arXiv subjects

J. Armstrong

Publications and source records attributed to J. Armstrong.

5 recordsLinked to original sources

arXiv2026

Demo: BeGREEN Intelligence Plane for AI-driven Energy Efficient O-RAN management

Cellular networks management is being enhanced by O-RAN architecture and AI/ML solutions, enabling automated intelligent control loops for RAN optimization across various use cases. Ensuring energy sustainability is crucial to minimizing the impact of mobile networks on global energy consumption. This demo showcases the BeGREEN Intelligence Plane, an AI-driven solution for energy-efficient management of O-RAN networks. The presented workflow focuses on controlling the operational status of emulated cells, highlighting the integration of key components such as the AI Engine and the optimizations achieved through rApps and xApps

cs.NI↗

arXiv2021

Continual learning of longitudinal health records

Continual learning denotes machine learning methods which can adapt to new environments while retaining and reusing knowledge gained from past experiences. Such methods address two issues encountered by models in non-stationary environments: ungeneralisability to new data, and the catastrophic forgetting of previous knowledge when retrained. This is a pervasive problem in clinical settings where patient data exhibits covariate shift not only between populations, but also continuously over time. However, while continual learning methods have seen nascent success in the imaging domain, they have been little applied to the multi-variate sequential data characteristic of critical care patient recordings. Here we evaluate a variety of continual learning methods on longitudinal ICU data in a series of representative healthcare scenarios. We find that while several methods mitigate short-term forgetting, domain shift remains a challenging problem over large series of tasks, with only replay based methods achieving stable long-term performance. Code for reproducing all experiments can be found at https://github.com/iacobo/continual

cs.LG↗

arXiv2018

Physical properties of beta Lyr A and its opaque accretion disk

Mass exchange and mass loss in close binaries can significantly affect their evolution, but a complete self-consistent theory of these processes is still to be developed. Processes such as radiative shielding due to a~hot-spot region, or a~hydrodynamical interaction of different parts of the gas stream have been studied previously. In order to test the respective predictions, it is necessary to carry out detailed observations of binaries undergoing the largescale mass exchange, especially for those that are in the rapid transfer phase. \bla is an archetype of such a system, having a long and rich observational history. Our goal for this first study is to quantitatively estimate the geometry and physical properties of the optically thick components, namely the Roche-lobe filling mass-losing star, and the accretion disk surrounding the mass-gaining star of \blae. A series of continuum visible and NIR spectro-interferometric observations by the NPOI, CHARA/MIRC and VEGA instruments covering the whole orbit of \bla acquired during a~two-week campaign in 2013 were complemented with \ubvr\ photometric observations acquired during a three-year monitoring of the system. We included NUV and FUV observations from OAO~A-2, IUE, and Voyager satellites.

astro-ph.SR↗

arXiv2018

Study of radon reduction in gases for rare event search experiments

The noble elements, argon and xenon, are frequently employed as the target and event detector for weakly interacting particles such as neutrinos and Dark Matter. For such rare processes, background radiation must be carefully minimized. Radon provides one of the most significant contaminants since it is an inevitable product of trace amounts of natural uranium. To design a purification system for reducing such contamination, the adsorption characteristics of radon in nitrogen, argon, and xenon carrier gases on various types of charcoals with different adsorbing properties and intrinsic radioactive purities have been studied in the temperature range of 190-295 K at flow rates of 0.5 and 2 standard liters per minute. Essential performance parameters for the various charcoals include the average breakthrough times ($τ$), dynamic adsorption coefficients (k$_a$) and the number of theoretical stages (n). It is shown that the k$_a$-values for radon in nitrogen, argon, and xenon increase as the temperature of the charcoal traps decreases, and that they are significantly larger in nitrogen and argon than in xenon gas due to adsorption saturation effects. It is found that, unlike in xenon, the dynamic adsorption coefficients for radon in nitrogen and argon strictly obey the Arrhenius law. The experimental results strongly indicate that nitric acid etched Saratech is the best candidate among all used charcoal brands. It allows reducing total radon concentration in the LZ liquid Xe detector to meet the ultimate goal in the search for Dark Matter.

physics.ins-det↗

arXiv2016

The singularities of $A_g^P$

We complete the proof given earlier by one of us that the singularities of the perfect cone compactification $A_g^P$ of the coarse moduli space $A_g$ of principally polarized abelian $g$-folds are terminal if $g\ge 6$, and canonical if $g=5$.

math.AG↗