training methods is thus an important diagnostic tool. In this paper, we find achievable information-theoretic lower bounds on loss in the presence of a test-time attacker for multi-class classifiers on any discrete dataset. We provide a general framework for finding the optimal 0-1 loss that revolves around the construction of a conflict hypergraph from the data and adversarial constraints. We further define other variants of the attacker-classifier game that determine the range of the optimal loss more efficiently than the full-fledged hypergraph construction. Our evaluation shows, for the first time, an analysis of the gap to optimal robustness for classifiers in the multi-class setting on benchmark datasets.

cs.LG↗

arXiv2023

Algebraic-Dynamical Perturbation Theory of Large-$U$ Hubbard Models. Single Particle Spectrum of Antiferromagnetic Mott Insulating States

In this work, we present an analytical framework for studying antiferromagnetic (AFM) Mott insulating states in the Hubbard model. We first derive an analytical solution for the single-particle Green's functions in the atomic limit. Within a second-order perturbation approach, we compute the ground state energy and show that the ground state is antiferromagnetically ordered. Then we derive an analytical solution for single-particle Green's functions when effects of the hopping term are considered in the N\'{e}el state. With the analytical solution, we compute the spectral functions and explain various properties of the AFM Mott insulating state as observed both experimentally and numerically: i) magnetic blueshift of the Mott gap; ii) the low energy part in the parental compounds of cuprate high $T_c$ superconductors, which corresponds to a single band Hubbard model description. This work comprehends the electronic properties of antiferromagnetic Mott states analytically and provides a foundation for future investigations of doped antiferromagnetic Mott insulators, aiming for the mechanism of cuprates high-$T_c$ superconductivity.

cond-mat.str-el↗

arXiv2023

Variational Equations-of-States for Interacting Quantum Hamiltonians

Variational methods are of fundamental importance and widely used in theoretical physics, especially for strongly interacting systems. In this work, we present a set of variational equations of state (VES) for pure states of an interacting quantum Hamiltonian. The VES can be expressed in terms of the variation of the density operators or static correlation functions. We derive the algebraic relationship between a known pure state density matrix and its variation, and obtain the VES by applying this relation to the averaged Heisenberg-equations-of-motion for the exact density matrix. Additionally, we provide a direct expression of the VES in terms of correlation functions to make it computable. We present three nontrivial applications of the VES: a perturbation calculation of correlation functions of the transverse field Ising model in arbitrary spatial dimensions, a study of a longitudinal field perturbation to the one-dimensional transverse field Ising model at the critical point and variational calculation of magnetization and ground state energy of the two-dimensional spin-1/2 Heisenberg model on a square lattice. For the second one, our results not only recover the scaling limit, but also indicate the possibility of continuous tuning of the critical exponents by adjusting the longitudinal fields differently from the scaling limit. For the Heisenberg model, we obtained results numerically comparable to established results with simple calculations. The VES approach provides a powerful and versatile tool for studying interacting quantum systems.

cond-mat.str-el↗

arXiv2023

Giant Nernst effect in the crossover between Fermi liquid and strange metal

The strange-metal state is a crucial problem in condensed matter physics highlighted by its ubiquity in almost all major correlated systems[1-7]. Its understanding could provide important insight into high-Tc superconductivity[2] and quantum criticality[8]. However, with the Fermi liquid theory failing in strange metals, understanding the highly unconventional behaviors has been a long-standing challenge. Fundamental aspects of strange metals remain elusive, including the nature of their charge carriers[1]. Here, we report the observation of a giant Nernst response in the strange-metal state in a two-dimensional superconductor 2M-WS2. A giant Nernst coefficient comparable to the vortex Nernst signal in superconducting cuprates, and its high sensitivity to carrier mobility, are found when the system enters the strange-metal state from the Fermi liquid state. The temperature and magnetic field dependence of the giant Nernst peak rule out the relevance of both Landau quasiparticles and superconductivity. Instead, the giant Nernst peak at the crossover indicates a dramatic change in carrier entropy when entering the strange-metal state. The presence of such an anomalous Nernst response is further confirmed in other iconic strange metals, suggesting its universality and places stringent experimental constraints on the mechanism of strange metals.

cond-mat.str-el↗

arXiv2022

Post-breach Recovery: Protection against White-box Adversarial Examples for Leaked DNN Models

