arXiv subjects

Mark Cheung

Publications and source records attributed to Mark Cheung.

13 recordsLinked to original sources

arXiv2026

Interplanetary scintillation-informed heliospheric modelling for the MeerKAT Pulsar Timing Array 4.5 yr dataset

Heliospheric density variations impart delays on pulse times of arrivals from millisecond pulsars. Improper modelling of these variations may affect gravitational wave detection and characterisation by pulsar timing arrays (PTAs). Currently, PTAs typically employ a time-varying, spherically symmetric heliosphere model, which does not capture the full spatial and temporal complexity of the heliosphere. Instead, we investigate whether a three-dimensional, time-dependent model of the inner heliosphere from interplanetary scintillation (IPS) measurements - the IPS-UCSD model - can be employed to mitigate the solar wind in PTA analyses. We applied the IPS-UCSD model to the MeerKAT PTA 4.5-year dataset to assess whether it could correct for heliospheric density variations, and the impact on GW sensitivity compared to a spherically-symmetric model. We find that the model does not accurately correct for heliosphere-induced timing distortions, leading to bias in recovered GW parameters. Using simulations, we show that the spherically symmetric heliosphere model also fails to fully capture heliospheric density variations like those in the IPS-UCSD model. However, if interstellar dispersion measure (DM) variations are also modelled, then the heliospheric model errors are partially absorbed by DM variations, reducing contamination of the GW signal. Therefore we find that a time-varying spherically symmetric model is sufficient to mitigate the effect of heliospheric time delays on recovered GW results at typical PTA radio frequencies, provided other signal components are also modelled. We propose that the most precisely timed pulsars may be used to improve data-driven heliospheric density models in the future.

astro-ph.HE↗

arXiv2026

Improving Solar EUV Irradiance Modeling with Differential Emission Measure Informed Spectra

Solar extreme ultraviolet (EUV) irradiance predominantly originates from the chromosphere and corona, and variations in EUV irradiance are a significant driver of space weather at Earth, causing increased satellite drag, radio communication disruptions, and reduced GPS accuracy. Since solar EUV irradiance measurements are intermittent and inhomogeneous, models are used to fill spectral and temporal gaps, typically relying on proxy-based methods that utilize linear correlations between emission originating from similar heights or temperatures in the solar atmosphere. Proxy methods can struggle to capture dynamic irradiance variations of the solar corona during flares in part due to the relative infrequency of flares on which to derive correlations and the steep temporal temperature gradients experienced during flares. This paper presents a hybrid method for modeling solar EUV irradiance utilizing forward-modeled physics-informed solar differential emission measures (DEMs) to capture the optically-thin coronal emission coupled with traditional proxy-based correlation methods to capture the chromospheric and continuum emission. The model is trained and tested using inputs from the Solar Dynamics Observatory (SDO) Atmospheric Imaging Assembly (AIA) and EUV Variability Experiment (EVE) and compared against a proxy-only method. The results demonstrate that including information from DEMs into a EUV irradiance model improves accuracy for flaring conditions, particularly for high-temperature coronal lines, where the model error is reduced by up to a factor of 3 for M- and X-class flares compared to a proxy-only model trained on the same data.

astro-ph.SR↗

arXiv2025

An Analysis of a Coronal Mass Ejection Leading Edge by Means of Multi-Spacecraft-in-Beam Phase Scintillation

A Coronal Mass Ejection (CME) was detected crossing the radio signals transmitted by the Mars Express (MEX) and Tianwen-1 (TIW) spacecraft at a solar elongation of $4.4^{o}$. The impact of the CME was clearly identifiable in the spacecraft signal SNR, Doppler noise and phase residuals observed at the University of Tasmania's Very Long Baseline Interferometry (VLBI) antenna in Ceduna, South Australia. The residual phases observed from the spacecraft were highly correlated with each other during the transit of the CME across the radio ray-path despite the spacecraft signals having substantially different Doppler trends. We analyse the auto- and cross-correlations between the spacecraft phase residuals, finding time-lags ranging between 3.18-14.43 seconds depending on whether the imprinted fluctuations were stronger on the uplink or the downlink radio ray-paths. We also examine the temporal evolution of the phase fluctuations to probe the finer structure of the CME and demonstrate that there was a clear difference in the turbulence regime of the CME leading edge and the background solar wind conditions several hours prior to the CME radio occultation. Finally, autocorrelation of the MEX two-way radio Doppler noise data from Ceduna and closed-loop Doppler data from ESA's New Norcia ground station antenna were used to constrain the location of the CME impact along the radio ray-path \add{to a region 0.2 AU from the Sun, at a heliospheric longitude consistent with CME origin at the Sun. The results presented demonstrate the potential of the multi-spacecraft-in-beam technique for studying CME structures in great detail, and providing measurements that complement the capabilities of future solar monitoring instruments.

