arXiv subjects

David Campbell

Publications and source records attributed to David Campbell.

15 recordsLinked to original sources

arXiv2026

Universal Dynamical Response to Slow Driving in Chaotic Systems

We propose a unified perspective on classical and quantum chaos based on the stability of a system's stationary states under slow driving. We probe this sensitivity via the system's susceptibility to the average protocol speed, which we call the ``speed-Fisher information," and relate it to irreversible entropy production in the system. We show that chaotic dynamics manifests as a divergence of the speed-Fisher information with the protocol time, and that this response is controlled by the perturbation's low-frequency spectral weight. This approach to chaos applies to both classical and quantum Hamiltonian systems, and naturally extends to non-Hamiltonian classical flows. We illustrate this framework with simple classical and quantum examples, along with a non-Hamiltonian flow that qualitatively exhibits analogous low-frequency spectral behavior.

cond-mat.stat-mech↗

arXiv2026

Defensive Refusal Bias: How Safety Alignment Fails Cyber Defenders

Safety alignment in large language models (LLMs), particularly for cybersecurity tasks, primarily focuses on preventing misuse. While this approach reduces direct harm, it obscures a complementary failure mode: denial of assistance to legitimate defenders. We study Defensive Refusal Bias -- the tendency of safety-tuned frontier LLMs to refuse assistance for authorized defensive cybersecurity tasks when those tasks include similar language to an offensive cyber task. Based on 2,390 real-world examples from the National Collegiate Cyber Defense Competition (NCCDC), we find that LLMs refuse defensive requests containing security-sensitive keywords at $2.72\times$ the rate of semantically equivalent neutral requests ($p < 0.001$). The highest refusal rates occur in the most operationally critical tasks: system hardening (43.8%) and malware analysis (34.3%). Interestingly, explicit authorization, where the user directly instructs the model that they have authority to complete the target task, increases refusal rates, suggesting models interpret justifications as adversarial rather than exculpatory. These findings are urgent for interactive use and critical for autonomous defensive agents, which cannot rephrase refused queries or retry. Our findings suggest that current LLM cybersecurity alignment relies on semantic similarity to harmful content rather than reasoning about intent or authorization. We call for mitigations that analyze intent to maximize defensive capabilities while still preventing harmful compliance.

cs.CR↗

arXiv2025

Diffusion as a Signature of Chaos

While classical chaos is defined via a system's sensitive dependence on its initial conditions (SDIC), this notion does not directly extend to quantum systems. Instead, recent works have established defining both quantum and classical chaos via the sensitivity to adiabatic deformations and measuring this sensitivity using the adiabatic gauge potential (AGP). Building on this formalism, we introduce the ``observable drift" as a probe of chaos in generic, non-Hamiltonian, classical systems. We show that this probe correctly characterizes classical systems that exhibit SDIC as chaotic. Moreover, this characterization is consistent with the measure-theoretic definition of chaos via weak mixing. Thus, we show that these two notions of sensitivity (to changes in initial conditions and to adiabatic deformations) can be probed using the same quantity, and therefore, are equivalent definitions of chaos. Numerical examples are provided via the tent map, the logistic map, and the Chirikov standard map.

nlin.CD↗

arXiv2025

Ultra-high-gain water-window X-ray laser driven by plasma photocathode wakefield acceleration

X-ray free-electron lasers are large and complex machines, limited by electron beam brightness. Here we show through start-to-end simulations how to realise compact, robust and tunable X-ray lasers in the water window, based on ultra-bright electron beams from plasma wakefield accelerators. First, an ultra-low-emittance electron beam is released by a plasma photocathode in a metre-scale plasma wakefield accelerator. By tuning the beam charge, space-charge forces create a balance between beam fields and wakefields that reduces beam energy spread and improves energy stability - both critical for beam extraction, transport, and focusing into a metre-scale undulator. Here, the resulting ultra-bright beams produce wavelength-tunable, coherent, femtosecond scale photon pulses at ultra-high gain. This regime enables reliable generation of millijoule-gigawatt-class X-ray laser pulses across the water window, offering tunability via the witness beam charge and robustness against variations in plasma wakefield strength. Our findings help democratise access to coherent, high-power, soft X-ray radiation.

