arXiv subjects

Alex Ho

Publications and source records attributed to Alex Ho.

8 recordsLinked to original sources

arXiv2026

Differentiable Cardiac Electrophysiology Simulations for Dynamical State and Parameter Estimation

The heart's contractions are triggered by action potential waves, which propagate through the cardiac muscle and exhibit diverse spatio-temporal dynamics during different heart rhythms. The dynamics are modeled with partial differential equations (PDEs) in cardiac electrophysiology simulations. However, fitting such models to measurement data to develop digital twins or patient-specific computer models is challenging. Here, we introduce differentiable cardiac electrophysiology simulations that can be fitted automatically to spatio-temporal measurement data of action potential waves in cardiac tissue. By comparing the simulated dynamics with the observation data, we define a loss function that is minimized via gradient-based optimization. Backpropagating the loss gradient through the differentiable PDE solver enables us to learn the parameters and recover the full dynamics, even with sparse, noisy, or partial observations. Implemented using both the finite-difference and smoothed particle hydrodynamics methods, our simulation framework can be applied to pixel-, voxel-, or point-based data, such as 2D or 3D slabs, or arbitrary shapes, such as the heart's ventricles. Using this methodology, we locate early activation sites inside a 3D bi-ventricular simulation geometry and fit a phenomenological model to imaging data of a voltage spiral wave in a cardiac monolayer cell culture. With experimental data, we employed a perceptual loss based on the Video Joint-Embedding Predictive Architecture, which enables fitting to noisy imaging data, and a generative diffusion model to estimate initial conditions and constrain solutions. Differentiable cardiac electrophysiology simulations could improve the diagnosis of rhythm abnormalities in patients and facilitate the development of personalized models or digital twins of the heart.

physics.med-ph↗

arXiv2026

Efficient Analysis of the Distilled Neural Tangent Kernel

Neural tangent kernel (NTK) methods are computationally limited by the need to evaluate large Jacobians across many data points. Existing approaches reduce this cost primarily through projecting and sketching the Jacobian. We show that NTK computation can also be reduced by compressing the data dimension itself using NTK-tuned dataset distillation. We demonstrate that the neural tangent space spanned by the input data can be induced by dataset distillation, yielding a 20-100$\times$ reduction in required Jacobian calculations. We further show that per-class NTK matrices have low effective rank that is preserved by this reduction. Building on these insights, we propose the distilled neural tangent kernel (DNTK), which combines NTK-tuned dataset distillation with state-of-the-art projection methods to reduce up NTK computational complexity by up to five orders of magnitude while preserving kernel structure and predictive performance.

cs.LG↗

arXiv2024

DisCo-DSO: Coupling Discrete and Continuous Optimization for Efficient Generative Design in Hybrid Spaces

We consider the challenge of black-box optimization within hybrid discrete-continuous and variable-length spaces, a problem that arises in various applications, such as decision tree learning and symbolic regression. We propose DisCo-DSO (Discrete-Continuous Deep Symbolic Optimization), a novel approach that uses a generative model to learn a joint distribution over discrete and continuous design variables to sample new hybrid designs. In contrast to standard decoupled approaches, in which the discrete and continuous variables are optimized separately, our joint optimization approach uses fewer objective function evaluations, is robust against non-differentiable objectives, and learns from prior samples to guide the search, leading to significant improvement in performance and sample efficiency. Our experiments on a diverse set of optimization tasks demonstrate that the advantages of DisCo-DSO become increasingly evident as the complexity of the problem increases. In particular, we illustrate DisCo-DSO's superiority over the state-of-the-art methods for interpretable reinforcement learning with decision trees.

cs.LG↗

arXiv2023

The accuracy of mutual potential approximations in simulations of binary asteroids

Simulations of asteroid binaries commonly use mutual gravitational potentials approximated by series expansions, leading to truncation errors, and also preventing correct computations of the forces and torques when the bodies are close. We make of a recently developed method where the mutual potential is calculated with the use of surface integrals and is exact for bodies of ellipsoidal shapes. The solutions produced by the surface integration method are compared with an approach that expands the mutual potential, truncated at second and fourth order. The approximate solutions are generated with the ``General Use Binary Asteroid Simulator'' (gubas). We find that the differences in the forces and torques are the largest when the bodies are nearly touching. These differences can exceed 1000% if the shape of the primary is highly elongated. Long term simulations show more than 100% difference in the dynamics if the bodies are initially close, while the differences are negligible if the bodies are initially far apart. For simulations with two triaxial ellipsoids, the computational efficiency of the surface integral method is comparable to fourth order approximations with gubas, and superior to potentials truncated to order eight or higher.

