arXiv subjects

Yuan

Publications and source records attributed to Yuan.

11 recordsLinked to original sources

arXiv2026

READY or Not: Reliable Enterprise Agent Deployment

An AI agent can perform well on benchmarks and still be unsuitable for deployment. Existing AI-agent benchmarks measure whether an agent can complete realistic professional work, whereas enterprise deployment asks a different question: whether an agent can meet a required reliability level, under acceptable human oversight, and at tolerable cost. We introduce Reliable Enterprise Agent Deployment (READY), a framework for qualifying AI agents for deployment on enterprise workflows. READY preserves each workflow's own definition of successful execution while applying a common qualification procedure. Given an agent, a workflow, and a class of candidate oversight policies, READY measures the reliability and operating cost of the human-AI system, selects the minimum-cost policy that satisfies a specified reliability target, and statistically qualifies it on held-out cases. The resulting deployment profile characterizes the supported operating point: reliability, human-oversight burden, and cost. READY is implemented as an open testbed that decouples workflow specification, execution, evaluation, and qualification, and runs on existing agent-evaluation infrastructure. In an end-to-end clinical-audit case study spanning 16 agent systems and 750 cases, READY reveals differences hidden by autonomous performance: two systems separated by only 0.3 points in autonomous accuracy (72.8% vs. 72.5%) require 39.2% versus 29.6% human review, respectively, to qualify at the same 76% reliability target under the evaluated oversight policy. READY thus shifts enterprise agent evaluation from how well can the agent perform the work? to under what conditions, and at what cost, can it be reliably deployed? By making those conditions explicit and statistically testable, READY provides a basis for comparing agent systems, setting oversight requirements, and making evidence-based deployment decisions.

cs.AI↗

arXiv2026

Scratched Lenses, Shifted Depth: Passive Camera-Side Optical Attacks

Physical adversarial attacks on vision systems are typically studied through scene manipulation, such as adversarial patches or projections, where the adversary controls what the camera observes. Camera-side attacks using stickers or auxiliary optics have also been explored, but they treat attacks as image-space perturbations from designed patterns. This misses how physical imperfections interact with scene-dependent lighting and optics. We identify a threat: passive lens-side damage that is persistent yet trigger-conditioned, producing optical artifacts that bias geometric inference under particular visual conditions. We instantiate this threat through Scratch-induced Lens Adversarial Streak Hijacking SLASH, a physical-world attack caused by small scratches on a camera lens or protective cover. Scratches interact with bright light sources and specular reflections to create structured streak artifacts that distort depth cues. Since the perturbation is fixed in the optical path but triggered by the scene, it is both persistent and selective. We formulate the attack in optical space, model the scratch pattern as a trigger-conditioned optical channel, and optimize one fixed configuration across diverse viewing conditions. We evaluate SLASH on monocular depth estimation and monocular 3D object detection in digital and real-world settings. Under the fixed-scratch constraint, directional depth shifts reach up to 32% relative error for monocular depth estimation, with consistent effects on monocular 3D object detection. Physical experiments confirm transfer to real camera recordings, inducing depth shifts above the model's natural prediction baseline. These findings reveal an attack surface where benign-looking hardware imperfections act as latent, scene-triggered adversarial mechanisms, challenging assumptions about physical robustness and motivating defenses for secure vision systems.

cs.CV↗

arXiv2026

AdvScene: Rethinking Adversarial Patch Evaluation Through Scene Robustness

Adversarial patches are physical patterns attached to real objects to mislead AI vision systems. Their real-world risk is not determined by a single successful prediction, but by whether they remain effective after deployment under changing viewpoints, distances, and scene conditions. We refer to this property as scene robustness, the effectiveness of a deployed patch across conditions in a real environment. Yet existing evaluations do not measure scene robustness well: real image benchmarks are realistic but fixed, while simulators are controllable but not grounded in a specific real scene. We present AdvScene, a scene-grounded framework for measuring the scene robustness of adversarial patches in reconstructed real environments. AdvScene reframes evaluation as operational measurement: given a fixed deployed patch, it characterizes the patch's operational envelope - where and when the attack succeeds - as a function of viewpoint, distance, and scene context. A key challenge is that the attack is typically defined only in a single anchor view, while evaluation requires a representation that remains faithful under viewpoint changes. We formalize this as a constrained lifting problem and introduce Adversarial Patch-to-Scene Embedding (APSE), which resolves cross-view ambiguity while preserving attack-critical appearance and enforcing locality, target-surface attachment, and cross-view consistency. We validate AdvScene using real-world physical data and conduct a comprehensive evaluation of existing adversarial patches. Our results show that AdvScene reveals substantial scene-dependent variation in attack effectiveness that is not captured by existing image-centric or simulator-based evaluations.

