arXiv subjects

Yihan Shen

Publications and source records attributed to Yihan Shen.

6 recordsLinked to original sources

arXiv2026

The Phantom of PCIe: Constraining Generative Artificial Intelligences for Practical Peripherals Trace Synthesizing

Peripheral Component Interconnect Express (PCIe) is the de facto interconnect standard for high-speed peripherals and CPUs. The development of PCIe devices for emerging applications requires realistic Transaction Layer Packet (TLP) traces that accurately simulate device-CPU interactions. While generative AI offers a promising avenue for synthesizing complex TLP sequences, it is prone to a critical challenge inherent in all generation tasks: hallucination. Naively applying these models often produces traces that violate fundamental PCIe protocol rules, such as ordering and causality, rendering them unusable for device simulation. To resolve this, our work introduces a methodology to bridge the gap between generative AI and high-fidelity device simulation. This paper presents Phantom, a framework that systematically addresses AI-generated hallucinations in TLP synthesis. Phantom achieves this by coupling a generative backbone with a novel post-processing filter that enforces PCIe-specific constraints, effectively eliminating invalid TLP sequences. We validate Phantom's effectiveness by synthesizing TLP traces for an actual PCIe network interface card. Experimental results show that Phantom produces practical, large-scale TLP traces, significantly outperforming existing models, with improvements of up to 1000$\times$ in task-specific metrics and up to 2.19$\times$ in Fréchet Inception Distance (FID) compared to backbone-only methods. The prototype implementation has been made open-source.

cs.LG↗

arXiv2026

Superradiant Suppression of Non-minimally Coupled Scalar fields for a Rotating Charged dS Black Hole in Conformal Weyl Gravity

In this study, we present an analytical investigation of the superradiant scattering of a massive charged conformally coupled scalar field in rotating charged $de~Sitter$ black hole spacetimes within two gravitational theories: General Relativity (GR) and fourth-order Conformal (Weyl-squared) Gravity (CWG). For the massless charged conformally coupled scalar, we exploit a recently discovered correspondence between the Heun equation and the semiclassical limit of Belavin-Polyakov-Zamolodchikov (BPZ) equations in two-dimensional conformal field theory to solve for the superradiant amplification factors as controlled expansions in a small parameter scaling. For the massive charged conformally coupled scalar, we use WKB methods to derive an order of magnitude approximation for the amplification factors in the cosmological region in terms of those in the region $r_+\ll r \ll r_c$ where $r_+$ and $r_c$ are the outer and cosmological event horizons, respectively. For both the massless and massive sectors, suppression of superradiant amplification in CWG relative to that in GR is observed across the parameter regimes studied. Particularly, in the massive sector, we find strong exponential suppression of superradiant amplification on the order of $e^{-2μΛ^{-1/2}}$ in the cosmological region.

hep-th↗

arXiv2025

Fast, Accurate Manifold Denoising by Tunneling Riemannian Optimization

Learned denoisers play a fundamental role in various signal generation (e.g., diffusion models) and reconstruction (e.g., compressed sensing) architectures, whose success derives from their ability to leverage low-dimensional structure in data. Existing denoising methods, however, either rely on local approximations that require a linear scan of the entire dataset or treat denoising as generic function approximation problems, often sacrificing efficiency and interpretability. We consider the problem of efficiently denoising a new noisy data point sampled from an unknown $d$-dimensional manifold $M \in \mathbb{R}^D$, using only noisy samples. This work proposes a framework for test-time efficient manifold denoising, by framing the concept of "learning-to-denoise" as "learning-to-optimize". We have two technical innovations: (i) online learning methods which learn to optimize over the manifold of clean signals using only noisy data, effectively "growing" an optimizer one sample at a time. (ii) mixed-order methods which guarantee that the learned optimizers achieve global optimality, ensuring both efficiency and near-optimal denoising performance. We corroborate these claims with theoretical analyses of both the complexity and denoising performance of mixed-order traversal. Our experiments on scientific manifolds demonstrate significantly improved complexity-performance tradeoffs compared to nearest neighbor search, which underpins existing provable denoising approaches based on exhaustive search.

cs.LG↗

arXiv2025

Local Averaging Accurately Distills Manifold Structure From Noisy Data

High-dimensional data are ubiquitous, with examples ranging from natural images to scientific datasets, and often reside near low-dimensional manifolds. Leveraging this geometric structure is vital for downstream tasks, including signal denoising, reconstruction, and generation. However, in practice, the manifold is typically unknown and only noisy samples are available. A fundamental approach to uncovering the manifold structure is local averaging, which is a cornerstone of state-of-the-art provable methods for manifold fitting and denoising. However, to the best of our knowledge, there are no works that rigorously analyze the accuracy of local averaging in a manifold setting in high-noise regimes. In this work, we provide theoretical analyses of a two-round mini-batch local averaging method applied to noisy samples drawn from a $d$-dimensional manifold $\mathcal M \subset \mathbb{R}^D$, under a relatively high-noise regime where the noise size is comparable to the reach $τ$. We show that with high probability, the averaged point $\hat{\mathbf q}$ achieves the bound $d(\hat{\mathbf q}, \mathcal M) \leq σ\sqrt{d\left(1+\frac{κ\mathrm{diam}(\mathcal {M})}{\log(D)}\right)}$, where $σ, \mathrm{diam(\mathcal M)},κ$ denote the standard deviation of the Gaussian noise, manifold's diameter and a bound on its extrinsic curvature, respectively. This is the first analysis of local averaging accuracy over the manifold in the relatively high noise regime where $σ\sqrt{D} \approx τ$. The proposed method can serve as a preprocessing step for a wide range of provable methods designed for lower-noise regimes. Additionally, our framework can provide a theoretical foundation for a broad spectrum of denoising and dimensionality reduction methods that rely on local averaging techniques.

stat.ML↗

arXiv2024

Variational Discretizations for Hamiltonian Systems

In this paper, we study the Lagrangian functions for a class of second-order differential systems arising from physics. For such systems, we present necessary and sufficient conditions for the existence of Lagrangian functions. Based on the variational principle and the splitting technique, we construct variational integrators and prove their equivalence to the composition of explicit symplectic methods. We apply the newly derived variational integrators to the Kepler problem and demonstrate their effectiveness in numerical simulations. Moreover, using the modified Lagrangian, we analyze the dynamical behavior of the numerical solutions in preserving the Laplace--Runge--Lenz (LRL) vector.

math.NA↗

arXiv2020

Can single crystal X-ray diffraction determine a structure uniquely?

The diffraction technique is widely used in the determination of crystal structures and is one of the bases for the modern science and technology. All related structure determination methods are based on the assumption that perfect single crystal X-ray diffraction (SXRD) can determine a structure uniquely. But as the structure factor phases are lost in SXRD and even more information is lost in powder X-ray diffraction (PXRD), this assumption is still questionable. In this work, we found that structures with certain characteristic can have its sister structure with exactly the same PXRD or even SXRD pattern. A computer program is developed to search the ICSD database, and about 1000 structures were identified to have this characteristic. The original structure and its sister structures can have different space groups, topologies, crystal systems etc. and some may even have multiple sisters. Our studies indicate that special caution is needed since a structure with reasonable atomic positions and perfect match of experimental diffraction intensities could still be wrong.

cond-mat.mtrl-sci↗