Searcharxiv⌕ Search

arXiv subjects

Christian Rose

Publications and source records attributed to Christian Rose.

At least 19 recordsLinked to original sources
arXiv2026

Gaussian upper bounds for averaged heat semigroups on graphs

Characterizations of pointwise Gaussian upper bounds on graphs with possibly unbounded geometry in terms of localized functional inequalities contain errors depending on the vertex degree. We introduce a new space-time averaged form of the heat semigroup in terms of time-averaged $\ell^{q/(q-1)}-\ell^q$-estimates and obtain Gaussian upper bounds from large-scale Faber-Krahn inequalities which avoid such errors. A main analytic ingredient is an integrated version of Davies' method which yields off-diagonal estimates for this averaged quantity, and which tends to pointwise Gaussian bounds as time tends to infinity. Conversely, on large scales volume doubling and Gaussian bounds on this averaged heat semigroup norms imply relative Faber-Krahn inequalities for subsets of prescribed relative measure. Our proof is based on a lower bound for Dirichlet eigenvalues in terms of these averages. The Faber-Krahn dimension is scale-dependent, but converges to the doubling dimension for increasing radii. This gives a characterization of asymptotic Gaussian heat kernel behavior in terms of Faber-Krahn inequalities.

math.AP↗
arXiv2026

SycoEval-EM: Sycophancy Evaluation of Large Language Models in Simulated Clinical Encounters for Emergency Care

Large language models (LLMs) deployed in clinical decision support may acquiesce to patient requests for care that conflicts with evidence-based guidelines. We developed SycoEval-EM, a multi-agent simulation framework to evaluate LLM robustness to adversarial patient persuasion in emergency medicine. Across 19 contemporary LLMs and 1,425 simulated clinical encounters spanning three Choosing Wisely scenarios, acquiescence rates ranged from 0% to 100%, revealing a bimodal distribution. Seven models maintained near-perfect guideline adherence, while six acquiesced in the majority of encounters. Vulnerability varied substantially across clinical scenarios. Acquiescence was highest for CT imaging requests, intermediate for antibiotic prescriptions for sinusitis, and lowest for opioid prescriptions for acute back pain. Model scale, recency, and performance on static medical benchmarks did not consistently predict robustness. All five persuasion tactics produced similar acquiescence rates, with no statistically significant differences after correction for multiple comparisons, suggesting a generalized susceptibility rather than tactic-specific weaknesses. LLM-as-judge evaluation was validated against two independent physician raters across 95 matched conversations and demonstrated near-perfect agreement for the primary outcome of acquiescence (Cohens kappa = 0.957). These findings indicate that static medical benchmarks are insufficient to predict safety performance under sustained social pressure and support incorporating multi-turn adversarial testing into clinical AI evaluation. Notably, two models achieved perfect guideline adherence across all encounters, demonstrating that robustness to patient pressure is attainable without sacrificing effective clinical communication.

cs.AI↗
arXiv2025

Off-diagonal upper heat kernel bounds on graphs with unbounded geometry

Results regarding off-diagonal Gaussian upper heat kernel bounds on discrete weighted graphs with possibly unbounded geometry are summarized and related. After reviewing uniform upper heat kernel bounds obtained by Carlen, Kusuoka, and Stroock, the universal Gaussian term on graphs found by Davies is addressed and related to corresponding results in terms of intrinsic metrics. Then we present a version of Grigor'yan's two-point method with Gaussian term involving an intrinsic metric. A discussion of upper heat kernel bounds for graph Laplacians with possibly unbounded but integrable weights on bounded combinatorial graphs preceeds the presentation of compatible bounds for anti-trees, an example of combinatorial graph with unbounded Laplacian. Characterizations of localized heat kernel bounds in terms of intrinsic metrics and universal Gaussian are reconsidered. Finally, the problem of optimality of the Gaussian term is discussed by relating Davies' optimal metric with the supremum over all intrinsic metrics.