physics.plasm-ph↗

arXiv2025

Adiabatic Gauge Potential as a Tool for Detecting Chaos in Classical Systems

The interplay between chaos and thermalization in weakly non-integrable systems is a rich and complex subject. Interest in this area is further motivated by a desire to develop a unified picture of chaos for both quantum and classical systems. In this work, we study the adiabatic gauge potential (AGP), an object typically studied in quantum mechanics that describes deformations of a quantum state under adiabatic variation of the Hamiltonian, in classical Fermi-Pasta-Ulam-Tsingou (FPUT) and Toda models. We show how the time variance of the AGP over a trajectory probes the long-time correlations of a generic observable and can be used to distinguish among nearly integrable, weakly chaotic, and strongly chaotic regimes. We draw connections between the evolution of the AGP and diffusion and derive a fluctuation-dissipation relation that connects its variance to long-time correlations of the observable. Within this framework, we demonstrate that strongly and weakly chaotic regimes correspond to normal and anomalous diffusion, respectively. The latter gives rise to a marked increase in the variance as the time interval is increased, and this behavior serves as the basis for our probe of the onset times of chaos, which is interpreted as a ``mixing" time. Numerical results are presented for FPUT and Toda systems that highlight integrable, weakly chaotic, and strongly chaotic regimes. Further, a hierarchy of $t_{\text{Lyapunov}} < t_{\text{chaos}} < t_{\text{thermalization}}$ is found in these models. We conclude by commenting on the wide applicability of our method to a broader class of systems.

nlin.CD↗

arXiv2024

FRODO: A novel approach to micro-macro multilevel regression

Within the field of hierarchical modelling, little attention is paid to micro-macro models: those in which group-level outcomes are dependent on covariates measured at the level of individuals within groups. Although such models are perhaps underrepresented in the literature, they have applications in economics, epidemiology, and the social sciences. Despite the strong mathematical similarities between micro-macro and measurement error models, few efforts have been made to apply the much better-developed methodology of the latter to the former. Here, we present a new empirical Bayesian technique for micro-macro data, called FRODO (Functional Regression On Densities of Observations). The method jointly infers group-specific densities for multilevel covariates and uses them as functional predictors in a functional linear regression, resulting in a model that is analogous to a generalized additive model (GAM). In doing so, it achieves a level of generality comparable to more sophisticated methods developed for errors-in-variables models, while further leveraging the larger group sizes characteristic of multilevel data to provide richer information about the within-group covariate distributions. After explaining the hierarchical structure of FRODO, its power and versatility are demonstrated on several simulated datasets, showcasing its ability to accommodate a wide variety of covariate distributions and regression models.

stat.ME↗

arXiv2024

A probabilistic diagnostic for Laplace approximations: Introduction and experimentation

Many models require integrals of high-dimensional functions: for instance, to obtain marginal likelihoods. Such integrals may be intractable, or too expensive to compute numerically. Instead, we can use the Laplace approximation (LA). The LA is exact if the function is proportional to a normal density; its effectiveness therefore depends on the function's true shape. Here, we propose the use of the probabilistic numerical framework to develop a diagnostic for the LA and its underlying shape assumptions, modelling the function and its integral as a Gaussian process and devising a "test" by conditioning on a finite number of function values. The test is decidedly non-asymptotic and is not intended as a full substitute for numerical integration - rather, it is simply intended to test the feasibility of the assumptions underpinning the LA with as minimal computation. We discuss approaches to optimize and design the test, apply it to known sample functions, and highlight the challenges of high dimensions.

