My Publications

Unlocking Telemetry Potential: Self-Supervised Learning for Continuous Clinical Electrocardiogram Monitoring

Applying machine learning to monitoring hospitalised patients has the potential to save lives. In this paper we take first steps toward that goal by demonstrating how deep learning can be successfully applied to the telemetry monitoring systems, in particular the electrocardiogram (ECG), which is ubiquitous in the modern hospital setting. By leveraging self-supervised learning we pretrain a series models across the large yet unlabeled telemetry dataset, and demonstrate improved model performance across many downstream tasks. We also provide qualitative demonstrations of how these models an effectively annotate ECG in real time based on scarce high quality labels, showing one path to AI-based hospital monitoring.

Tom Kite, Uzair Tahamid Siam, Brian Ayers, Nicholas Houstis, Aaron D Aguirre

Bridging the gap: spectral distortions meet gravitational waves

In this paper we provide a new method to constrain the energy density of gravitational waves at high redshifts using measurements of the µ-type spectral distortion of the CMB. We show that the µ-distortion is sensitive to gravitational waves with frequencies \(10^{-16} \lesssim f \lesssim 10^{-8}\) Hz, which is inaccessible to other probes. In particular, this range of frequencies bridges the gap between cosmological large-scale and astrophysical small-scale probes. We use this method to place new constraints on the energy density of gravitational waves at high redshifts.

Thomas Kite, Andrea Ravenni, Subodh P. Patil, Jens Chluba

Clarifying transfer function approximations for the large-scale gravitational wave background in ΛCDM

In this paper we explore the solutions of gravitational wave transfer functions in the primordial universe. In doing so we clarify some misconceptions in the literature, show the ranges of validity of various analytic approximations and explore new aspects of the numerical solutions, such as the universal expansion due to dark energy. We also provide a new, simple analytic approximation for the transfer function that is accurate to within ~5% for all frequencies and redshifts of interest.

Thomas Kite, Andrea Ravenni, Jens Chluba and Subodh P. Patil

Spectro-spatial evolution of the CMB I: discretisation of the thermalisation Green’s function

This paper is the first in a series of three papers on the spectro-spatial evolution of the CMB. In this paper we present a new method to solve the Kompaneets equation for the CMB photon distribution function. The method constructs a discrete set of characteristic photon spectra, which allows numerical solvers to reduce the size of the equation set by two or three orders of magnitude while also providing better physical insight to the system. This paper motivates the discretisation and shows the new method converges to expected results for the average spectrum.

Jens Chluba, Tom Kite and Andrea Ravenni

Spectro-spatial evolution of the CMB II: generalised Boltzmann hierarchy

This paper is the second in the spectro-spatial series. In this work we apply the discretised basis of spectral shapes to the perturbation theory required for deriving CMB power spectra. With this method we can, for the first time, properly describe the interplay between first order perturbations in photon frequency and spatial inhomogeneities.

Jens Chluba, Andrea Ravenni and Tom Kite

Spectro-spatial evolution of the CMB III: transfer functions, power spectra and Fisher forecasts

This paper is the final installment in the spectro-spatial series. Here we take the bottom line equations from the previous paper and numerically solve them. We distill the physical meaning and origin of each term in order to properly build up the physics in the solutions. We show for the first time distorted CMB power spectra and transfer functions following average energy injections at arbitrary times in the early universe. We also show that forecasts for future CMB imagers will be able to constrain average energy injection beyond the limits set by COBE/FIRAS.

Tom Kite, Andrea Ravenni and Jens Chluba

Disentangling the primordial nature of stochastic gravitational wave backgrounds with CMB spectral distortions

In this paper we demonstrate the unique ability of spectral distortion (SD) measurements to disentangle the primordial nature of a stochastic gravitational wave background (SGWB) from its possible astrophysical foreground. We show this not through the SD induced from SGWB, but rather the limits that SD can place on the primordial power spectrum itself, which in turn limits the formation of SWGB through a variety of mechanisms. We show the potential of SD measurements applied to some concrete scenarios, like secondary induced gravitational waves (SIGW) and blackhole formation.

Bryce Cyr, Tom Kite, Jens Chluba, J. Colin Hill, Donghui Jeong, Sandeep Kumar Acharya, Boris Bolliet and Subodh P. Patil

PhD thesis: Primordial Spectral Distortions Originating from Cosmological Perturbations

My doctoral thesis is written in journal format, and as such it covers the introductory material necessary to understand my papers across three introductory chapters before delving into the papers themselves. A final conclusion discusses the implications of the work and suggests future directions.

Moriond conference proceeding: Bridging the Gap: Spectral Distortions meet Gravitational Waves

This work is a short summary of my first paper, and formed my contribution to Moriond Cosmology 2022. It introduces spectral distortions themselves from a more introductory point of view.

My CV My Publications LinkedIn Github Instagram