Searcharxiv⌕ Search

arXiv subjects

Frank W. Bentrem

Publications and source records attributed to Frank W. Bentrem.

13 recordsLinked to original sources
arXiv2021

Leveraging Sharing Communities to Achieve Federated Learning for Cybersecurity

Automated cyber threat detection in computer networks is a major challenge in cybersecurity. The cyber domain has inherent challenges that make traditional machine learning techniques problematic, specifically the need to learn continually evolving attacks through global collaboration while maintaining data privacy, and the varying resources available to network owners. We present a scheme to mitigate these difficulties through an architectural approach using community model sharing with a streaming analytic pipeline. Our streaming approach trains models incrementally as each log record is processed, thereby adjusting to concept drift resulting from changing attacks. Further, we designed a community sharing approach which federates learning through merging models without the need to share sensitive cyber-log data. Finally, by standardizing data and Machine Learning processes in a modular way, we provide network security operators the ability to manage cyber threat events and model sensitivity through community member and analytic method weighting in ways that are best suited for their available resources and data.

cs.CR↗
arXiv2010

Discovering the Value of Multidisciplinary Approaches to Research: Insights from a Sabbatical

In this informal report, I outline my research efforts, collaborations, and other experiences while participating in the Naval Research Laboratory (NRL)'s Advanced Graduate Research Program (AGRP), aka Sabbatical, from October 2008 through September 2009. This report is in no way intended to present the technical details of the various research projects, but rather a broad overview of the small ways my efforts may have contributed to ongoing research. I wish to convey to the reader the value of multidisciplinary approaches to scientific research and how the AGRP facilitates these opportunities.

physics.pop-ph↗
arXiv2010

Critical Field Strength in an Electroclinic Liquid Crystal Elastomer

We elucidate the polymer dynamics of a liquid crystal elastomer based on the time-dependent response of the pendent liquid crystal mesogens. The molecular tilt and switching time of mesogens are analyzed as a function of temperature and cross-linking density upon application of an electric field. We observe an unexpected maximum in the switching time of the liquid crystal mesogens at intermediate field strength. Analysis of the molecular tilt over multiple time regimes correlates the maximum response time with a transition to entangled polymer dynamics at a critical field strength.

cond-mat.soft↗
arXiv2010

A Q-Ising model application for linear-time image segmentation

A computational method is presented which efficiently segments digital grayscale images by directly applying the Q-state Ising (or Potts) model. Since the Potts model was first proposed in 1952, physicists have studied lattice models to gain deep insights into magnetism and other disordered systems. For some time, researchers have realized that digital images may be modeled in much the same way as these physical systems (i.e., as a square lattice of numerical values). A major drawback in using Potts model methods for image segmentation is that, with conventional methods, it processes in exponential time. Advances have been made via certain approximations to reduce the segmentation process to power-law time. However, in many applications (such as for sonar imagery), real-time processing requires much greater efficiency. This article contains a description of an energy minimization technique that applies four Potts (Q-Ising) models directly to the image and processes in linear time. The result is analogous to partitioning the system into regions of four classes of magnetism. This direct Potts segmentation technique is demonstrated on photographic, medical, and acoustic images.

physics.data-an↗
arXiv2009

Almost rolling motion: An investigation of rolling grooved cylinders

We examine the dynamics of cylinders that are grooved to form N teeth for rolling motion down an inclined plane. The grooved cylinders are experimentally found to reach a terminal velocity. This result can be explained by the inclusion of inelastic processes which occur whenever a tooth hits the surface. The fraction of the angular velocity that is lost during an inelastic collision is phenomenologically found to be proportional to (2*sin^2*pi/N)-(alpha*sin^3*pi/N), and the method of least squares is used to find the constant alpha=0.98. The adjusted theoretical results for the time of rolling as well as for terminal velocity are found to be in good agreement with the experimental results.

physics.class-ph↗
arXiv2008

Sensing Shallow Seafloor and Sediment Properties, Recent History

Near surface seafloor properties are needed for recreational, commercial, and military applications. Construction projects on the ocean seafloors often require extensive knowledge about strength, deformability, hydraulic, thermal, acoustic, and seismic characteristics for locating stable environments and ensuring proper functioning of structures, pipelines, and other installations on the surface of and buried into the marine sediments. The military is also interested in a variety of seafloor properties as they impact sound propagation, mine impact burial, trafficability, bearing capacity, time-dependent settlement, and stability of objects on the seafloor. Point measurements of sediment properties are done using core samplers and sediment grab devices (with subsequent lab analysis) and in-situ probes. These techniques are expensive in terms of ship time and provide limited area coverage. Sub-bottom acoustic and electromagnetic sensors can provide profiles of near surface sediment information with improved coverage rates. Fusion techniques are being developed to provide areal extent of sediment information from multiple sensors. This paper examines the recent history of techniques used to measure sediment properties in the upper portions of the seafloor and in shallow (<100m) water.

physics.ao-ph↗
arXiv2006

Thermal Roughening and Deroughening at Polymer Interfaces in Electrophoretic Deposition