astro-ph.SR↗

arXiv2025

Reconstructing Fine-Grained Network Data using Autoencoder Architectures with Domain Knowledge Penalties

The ability to reconstruct fine-grained network session data, including individual packets, from coarse-grained feature vectors is crucial for improving network security models. However, the large-scale collection and storage of raw network traffic pose significant challenges, particularly for capturing rare cyberattack samples. These challenges hinder the ability to retain comprehensive datasets for model training and future threat detection. To address this, we propose a machine learning approach guided by formal methods to encode and reconstruct network data. Our method employs autoencoder models with domain-informed penalties to impute PCAP session headers from structured feature representations. Experimental results demonstrate that incorporating domain knowledge through constraint-based loss terms significantly improves reconstruction accuracy, particularly for categorical features with session-level encodings. By enabling efficient reconstruction of detailed network sessions, our approach facilitates data-efficient model training while preserving privacy and storage efficiency.

cs.LG↗

arXiv2025

Comparison of plasma dynamics in Coronal Holes and Quiet Sun using flux emergence simulations

This paper presents a comparison of plasma dynamics in Coronal Holes (CHs) and Quiet Sun (QS) through 2.5D MHD flux emergence simulations. The magnetic reconnection between the emerging and the pre-existing flux leads to the formation of cool, dense plasmoids with hot boundaries, and hot & cool jets with velocities $\approx50$ km s$^{-1}$. We perform spectral synthesis in spectral lines probing transition region and coronal temperatures. CHs show reduced intensities, excess upflows (downflows), and widths during the jetting (downflow) period when compared to QS. During the jetting and downflow periods, velocity and line width of the hot spectral lines in CHs show a strong positive correlation with the vertical magnetic field at z = 0, while the intensity of the cooler lines shows a weak correlation, which is not seen in QS. During the jetting period in CH, we find upflows in Si IV to be correlated (anti-correlated) with upflows (downflows) in other lines, and downflows in CH in Si IV to be correlated (anti-correlated) with upflows (downflows) in other lines when compared to QS. During downflow, we find no strong correlation between Si IV and other line velocities. The correlation during the jetting period occurs due to coincident, co-spatial origins of the hot and cool jet, while the lack of correlation during the downflow phase suggests a decoupling of hot and cool plasma. These results demonstrate that flux emergence and reconnection with pre-existing flux in the atmosphere support a unified scenario for solar wind formation and coronal heating.

astro-ph.SR↗

arXiv2023

Particle acceleration in solar flares with imaging-spectroscopy in soft X-rays

Particles are accelerated to very high, non-thermal energies during explosive energy-release phenomena in space, solar, and astrophysical plasma environments. In the case of solar flares, it has been established that magnetic reconnection plays an important role for releasing the magnetic energy, but it remains unclear if or how magnetic reconnection can further explain particle acceleration during flares. Here we argue that the key issue is the lack of understanding of the precise context of particle acceleration but it can be overcome, in the near future, by performing imaging-spectroscopy in soft X-rays (SXRs). Such observations should be complemented by observations in other wavelengths such as extreme-ultraviolets (EUVs), microwaves, hard X-rays (HXRs), and gamma-rays. Also, numerical simulations will be crucial for further narrowing down the particle acceleration mechanism in the context revealed by the observations. Of all these efforts, imaging-spectroscopy in SXRs, if successfully applied to large limb flares, will be a milestone in our challenge of understanding electron acceleration in solar flares and beyond, i.e. the Plasma Universe.

astro-ph.SR↗

arXiv2020

Edge Entropy as an Indicator of the Effectiveness of GNNs over CNNs for Node Classification

Graph neural networks (GNNs) extend convolutional neural networks (CNNs) to graph-based data. A question that arises is how much performance improvement does the underlying graph structure in the GNN provide over the CNN (that ignores this graph structure). To address this question, we introduce edge entropy and evaluate how good an indicator it is for possible performance improvement of GNNs over CNNs. Our results on node classification with synthetic and real datasets show that lower values of edge entropy predict larger expected performance gains of GNNs over CNNs, and, conversely, higher edge entropy leads to expected smaller improvement gains.

cs.LG↗

arXiv2020

Graph Signal Processing and Deep Learning: Convolution, Pooling, and Topology

Deep learning, particularly convolutional neural networks (CNNs), have yielded rapid, significant improvements in computer vision and related domains. But conventional deep learning architectures perform poorly when data have an underlying graph structure, as in social, biological, and many other domains. This paper explores 1)how graph signal processing (GSP) can be used to extend CNN components to graphs in order to improve model performance; and 2)how to design the graph CNN architecture based on the topology or structure of the data graph.