astro-ph.EP↗

arXiv2022

Dynamics of asteroid systems post rotational fission

Asteroid binaries found amongst the Near-Earth objects are believed to have formed from rotational fission. In this paper, we aim to study the dynamical evolution of asteroid systems the moment after fission. The initial condition is modelled as a contact binary, similar to that of Boldrin et al. (2016). Both bodies are modelled as ellipsoids, and the secondary is given an initial rotation angle about its body-fixed $y$-axis. Moreover, we consider six different cases, three where the density of the secondary varies, and three where we vary its shape. The simulations consider 45 different initial tilt angles of the secondary, each with 37 different mass ratios. We start the dynamical simulations at the moment the contact binary reaches a spin fission limit, and our model ensures that the closest distance between the surfaces of the two bodies is always kept at 1 cm. The forces, torques and gravitational potential between the two bodies are modelled using a newly developed surface integration scheme, giving exact results for two ellipsoids. We find that more than 80% of the simulations end with the two bodies impacting, and collisions between the bodies are more common when the density of the secondary is lower, or when it becomes more elongated. When comparing with data on asteroid pairs from Pravec et al. (2019) we find that variations in density and shape of the secondary can account for some of the spread seen in the rotation period for observed pairs. Furthermore, the secondary may also reach a spin limit for surface disruption, creating a ternary/multiple system. We find that secondary fission typically occurs within the first five hours after the contact binary separates, and is more common when the secondary is less dense or more elongated.

astro-ph.EP↗

arXiv2021

Learning Temporally Causal Latent Processes from General Temporal Data

Our goal is to recover time-delayed latent causal variables and identify their relations from measured temporal data. Estimating causally-related latent variables from observations is particularly challenging as the latent variables are not uniquely recoverable in the most general case. In this work, we consider both a nonparametric, nonstationary setting and a parametric setting for the latent processes and propose two provable conditions under which temporally causal latent processes can be identified from their nonlinear mixtures. We propose LEAP, a theoretically-grounded framework that extends Variational AutoEncoders (VAEs) by enforcing our conditions through proper constraints in causal process prior. Experimental results on various datasets demonstrate that temporally causal latent processes are reliably identified from observed variables under different dependency structures and that our approach considerably outperforms baselines that do not properly leverage history or nonstationarity information. This demonstrates that using temporal information to learn latent processes from their invertible nonlinear mixtures in an unsupervised manner, for which we believe our work is one of the first, seems promising even without sparsity or minimality assumptions.

stat.ML↗

arXiv2021

Extended two-body problem for rotating rigid bodies

A new technique that utilizes surface integrals to find the force, torque and potential energy between two non-spherical, rigid bodies is presented. The method is relatively fast, and allows us to solve the full rigid two-body problem for pairs of spheroids and ellipsoids with 12 degrees of freedom. We demonstrate the method with two dimensionless test scenarios, one where tumbling motion develops, and one where the motion of the bodies resemble spinning tops. We also test the method on the asteroid binary (66391) 1999 KW4, where both components are modelled either as spheroids or ellipsoids. The two different shape models have negligible effects on the eccentricity and semi-major axis, but have a larger impact on the angular velocity along the $z$-direction. In all cases, energy and total angular momentum is conserved, and the simulation accuracy is kept at the machine accuracy level.

astro-ph.EP↗

arXiv2018

Probing modified gravity in cosmic filaments

Multiple modifications of general relativity (GR) have been proposed in the literature in order to understand the nature of the accelerated expansion of the Universe. However, thus far all the predictions of GR have been confirmed with constantly increasing accuracy. In this work, we study the imprints of a particular class of models -- "screened" modified gravity theories -- on cosmic filaments. We have utilized the $N$-body code ISIS/RAMSES to simulate the symmetron model and the Hu-Sawicky $f(R)$ model, and we post-process the output with DisPerSE to identify the filaments of the cosmic web. We investigated how the global properties of the filaments -- such as their lengths, masses, and thicknesses -- as well as their radial density and speed profiles change under different gravity theories. We find that filaments are, on average, shorter and denser in modified gravity models compared to in $\Lambda$CDM. We also find that the speed profiles of the filaments are enhanced, consistent with theoretical expectations. Overall, our results suggest that cosmic filaments can be an effective complementary probe of screened modified gravity theories on Mpc scales.

astro-ph.CO↗