cs.CR↗

arXiv2023

Generating Signed Permutations by Twisting Two-Sided Ribbons

We provide a simple and natural solution to the problem of generating all $2^n \cdot n!$ signed permutations of $[n] = \{1,2,\ldots,n\}$. Our solution provides a pleasing generalization of the most famous ordering of permutations: plain changes (Steinhaus-Johnson-Trotter algorithm). In plain changes, the $n!$ permutations of $[n]$ are ordered so that successive permutations differ by swapping a pair of adjacent symbols, and the order is often visualized as a weaving pattern involving $n$ ropes. Here we model a signed permutation using $n$ ribbons with two distinct sides, and each successive configuration is created by twisting (i.e., swapping and turning over) two neighboring ribbons or a single ribbon. By greedily prioritizing $2$-twists of the largest symbol before $1$-twists of the largest symbol, we create a signed version of plain change's memorable zig-zag pattern. We provide a loopless algorithm (i.e., worst-case $\mathcal{O}(1)$-time per object) by extending the well-known mixed-radix Gray code algorithm.

cs.DS↗

arXiv2023

High-resolution acoustic field mapping of GHz phononic crystals with atomic force microscopy

On-chip technology based on acoustic waves is a strong asset in modern telecommunication with the prospects of becoming a cornerstone of next-generation devices. In this context, mapping and manipulating acoustic waves through coherent scattering is pivotal for a non-trivial control of the flow of acoustic energy, which could consequently enable advanced information manipulation. To this end, here a technique for mapping acoustic fields is introduced and used for characterizing $\mu$m-sized phononic crystals defined in GaAs slabs and excited by GHz surface acoustic waves (SAWs) based on atomic force microscopy. It is shown that incoherent scattering excites a wide distribution of modes, which enables the mapping of the dispersion relation of the two-dimensional structures, while the phononic crystal symmetry directly correlates with coherent scattering effects. Enabling the use of acoustic atomic force microscopy and understanding the role of scattering are of paramount importance for the versatile use of GHz acoustic waves in technological applications, setting the baseline for advanced operations like hyperspectral filtering, beam steering or spatial-division multiplexing.

physics.app-ph↗

arXiv2022

Chemo-dynamical substructure in the M31 inner halo globular clusters: Further evidence for a recent accretion event

Based upon a metallicity selection, we identify a significant sub-population of the inner halo globular clusters in the Andromeda Galaxy which we name the Dulais Structure. It is distinguished as a co-rotating group of 10-20 globular clusters which appear to be kinematically distinct from, and on average more metal-poor than, the majority of the inner halo population. Intriguingly, the orbital axis of this Dulais Structure is closely aligned with that of the younger accretion event recently identified using a sub-population of globular clusters in the outer halo of Andromeda, and this is strongly suggestive of a causal relationship between the two. If this connection is confirmed, a natural explanation for the kinematics of the globular clusters in the Dulais Structure is that they trace the accretion of a substantial progenitor (~10^11 Msun) into the halo of Andromeda during the last few billion years, that may have occurred as part of a larger group infall.

astro-ph.GA↗

arXiv2018

Eigenstates Transition Without Undergoing an Adiabatic Process

We introduce a class of non-Hermitian Hamiltonians that offers a dynamical approach to short-cut to adiabaticity (DASA). In particular, in our proposed 2 * 2 Hamiltonians, one eigenvalue is absolutely real and the other one is complex. This specific form of the eigenvalues helps us to exponentially decay the population in an undesired eigenfunction or amplify the population in the desired state while keeping the probability amplitude in the other eigenfunction conserved. This provides us with a powerful method to have a diabatic process with the same outcome as its corresponding adiabatic process. In contrast to standard shortcuts to adiabaticity, our Hamiltonians have a much simpler form with a lower thermodynamic cost. Furthermore, we show that DASA can be extended to higher dimensions using the parameters associated with our 2 * 2 Hamiltonians. Our proposed Hamiltonians not only have application in DASA but also can be used for tunable mode selection and filtering in acoustics, electronics, and optics.