math.AP↗
arXiv2025

Optimal Poincar\'e-Hardy-type Inequalities on Manifolds and Graphs

We review a method to obtain optimal Poincar\'e-Hardy-type inequalities on the hyperbolic spaces, and discuss briefly generalisations to certain classes of Riemannian manifolds. Afterwards, we recall a corresponding result on homogeneous regular trees and provide a new proof using the aforementioned method. The same strategy will then be applied to obtain new optimal Hardy-type inequalities on weakly spherically symmetric graphs which include fast enough growing trees and anti-trees. In particular, this yields optimal weights which are larger at infinity than the optimal weights classically constructed via the Fitzsimmons ratio of the square root of the minimal positive Green's function.

math.AP↗
arXiv2024

Spectral comparison results for Laplacians on discrete graphs

In the recent literature, various authors have studied spectral comparison results for Schr\"odinger operators with discrete spectrum in different settings including Euclidean domains and quantum graphs. In this note we derive such spectral comparison results in a rather general framework for general and possibly infinite discrete graphs. Along the way, we establish a discrete version of the local Weyl law whose proof does neither involve any Tauberian theorem nor the Weyl law as used in the continuous case.

math.SP↗
arXiv2024

Gaussian upper heat kernel bounds and Faber-Krahn inequalities on graphs

We investigate the equivalence of relative Faber-Krahn inequalities and the conjunction of Gaussian upper heat kernel bounds and volume doubling on large scales on graphs. For the normalizing measure, we obtain their equivalence up to constants by imposing comparability of small balls and the vertex degree at their centers. Removing this comparability assumption on the cost of a local regularity condition entering the equivalence and allowing for a variable dimension lead to a further generalization. The variable dimension converges to the doubling dimension for increasing ball radius. If the counting measure or arbitrary measures are considered, the local regularity condition contains the vertex degree. Furthermore, correction functions for the Gaussian, doubling, and Faber-Krahn dimension depending on the vertex degree are introduced. For the Gaussian and doubling, the variable correction functions always tend to one at infinity. Moreover, the variable Faber-Krahn dimension can be related to the doubling dimension and the vertex degree growth.

math.AP↗
arXiv2024

Gaussian upper bounds, volume doubling and Sobolev inequalities on graphs

We investigate the equivalence of Sobolev inequalities and the conjunction of Gaussian upper heat kernel bounds and volume doubling on large scales on graphs. For the normalizing measure, we obtain their equivalence up to constants by imposing comparability of small balls and the vertex degree at their centers. If arbitrary measures are considered, we incorporate a new local regularity condition. Furthermore, new correction functions for the Gaussian, doubling, and Sobolev dimension are introduced. For the Gaussian and doubling, the variable correction functions always tend to one at infinity. Moreover, the variable Sobolev dimension can be related to the doubling dimension and the vertex degree growth.

math.AP↗
arXiv2023

Unique continuation estimates on manifolds with Ricci curvature bounded below

We prove quantitative unique continuation estimates for relatively dense sets and spectral subspaces associated to small energies of Schr\"odinger operators on Riemannian manifolds with Ricci curvature bounded below. The upper bound for the energy range and the constant appearing in the estimate are given in terms of the lower bound of the Ricci curvature and the parameters of the relatively dense set.

math.AP↗
arXiv2022

Anchored heat kernel upper bounds on graphs with unbounded geometry and anti-trees

We derive Gaussian heat kernel bounds on graphs with respect to a fixed origin for large times under the assumption of a Sobolev inequality and volume doubling on large balls. The upper bound from our previous work [KR22] is affected by a new correction term measuring the distance to the origin. The main result is then applied to anti-trees with unbounded vertex degree, yielding Gaussian upper bounds for this class of graphs for the first time. In order to prove this, we show that isoperimetric estimates with respect to intrinsic metrics yield Sobolev inequalities. Finally, we prove that anti-trees are Ahlfors regular and that they satisfy an isoperimetric inequality of a larger dimension.

