arXiv subjects

David Bowman

Publications and source records attributed to David Bowman.

10 recordsLinked to original sources

arXiv2026

Harnack inequality for non-uniformly elliptic equations in non-divergence form

We study regularity properties for solutions to the nakedly degenerate elliptic equation $a_{ij}\partial_{ij}u =0$, where the coefficients satisfy $I \ge a_{ij}(x) \ge \lambda(x) I$ and the only assumption is that $\lambda^{-1} \in L^p$. We prove an improvement of oscillation and a Liouville theorem for $p>d-1$, and a Harnack inequality for $p$ sufficiently large depending on dimension. Along the way, we obtain a new $\log-L^\varepsilon$ Weak Harnack inequality for supersolutions. Then, touching subsolutions by double exponential blow-up barriers, we also derive a logarithmic local maximum principle that is new even in the uniformly elliptic case. Both of these results hold for $p>d-1$. Finally, we construct examples showing that there cannot be Harnack or Weak Harnack inequalities in the regime $p<d-1$, nor can there be power-type $L^\varepsilon$ inequalities in the case of any $p<\infty$.

math.AP↗

arXiv2025

Hypoelliptic Regularization in the Obstacle Problem for the Kolmogorov Operator

We study the obstacle problem associated with the Kolmogorov operator $\Delta_v - \partial_t - v\cdot\nabla_x$, which arises from the theory of optimal control in Asian-American options pricing models. Our first main contribution is to improve the known regularity of solutions, from $C^{0,1}_t \cap C^{0,2/3}_x \cap C^{1,1}_v$ to $C^{0,1}_{t,x} \cap C^{1,1}_v$. The previous result in the literature, which has been called optimal, corresponds to $C^{1,1}$ regularity with respect to the Kolmogorov distance. This is the expected regularity for solutions to obstacle problems. Our unexpected improvement of regularity in the $x$ variable is obtained using Bernstein's technique and an approach drawing on ideas from Evans-Krylov theory. We then use this improvement in regularity of the solution to prove the first known free boundary regularity result. We show that under a standard thickness condition, the free boundary is a $C^{0,1/2}_{t,x} \cap C^{0,1}_v$ regular surface. This result constitutes the first step in the program of free boundary regularity. Critically, our arguments rely on a new monotonicity formula and a commutator estimate that are only made possible by the solution's enhanced regularity in $x$.

math.AP↗

arXiv2024

Global Existence for an Isotropic Landau Model

Following the recent ideas of Guillen and Silvestre in $[9]$, we prove that the Fisher information is non-increasing along the flow of the isotropic Landau equation. We then use this fact to deduce global existence for the equation $\partial_t f = (-\Delta)^{-1}f \cdot \Delta f + f^2$ under a relatively lax set of conditions on the initial data. In particular, we remove the restrictive radially decreasing assumption of previous works.

math.AP↗

arXiv2023

On Autonomous Agents in a Cyber Defence Environment

Autonomous Cyber Defence is required to respond to high-tempo cyber-attacks. To facilitate the research in this challenging area, we explore the utility of the autonomous cyber operation environments presented as part of the Cyber Autonomy Gym for Experimentation (CAGE) Challenges, with a specific focus on CAGE Challenge 2. CAGE Challenge 2 required a defensive Blue agent to defend a network from an attacking Red agent. We provide a detailed description of the this challenge and describe the approaches taken by challenge participants. From the submitted agents, we identify four classes of algorithms, namely, Single- Agent Deep Reinforcement Learning (DRL), Hierarchical DRL, Ensembles, and Non-DRL approaches. Of these classes, we found that the hierarchical DRL approach was the most capable of learning an effective cyber defensive strategy. Our analysis of the agent policies identified that different algorithms within the same class produced diverse strategies and that the strategy used by the defensive Blue agent varied depending on the strategy used by the offensive Red agent. We conclude that DRL algorithms are a suitable candidate for autonomous cyber defence applications.

cs.CR↗

arXiv2021

Deep hierarchical reinforcement agents for automated penetration testing

Penetration testing the organised attack of a computer system in order to test existing defences has been used extensively to evaluate network security. This is a time consuming process and requires in-depth knowledge for the establishment of a strategy that resembles a real cyber-attack. This paper presents a novel deep reinforcement learning architecture with hierarchically structured agents called HA-DRL, which employs an algebraic action decomposition strategy to address the large discrete action space of an autonomous penetration testing simulator where the number of actions is exponentially increased with the complexity of the designed cybersecurity network. The proposed architecture is shown to find the optimal attacking policy faster and more stably than a conventional deep Q-learning agent which is commonly used as a method to apply artificial intelligence in automatic penetration testing.