stat.ME↗

arXiv2024

Periodic Orbits in Fermi-Pasta-Ulam-Tsingou Systems

The FPUT paradox is the phenomenon whereby a one-dimensional chain of oscillators with nonlinear couplings shows non-ergodic behavior. The trajectory of the system in phase space, with a long wavelength initial condition, closely follows that of the Toda model over short times, as both systems seem to relax quickly to a non-thermal, metastable state. Over longer times, resonances in the FPUT spectrum drive the system towards equilibrium, away from the Toda trajectory. Similar resonances are observed in $q$-breather spectra, suggesting that $q$-breathers are involved in the route towards thermalization. In this article we investigate such resonances and show that they occur due to exact overlaps of $q$-breather frequencies of the type $m\Omega_1 = \Omega_k$. The resonances appear as peaks in the energy spectrum. Further, they give rise to new composite periodic orbits, which exist simultaneously with the original $q$-breathers. We find that such resonances are absent in integrable systems, as a consequence of the (infinite number of) conservation laws associated with integrability.

nlin.PS↗

arXiv2024

The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning

The White House Executive Order on Artificial Intelligence highlights the risks of large language models (LLMs) empowering malicious actors in developing biological, cyber, and chemical weapons. To measure these risks of malicious use, government institutions and major AI labs are developing evaluations for hazardous capabilities in LLMs. However, current evaluations are private, preventing further research into mitigating risk. Furthermore, they focus on only a few, highly specific pathways for malicious use. To fill these gaps, we publicly release the Weapons of Mass Destruction Proxy (WMDP) benchmark, a dataset of 3,668 multiple-choice questions that serve as a proxy measurement of hazardous knowledge in biosecurity, cybersecurity, and chemical security. WMDP was developed by a consortium of academics and technical consultants, and was stringently filtered to eliminate sensitive information prior to public release. WMDP serves two roles: first, as an evaluation for hazardous knowledge in LLMs, and second, as a benchmark for unlearning methods to remove such hazardous knowledge. To guide progress on unlearning, we develop RMU, a state-of-the-art unlearning method based on controlling model representations. RMU reduces model performance on WMDP while maintaining general capabilities in areas such as biology and computer science, suggesting that unlearning may be a concrete path towards reducing malicious use from LLMs. We release our benchmark and code publicly at https://wmdp.ai

cs.LG↗

arXiv2023

Experimental study of developing free-falling annular flow in a large-scale vertical pipe

Annular flow is the primary characteristic of unsteady wastewater flow, which initiates entrained air and sets up the air pressure regime within the system - an important design consideration. This paper reports on an experimental investigation of free-falling annular flow in a vertical pipe with different inlets at extended flow ranges up to Re = 3 x 10e4, similar to those in Building Drainage Systems (BDS). In the experimental setup, a vertical pipe system (5 m) was used to record velocity profiles and film thickness in the developing region through Particle Image Velocimetry (PIV) measurements. Entrained droplets were collected through a separator, and the entrainment fraction was calculated at different flow conditions. The study reports on the development process of the film velocity and thickness along the vertical pipe, which agrees well with empirical predictions. The results of the droplet entrainment of a vertical annular flow show the development process to the steady state. Additionally, a Tee-junction inlet in drainage system generates a higher and different entrainment profile.

physics.flu-dyn↗

arXiv2019

Efficient Follow-Up of Exoplanet Transits Using Small Telescopes

This paper is to introduce an online tool for the prediction of exoplanet transit light curves. Small telescopes can readily capture exoplanet transits under good weather conditions when the combination of a bright star and a large transiting exoplanet results in a significant depth of transit. However, in reality there are many considerations that need to be made in order to obtain useful measurements. This paper and accompanying website layout a procedure based on time series differential photometry that has been successfully employed using 0.4m aperture telescopes to predict the expected precision for a whole light curve. This enables robust planning to decide whether the observation of a particular exoplanet transit should be attempted and in particular to be able to readily see when it should not to be attempted. This may result in a significant increase in the number of transit observations captured by non-specialists. The technique and website are also appropriate for planning a variety of variable star observations where a prediction of the light curve can be made.