eess.SP↗

arXiv2020

Pooling in Graph Convolutional Neural Networks

Graph convolutional neural networks (GCNNs) are a powerful extension of deep learning techniques to graph-structured data problems. We empirically evaluate several pooling methods for GCNNs, and combinations of those graph pooling methods with three different architectures: GCN, TAGCN, and GraphSAGE. We confirm that graph pooling, especially DiffPool, improves classification accuracy on popular graph classification datasets and find that, on average, TAGCN achieves comparable or better accuracy than GCN and GraphSAGE, particularly for datasets with larger and sparser graph structures.

eess.SP↗

arXiv2019

Using U-Nets to Create High-Fidelity Virtual Observations of the Solar Corona

Understanding and monitoring the complex and dynamic processes of the Sun is important for a number of human activities on Earth and in space. For this reason, NASA's Solar Dynamics Observatory (SDO) has been continuously monitoring the multi-layered Sun's atmosphere in high-resolution since its launch in 2010, generating terabytes of observational data every day. The synergy between machine learning and this enormous amount of data has the potential, still largely unexploited, to advance our understanding of the Sun and extend the capabilities of heliophysics missions. In the present work, we show that deep learning applied to SDO data can be successfully used to create a high-fidelity virtual telescope that generates synthetic observations of the solar corona by image translation. Towards this end we developed a deep neural network, structured as an encoder-decoder with skip connections (U-Net), that reconstructs the Sun's image of one instrument channel given temporally aligned images in three other channels. The approach we present has the potential to reduce the telemetry needs of SDO, enhance the capabilities of missions that have less observing channels, and transform the concept development of future missions.

astro-ph.SR↗

arXiv2019

Auto-Calibration of Remote Sensing Solar Telescopes with Deep Learning

As a part of NASA's Heliophysics System Observatory (HSO) fleet of satellites,the Solar Dynamics Observatory (SDO) has continuously monitored the Sun since2010. Ultraviolet (UV) and Extreme UV (EUV) instruments in orbit, such asSDO's Atmospheric Imaging Assembly (AIA) instrument, suffer time-dependent degradation which reduces instrument sensitivity. Accurate calibration for (E)UV instruments currently depends on periodic sounding rockets, which are infrequent and not practical for heliophysics missions in deep space. In the present work, we develop a Convolutional Neural Network (CNN) that auto-calibrates SDO/AIA channels and corrects sensitivity degradation by exploiting spatial patterns in multi-wavelength observations to arrive at a self-calibration of (E)UV imaging instruments. Our results remove a major impediment to developing future HSOmissions of the same scientific caliber as SDO but in deep space, able to observe the Sun from more vantage points than just SDO's current geosynchronous orbit.This approach can be adopted to perform autocalibration of other imaging systems exhibiting similar forms of degradation

astro-ph.SR↗

arXiv2017

Order out of Randomness : Self-Organization Processes in Astrophysics

Self-organization is a property of dissipative nonlinear processes that are governed by an internal driver and a positive feedback mechanism, which creates regular geometric and/or temporal patterns and decreases the entropy, in contrast to random processes. Here we investigate for the first time a comprehensive number of 16 self-organization processes that operate in planetary physics, solar physics, stellar physics, galactic physics, and cosmology. Self-organizing systems create spontaneous {\sl order out of chaos}, during the evolution from an initially disordered system to an ordered stationary system, via quasi-periodic limit-cycle dynamics, harmonic mechanical resonances, or gyromagnetic resonances. The internal driver can be gravity, rotation, thermal pressure, or acceleration of nonthermal particles, while the positive feedback mechanism is often an instability, such as the magneto-rotational instability, the Rayleigh-B\'enard convection instability, turbulence, vortex attraction, magnetic reconnection, plasma condensation, or loss-cone instability. Physical models of astrophysical self-organization processes involve hydrodynamic, MHD, and N-body formulations of Lotka-Volterra equation systems.

astro-ph.SR↗

arXiv2016

Contrastive Structured Anomaly Detection for Gaussian Graphical Models

Gaussian graphical models (GGMs) are probabilistic tools of choice for analyzing conditional dependencies between variables in complex systems. Finding changepoints in the structural evolution of a GGM is therefore essential to detecting anomalies in the underlying system modeled by the GGM. In order to detect structural anomalies in a GGM, we consider the problem of estimating changes in the precision matrix of the corresponding Gaussian distribution. We take a two-step approach to solving this problem:- (i) estimating a background precision matrix using system observations from the past without any anomalies, and (ii) estimating a foreground precision matrix using a sliding temporal window during anomaly monitoring. Our primary contribution is in estimating the foreground precision using a novel contrastive inverse covariance estimation procedure. In order to accurately learn only the structural changes to the GGM, we maximize a penalized log-likelihood where the penalty is the $l_1$ norm of difference between the foreground precision being estimated and the already learned background precision. We modify the alternating direction method of multipliers (ADMM) algorithm for sparse inverse covariance estimation to perform contrastive estimation of the foreground precision matrix. Our results on simulated GGM data show significant improvement in precision and recall for detecting structural changes to the GGM, compared to a non-contrastive sliding window baseline.

stat.ML↗