cs.AI↗

arXiv2021

CybORG: A Gym for the Development of Autonomous Cyber Agents

Autonomous Cyber Operations (ACO) involves the development of blue team (defender) and red team (attacker) decision-making agents in adversarial scenarios. To support the application of machine learning algorithms to solve this problem, and to encourage researchers in this field to attend to problems in the ACO setting, we introduce CybORG, a work-in-progress gym for ACO research. CybORG features a simulation and emulation environment with a common interface to facilitate the rapid training of autonomous agents that can then be tested on real-world systems. Initial testing demonstrates the feasibility of this approach.

cs.CR↗

arXiv2020

CybORG: An Autonomous Cyber Operations Research Gym

Autonomous Cyber Operations (ACO) involves the consideration of blue team (defender) and red team (attacker) decision-making models in adversarial scenarios. To support the application of machine learning algorithms to solve this problem, and to encourage such practitioners to attend to problems in the ACO setting, a suitable gym (toolkit for experiments) is necessary. We introduce CybORG, a work-in-progress gym for ACO research. Driven by the need to efficiently support reinforcement learning to train adversarial decision-making models through simulation and emulation, our design differs from prior related work. Our early evaluation provides some evidence that CybORG is appropriate for our purpose and may provide a basis for advancing ACO research towards practical applications.

cs.CR↗

arXiv2016

Holomorphic Flexibility Properties of Spaces of Elliptic Functions

Let $X$ be an elliptic curve and $\mathbb{P}$ the Riemann sphere. Since $X$ is compact, it is a deep theorem of Douady that the set $\mathcal{O}(X,\mathbb{P})$ consisting of holomorphic maps $X\to \mathbb{P}$ admits a complex structure. If $R_n$ denotes the set of maps of degree $n$, then Namba has shown for $n\geq2$ that $R_n$ is a $2n$-dimensional complex manifold. We study holomorphic flexibility properties of the spaces $R_2$ and $R_3$. Firstly, we show that $R_2$ is homogeneous and hence an Oka manifold. Secondly, we present our main theorem, that there is a $6$-sheeted branched covering space of $R_3$ that is an Oka manifold. It follows that $R_3$ is $\mathbb{C}$-connected and dominable. We show that $R_3$ is Oka if and only if $\mathbb{P}_2\backslash C$ is Oka, where $C$ is a cubic curve that is the image of a certain embedding of $X$ into $\mathbb{P}_2$. We investigate the strong dominability of $R_3$ and show that if $X$ is not biholomorphic to $\mathbb{C}/Γ_0$, where $Γ_0$ is the hexagonal lattice, then $R_3$ is strongly dominable. As a Lie group, $X$ acts freely on $R_3$ by precomposition by translations. We show that $R_3$ is holomorphically convex and that the quotient space $R_3/X$ is a Stein manifold. We construct an alternative $6$-sheeted Oka branched covering space of $R_3$ and prove that it is isomorphic to our first construction in a natural way. This alternative construction gives us an easier way of interpreting the fibres of the branched covering map.

math.CV↗

arXiv2015

Multi-wavelength holography with a single Spatial Light Modulator for ultracold atom experiments

We demonstrate a method to create arbitrary intensity distributions of multiple wavelengths of light, which can be useful for ultracold atom experiments, by using regional phase-calculation algorithms to find a single hologram which is illuminated with overlapped laser beams. The regionality of the algorithms is used to program spatially distinct features in the calculated intensity distribution, which then overlap in the Fourier plane due to the dependence of diffraction angle on wavelength. This technique is easily integrated into cold atom experiments, requiring little optical access. We demonstrate the method and two possible experimental scenarios by generating light patterns with 670nm, 780nm and 1064nm laser light which are accurate to the level of a few percent.

cond-mat.quant-gas↗

arXiv2007

Explanation for Anomalous Shock Temperatures Measured by Neutron Resonance Spectroscopy

Neutron resonance spectrometry (NRS) has been used to measure the temperature inside Mo samples during shock loading. The temperatures obtained were significantly higher than predicted assuming ideal hydrodynamic loading. The effect of plastic flow and non-ideal projectile behavior were assessed. Plastic flow was calculated self-consistently with the shock jump conditions: this is necessary for a rigorous estimate of the locus of shock states accessible. Plastic flow was estimated to contribute a temperature rise of 53K compared with hydrodynamic flow. Simulations were performed of the operation of the explosively-driven projectile system used to induce the shock in the Mo sample. The simulations predicted that the projectile was significantly curved on impact, and still accelerating. The resulting spatial variations in load, including radial components of velocity, were predicted to increase the apparent temperature that would be deduced from the width of the neutron resonance by 160K. These corrections are sufficient to reconcile the apparent temperatures deduced using NRS with the accepted properties of Mo, in particular its equation of state.

cond-mat.other↗