Breakthrough Listen: publications

Silent Speech Recognition with Wearable Magnetometers

(2025)

Authors:

Debadatta Dash, Evan Kittle, Isabel Gerrard, Richard Csaky, Gabriel Gonzalez, David Taylor, Juan Pablo Llinas, Dominic Labanowski, Nishita Deka, Richy Yun

Statistics of travelling ionospheric disturbances observed using the LOFAR radio telescope

Journal of Space Weather and Space Climate EDP Sciences 15 (2025) 6-6

Authors:

Ben Boyde, Alan G Wood, Gareth Dorrian, Francesco de Gasperin, Maaijke Mevius

Abstract:

A climatology of Travelling Ionospheric Disturbances (TIDs) observed using the LOw Frequency ARray (LOFAR) has been created based on 2723 h of astronomical observations. Radio telescopes such as LOFAR must contend with many causes of signal distortions, including the ionosphere. To produce accurate astronomical images, calibration solutions are derived to mitigate these distortions as much as possible. These calibration solutions provide extremely precise measurements of ionospheric variations across the LOFAR network, enabling TIDs to be detected which may be inaccessible to more traditional techniques. Waves are detected by LOFAR under all observing conditions, with no clear dependence on solar or geomagnetic activity. The vast majority of the observed waves travel in the opposite direction to the climatological thermospheric winds, suggesting that they are caused by upward propagating atmospheric gravity waves which are filtered by the wind. Waves of different periods display slightly different propagation directions, with waves of shorter periods consistent with the winds at lower altitudes within the thermosphere (180 km for 10–15 min periods compared to 220 km for 20–27 min periods). This suggests that either the shorter period waves are being detected at lower altitudes or that they are simply more sensitive to the winds at lower altitudes. This indicates that observations made using LOFAR may enable the investigation of vertical coupling within the neutral atmosphere. The shortest period waves in the dataset (<∼10) display distinct characteristics, suggesting they may be from a distinct population such as previously reported disturbances in the plasmasphere. The short period waves are compared to previous observations using other radio telescopes, showing that plasmaspheric disturbances likely account for some of the shortest period waves (<∼5 min) but there are still a large number of waves at these periods which are of uncertain origin.

Supernova bounds on new scalars from resonant and soft emission

JHEP 04 (2025) 013

Authors:

Edward Hardy, Anton Sokolov, Henry Stubbs

Abstract:

What We Don't C: Representations for scientific discovery beyond VAEs

Machine Learning and the Physical Sciences workshop at NeurIPS 2025

Authors:

Brian Rogers, Micah Bowles, Chris J. Lintott, Steve Croft

Abstract:

Accessing information in learned representations is critical for scientific discovery in high-dimensional domains. We introduce a novel method based on latent flow matching with classifier-free guidance that disentangles latent subspaces by explicitly separating information included in conditioning from information that remains in the residual representation. Across three experiments -- a synthetic 2D Gaussian toy problem, colored MNIST, and the Galaxy10 astronomy dataset -- we show that our method enables access to meaningful features of high dimensional data. Our results highlight a simple yet powerful mechanism for analyzing, controlling, and repurposing latent representations, providing a pathway toward using generative models for scientific exploration of what we don't capture, consider, or catalog.

Commensal Transient Searches with MeerKAT in Gamma-Ray Burst and Supernova Fields

(2024)

Authors:

SI Chastain, AJ van der Horst, A Horesh, A Rowlinson, A Andersson, R Diretse, M Vaccari, RP Fender, PA Woudt

51��