Server breaches are an unfortunate reality on today's Internet. In the context of deep neural network (DNN) models, they are particularly harmful, because a leaked model gives an attacker "white-box" access to generate adversarial examples, a threat model that has no practical robust defenses. For practitioners who have invested years and millions into proprietary DNNs, e.g. medical imaging, this seems like an inevitable disaster looming on the horizon. In this paper, we consider the problem of post-breach recovery for DNN models. We propose Neo, a new system that creates new versions of leaked models, alongside an inference time filter that detects and removes adversarial examples generated on previously leaked models. The classification surfaces of different model versions are slightly offset (by introducing hidden distributions), and Neo detects the overfitting of attacks to the leaked model used in its generation. We show that across a variety of tasks and attack methods, Neo is able to filter out attacks from leaked models with very high accuracy, and provides strong protection (7--10 recoveries) against attackers who repeatedly breach the server. Neo performs well against a variety of strong adaptive attacks, dropping slightly in # of breaches recoverable, and demonstrates potential as a complement to DNN defenses in the wild.

cs.CR↗

arXiv2022

Spontaneous magnetism after a "remote" quench: a proposal to test Copenhagen interpretation

A change of quantum states for a quantum particle may lead to a change of physical field it exerts to the environment. We discuss such Gedankenexperiment for measuring the magnetic dipole fields associated with the electronic spins. When entangled, electrons are no longer free Dirac electrons but become an excited state in a quantum-electrodynamics sense. A measurement of magnetism associated with entanglement-collapse of electrons becomes a test for Copenhagen interpretation in the reign of quantum-electrodynamics. This proposal is equally applicable to other particles and their associated fields and interactions, such as neutrons and the electroweak force.

quant-ph↗

arXiv2022

Algebraic-Dynamical Theory for Quantum Spin-1/2: the Two-spin limit and Implications for Lattice Models

Recently, an {\it algebraic-dynamical theory} (ADT) for strongly interacting many-body quantum Hamiltonians in W. Ding, arXiv: 2202.12082 (2022). By introducing the complete operator basis set, ADT proposes a generic framework for systematically constructing dynamical theories for interacting quantum Hamiltonians, using quantum entanglement as the organizing principle. In this work, we study exact ADT solutions of interacting two-spin problems which can be used as "free theories" for perturbation study on relevant solvable limits. Then we perform ADT perturbation calculations and obtain the correct ground state under the perturbation, which shows that the ADT framework is capable of constructing correct dynamical perturbation theories for strongly interacting quantum spin models. We also discuss the implication for relevant lattice models.

cond-mat.str-el↗

arXiv2022

Algebraic-Dynamical Theory for Quantum Many-body Hamiltonians: A Formalized Approach To Strongly Interacting Systems

Non-commutative algebras and entanglement are two of the most important hallmarks of many-body quantum systems. Dynamical perturbation methods are the most widely used approaches for quantum many-body systems. While study of entanglement-based numerical methods are booming recently, the traditional dynamical perturbation methods have not benefited from study of quantum entanglement. In this work, we formulate an algebraic-dynamical theory (ADT) by combining the power of quantum algebras and dynamical methods in which quantum entanglement naturally emerges as the organizing principle. We start by introducing a complete operator basis set (COBS), with which an arbitrary state, either pure or mixed, can be represented by the expectation values of COBS. Then we establish a complete mapping from a given state to a complete set of dynamical correlation functions of the state through the Heisenberg- and Schwinger-Dyson-equations-of-motion (SDEOM). The completeness of COBS and the mapping ensures ADT to be a mathematically complete framework in principle. Applying ADT to many-body systems on lattices, we find that the quantum entanglement is represented by the cumulant structure of expectation values of the many-body COBS. The cumulant structure of the state forms a hierarchy in correlations. More importantly, such static correlational hierarchy is inherited by the dynamical correlations and their SDEOM. We propose that the dynamical hierarchy is also carried into any perturbative calculation on that state. We demonstrate the validity of such perturbation hierarchy with an explicit example, in which we show that a single-particle-type perturbative calculation fails while a many-body perturbation following the hierarchy succeeds. We also discuss the computation and approximation schemes of ADT and its implications to other strong coupling theories like parton and slave particle methods.

quant-ph↗

arXiv2022

Calibration with Privacy in Peer Review

Reviewers in peer review are often miscalibrated: they may be strict, lenient, extreme, moderate, etc. A number of algorithms have previously been proposed to calibrate reviews. Such attempts of calibration can however leak sensitive information about which reviewer reviewed which paper. In this paper, we identify this problem of calibration with privacy, and provide a foundational building block to address it. Specifically, we present a theoretical study of this problem under a simplified-yet-challenging model involving two reviewers, two papers, and an MAP-computing adversary. Our main results establish the Pareto frontier of the tradeoff between privacy (preventing the adversary from inferring reviewer identity) and utility (accepting better papers), and design explicit computationally-efficient algorithms that we prove are Pareto optimal.

cs.CR↗