math.AP↗
arXiv2022

Gaussian upper bounds for heat kernels on graphs with unbounded geometry

We prove large-time Gaussian upper bounds for continuous-time heat kernels of Laplacians on graphs with unbounded geometry. Our estimates hold for centers of large balls satisfying a Sobolev inequality and volume doubling. Distances are measured with respect to an intrinsic metric with finite distance balls and finite jump size. The Gaussian decay is given by Davies' function which is natural and sharp in the graph setting. Furthermore, we find a new polynomial correction term which does not blow up at zero. Although our main focus is unbounded Laplacians, the results are new even for the normalized Laplacian. In the case of unbounded vertex degree or degenerating measure, the estimates are affected by new error terms reflecting the unboundedness of the geometry.

math.AP↗
arXiv2021

Integral Ricci curvature and the mass gap of Dirichlet Laplacians on domains

We obtain a fundamental gap estimate for classes of bounded domains with quantitative control on the boundary in a complete manifold with integral bounds on the negative part of the Ricci curvature. This extends the result of \cite{Oden-Sung-Wang99} to $L^p$-Ricci curvature assumptions, $p>n/2$. To achieve our result, it is shown that the domains under consideration are John domains, what enables us to obtain an estimate on the first nonzero Neumann eigenvalue, which is of independent interest.

math.DG↗
arXiv2020

Quantitative unique continuation for spectral subspaces of Schr\"odinger operators with singular potentials

Recent (scale-free) quantitative unique continuation estimates for spectral subspaces of Schr\"odinger operators are extended to allow singular potentials such as certain $L^p$-functions. The proof is based on accordingly adapted Carleman estimates. Applications include Wegner and initial length scale estimates for random Schr\"odinger operators and control theory for the controlled heat equation with singular heat generation term.

math.AP↗
arXiv2020

Quantitative Sobolev extensions and the Neumann heat kernel for integral Ricci curvature conditions

We prove the existence of Sobolev extension operators for certain uniform classes of domains in a Riemannian manifold with an explicit uniform bound on the norm depending only on the geometry near their boundaries. We use this quantitative estimate to obtain uniform Neumann heat kernel upper bounds and gradient estimates for positive solutions of the Neumann heat equation assuming integral Ricci curvature conditions and geometric conditions on the boundary. Those estimates also imply quantitative lower bounds on the first Neumann eigenvalue of the considered domains.

math.DG↗
arXiv2020

Eigenvalue estimates for Kato-type Ricci curvature conditions

We prove that optimal lower eigenvalue estimates of Zhong-Yang type as well as a Cheng-type upper bound for the first eigenvalue hold on closed manifolds assuming only a Kato condition on the negative part of the Ricci curvature. This generalizes all earlier results on $L^p$-curvature assumptions. Moreover, we introduce the Kato condition on compact manifolds with boundary with respect to the Neumann Laplacian, leading to Harnack estimates for the Neumann heat kernel and lower bounds for all Neumann eigenvalues, what provides a first insight in handling variable Ricci curvature assumptions in this case.

math.DG↗
arXiv2019

Spectrally positive Bakry-\'Emery Ricci curvature on graphs

We investigate analytic and geometric implications of non-constant Ricci curvature bounds. We prove a Lichnerowicz eigenvalue estimate and finiteness of the fundamental group assuming that $L+2 Ric$ is a positive operator where $L$ is the graph Laplacian. Assuming that the negative part of the Ricci curvature is small in Kato sense, we prove diameter bounds, elliptic Harnack inequality and Buser inequality. This article seems to be the first one establishing these results while allowing for some negative curvature.

math.DG↗
arXiv2019

Almost positive Ricci curvature in Kato sense -- an extension of Myers' theorem

It is shown that if the Kato constant of the negative part of the Ricci curvature below a positive level is small, then the volume of the corresponding manifold can be bounded above in terms of the Kato constant and the total Ricci curvature. Together with the results from [5] and [6], this yields a generalization of the famous Bonnet-Myers theorem. Connections to some earlier generalizations are discussed.

math.DG↗