astro-ph.EP↗

arXiv2013

An ANOVA Test for Parameter Estimability using Data Cloning with Application to Statistical Inference for Dynamic Systems

Models for complex systems are often built with more parameters than can be uniquely identified by available data. Because of the variety of causes, identifying a lack of parameter identifiability typically requires mathematical manipulation of models, monte carlo simulations, and examination of the Fisher Information Matrix. A simple test for parameter estimability is introduced, using Data Cloning, a Markov Chain Monte Carlo based algorithm. Together, Data cloning and the ANOVA based test determine if the model parameters are estimable and if so, determine their maximum likelihood estimates and provide asymptotic standard errors. When not all model parameters are estimable, the Data Cloning results and the ANOVA test can be used to determine estimable parameter combinations or infer identifiability problems in the model structure. The method is illustrated using three different real data systems that are known to be difficult to analyze.

stat.ME↗

arXiv2013

Materials design from non-equilibrium steady states: driven graphene as a tunable semiconductor with topological properties

Controlling the properties of materials by driving them out of equilibrium is an exciting prospect that has only recently begun to be explored. In this paper we give a striking theoretical example of such materials design: a tunable gap in monolayer graphene is generated by exciting a particular optical phonon. We show that the system reaches a steady state whose transport properties are the same as if the system had a static electronic gap, controllable by the driving amplitude. Moreover, the steady state displays topological phenomena: there are chiral edge currents, which circulate a fractional charge e/2 per rotation cycle, with frequency set by the optical phonon frequency.

cond-mat.mes-hall↗

arXiv2009

Building on Quicksand

Reliable systems have always been built out of unreliable components. Early on, the reliable components were small such as mirrored disks or ECC (Error Correcting Codes) in core memory. These systems were designed such that failures of these small components were transparent to the application. Later, the size of the unreliable components grew larger and semantic challenges crept into the application when failures occurred. As the granularity of the unreliable component grows, the latency to communicate with a backup becomes unpalatable. This leads to a more relaxed model for fault tolerance. The primary system will acknowledge the work request and its actions without waiting to ensure that the backup is notified of the work. This improves the responsiveness of the system. There are two implications of asynchronous state capture: 1) Everything promised by the primary is probabilistic. There is always a chance that an untimely failure shortly after the promise results in a backup proceeding without knowledge of the commitment. Hence, nothing is guaranteed! 2) Applications must ensure eventual consistency. Since work may be stuck in the primary after a failure and reappear later, the processing order for work cannot be guaranteed. Platform designers are struggling to make this easier for their applications. Emerging patterns of eventual consistency and probabilistic execution may soon yield a way for applications to express requirements for a "looser" form of consistency while providing availability in the face of ever larger failures. This paper recounts portions of the evolution of these trends, attempts to show the patterns that span these changes, and talks about future directions as we continue to "build on quicksand".

cs.DC↗

arXiv2009

Comparing Bibliometric Statistics Obtained from the Web of Science and Scopus

For more than 40 years, the Institute for Scientific Information (ISI, now part of Thomson Reuters) produced the only available bibliographic databases from which bibliometricians could compile large-scale bibliometric indicators. ISI's citation indexes, now regrouped under the Web of Science (WoS), were the major sources of bibliometric data until 2004, when Scopus was launched by the publisher Reed Elsevier. For those who perform bibliometric analyses and comparisons of countries or institutions, the existence of these two major databases raises the important question of the comparability and stability of statistics obtained from different data sources. This paper uses macro-level bibliometric indicators to compare results obtained from the WoS and Scopus. It shows that the correlations between the measures obtained with both databases for the number of papers and the number of citations received by countries, as well as for their ranks, are extremely high (R2 > .99). There is also a very high correlation when countries' papers are broken down by field. The paper thus provides evidence that indicators of scientific production and citations at the country level are stable and largely independent of the database.

cs.IR↗