quant-ph↗

arXiv2012

Virtual Vector Machine for Bayesian Online Classification

In a typical online learning scenario, a learner is required to process a large data stream using a small memory buffer. Such a requirement is usually in conflict with a learner's primary pursuit of prediction accuracy. To address this dilemma, we introduce a novel Bayesian online classi cation algorithm, called the Virtual Vector Machine. The virtual vector machine allows you to smoothly trade-off prediction accuracy with memory size. The virtual vector machine summarizes the information contained in the preceding data stream by a Gaussian distribution over the classi cation weights plus a constant number of virtual data points. The virtual data points are designed to add extra non-Gaussian information about the classi cation weights. To maintain the constant number of virtual points, the virtual vector machine adds the current real data point into the virtual point set, merges two most similar virtual points into a new virtual point or deletes a virtual point that is far from the decision boundary. The information lost in this process is absorbed into the Gaussian distribution. The extra information provided by the virtual points leads to improved predictive accuracy over previous online classification algorithms.

cs.LG↗

arXiv2012

Sparse-posterior Gaussian Processes for general likelihoods

Gaussian processes (GPs) provide a probabilistic nonparametric representation of functions in regression, classification, and other problems. Unfortunately, exact learning with GPs is intractable for large datasets. A variety of approximate GP methods have been proposed that essentially map the large dataset into a small set of basis points. Among them, two state-of-the-art methods are sparse pseudo-input Gaussian process (SPGP) (Snelson and Ghahramani, 2006) and variablesigma GP (VSGP) Walder et al. (2008), which generalizes SPGP and allows each basis point to have its own length scale. However, VSGP was only derived for regression. In this paper, we propose a new sparse GP framework that uses expectation propagation to directly approximate general GP likelihoods using a sparse and smooth basis. It includes both SPGP and VSGP for regression as special cases. Plus as an EP algorithm, it inherits the ability to process data online. As a particular choice of approximating family, we blur each basis point with a Gaussian distribution that has a full covariance matrix representing the data distribution around that basis point; as a result, we can summarize local data manifold information with a small set of basis points. Our experiments demonstrate that this framework outperforms previous GP classification methods on benchmark datasets in terms of minimizing divergence to the non-sparse GP solution as well as lower misclassification rate.

cs.LG↗

arXiv2012

Sparse matrix-variate Gaussian process blockmodels for network modeling

We face network data from various sources, such as protein interactions and online social networks. A critical problem is to model network interactions and identify latent groups of network nodes. This problem is challenging due to many reasons. For example, the network nodes are interdependent instead of independent of each other, and the data are known to be very noisy (e.g., missing edges). To address these challenges, we propose a new relational model for network data, Sparse Matrix-variate Gaussian process Blockmodel (SMGB). Our model generalizes popular bilinear generative models and captures nonlinear network interactions using a matrix-variate Gaussian process with latent membership variables. We also assign sparse prior distributions on the latent membership variables to learn sparse group assignments for individual network nodes. To estimate the latent variables efficiently from data, we develop an efficient variational expectation maximization method. We compared our approaches with several state-of-the-art network models on both synthetic and real-world network datasets. Experimental results demonstrate SMGBs outperform the alternative approaches in terms of discovering latent classes or predicting unknown interactions.

cs.LG↗

arXiv2011

Infinite Tucker Decomposition: Nonparametric Bayesian Models for Multiway Data Analysis

Tensor decomposition is a powerful computational tool for multiway data analysis. Many popular tensor decomposition approaches---such as the Tucker decomposition and CANDECOMP/PARAFAC (CP)---amount to multi-linear factorization. They are insufficient to model (i) complex interactions between data entities, (ii) various data types (e.g. missing data and binary data), and (iii) noisy observations and outliers. To address these issues, we propose tensor-variate latent nonparametric Bayesian models, coupled with efficient inference methods, for multiway data analysis. We name these models InfTucker. Using these InfTucker, we conduct Tucker decomposition in an infinite feature space. Unlike classical tensor decomposition models, our new approaches handle both continuous and binary data in a probabilistic framework. Unlike previous Bayesian models on matrices and tensors, our models are based on latent Gaussian or $t$ processes with nonlinear covariance functions. To efficiently learn the InfTucker from data, we develop a variational inference technique on tensors. Compared with classical implementation, the new technique reduces both time and space complexities by several orders of magnitude. Our experimental results on chemometrics and social network datasets demonstrate that our new models achieved significantly higher prediction accuracy than the most state-of-art tensor decomposition

cs.LG↗