Thermal scaling and relaxation of the interface width in an electrophoretic deposition of polymer chains is examined by a three-dimensional Monte Carlo simulation on a discrete lattice. Variation of the equilibrium interface width $W_r$ with the temperature $T$ shows deroughening $W_r \propto T^{-δ}$, with $δ\sim 1/4$, at low temperatures and roughening $W_r \propto T^ν$, with $ν\sim 0.4$ at high temperatures. The roughening-deroughening transition temperature $T_t$ increases with longer chain lengths and is reduced by using the slower segmental dynamics.

cond-mat.soft↗
arXiv2006

High-Frequency Acoustic Sediment Classification in Shallow Water

A geoacoustic inversion technique for high-frequency (12 kHz) multibeam sonar data is presented as a means to classify the seafloor sediment in shallow water (40--300 m). The inversion makes use of backscattered data at a variety of grazing angles to estimate mean grain size. The need for sediment type and the large amounts of multibeam data being collected with the Naval Oceanographic Office's Simrad EM 121A systems, have fostered the development of algorithms to process the EM 121A acoustic backscatter into maps of sediment type. The APL-UW (Applied Physics Laboratory at the University of Washington) backscattering model is used with simulated annealing to invert for six geoacoustic parameters. For the inversion, three of the parameters are constrained according to empirical correlations with mean grain size, which is introduced as an unconstrained parameter. The four unconstrained (free) parameters are mean grain size, sediment volume interaction, and two seafloor roughness parameters. Acoustic sediment classification is performed in the Onslow Bay region off the coast of North Carolina using data from the 12kHz Simrad EM 121A multibeam sonar system. Raw hydrophone data is beamformed into 122 beams with a 120-degree swath on the ocean floor, and backscattering strengths are calculated for each beam and for each ping. Ground truth consists of 68 grab samples in the immediate vicinity of the sonar survey, which have been analyzed for mean grain size. Mean grain size from the inversion shows 90\% agreement with the ground truth and may be a useful tool for high-frequency acoustic sediment classification in shallow water.

physics.ao-ph↗
arXiv2006

Roughening, deroughening, and nonuniversal scaling of the interface width in electrophoretic deposition of polymer chains

Growth and roughness of the interface of deposited polymer chains driven by a field onto an impenetrable adsorbing surface are studied by computer simulations in (2+1) dimensions. The evolution of the interface width W shows a crossover from short-time growth described by the exponent beta1 to a long-time growth with exponent beta2 (>beta1). The saturated width increases, i.e., the interface roughens, with the molecular weight Lc, but the roughness exponent alpha (from Ws~L^alpha) becomes negative in contrast to models for particle deposition; alpha depends on the chain length--a nonuniversal scaling with the substrate length L. Roughening and deroughening occur as the field E and the temperature T compete such that Ws=(A+BT)E^-1/2.

cond-mat.soft↗
arXiv2006

Density and conformation with relaxed substrate, bulk, and interface electrophoretic deposition of polymer chains

Characteristics of relaxed density profile and conformation of polymer chains are studied by a Monte Carlo simulation on a discrete lattice in three dimensions using different segmental (kink-jump $K$, crank-shaft $C$, reptation $R$) dynamics. Three distinct density regimes, substrate, bulk, and interface, are identified. With the $KC$ segmental dynamics we find that the substrate coverage grows with a power-law, $d_s \propto t^γ$ with a field dependent nonuniversal exponent $γ= 0.23 + 0.7 E$. The bulk volume fraction $d_b$ and the substrate polymer density ($d_s$) increases exponentially with the field ($d_b \propto E^{0.4}$, $d_s \propto E^{0.2}$) in the low field regime. The interface polymer density $d_f$ increases with the molecular weight. With the $KCR$ segmental dynamics, bulk and substrate density decreases linearly with the temperature at high temperatures. The bulk volume fraction is found to decay with the molecular weight, $d_b \propto L_c^{-0.11}$. The radius of gyration remains Gaussian in all density regions.

cond-mat.soft↗
arXiv2006

Interface relaxation in electrophoretic deposition of polymer chains: Effects of segmental dynamics, molecular weight, and field

Using different segmental dynamics and relaxation, characteristics of the interface growth is examined in an electrophoretic deposition of polymer chains on a three (2+1) dimensional discrete lattice with a Monte Carlo simulation. Incorporation of faster modes such as crankshaft and reptation movements along with the relatively slow kink-jump dynamics seems crucial in relaxing the interface width. As the continuously released polymer chains are driven (via segmental movements) and deposited, the interface width $W$ grows with the number of time steps $t$, $W \propto t^β,$ ($β\sim 0.4$--$0.8)$, which is followed by its saturation to a steady-state value $W_s$. Stopping the release of additional chains after saturation while continuing the segmental movements relaxes the saturated width to an equilibrium value ($W_s \to W_r$). Scaling of the relaxed interface width $W_r$ with the driving field $E$, $W_r \propto E^{-1/2}$ remains similar to that of the steady-state $W_s$ width. In contrast to monotonic increase of the steady-state width $W_s$, the relaxed interface width $W_r$ is found to decay (possibly as a stretched exponential) with the molecular weight.

cond-mat.soft↗