Showing 1–50 of 297 results for author: Williams, H

Integrating Novelty and Surprise for Experience Prioritization and Exploration in Image-Based Reinforcement Learning

Authors: Hoda Yamani, Henry Williams, Bruce A. MacDonald

Abstract: Sample efficiency is a central challenge in reinforcement learning (RL), particularly in image-based domains where agents must learn from high-dimensional visual inputs. Traditional sampling often relies on random or suboptimal experience selection, leading to redundant updates and slow learning. Improving efficiency requires mechanisms that prioritize informative experiences while also encouragin… ▽ More Sample efficiency is a central challenge in reinforcement learning (RL), particularly in image-based domains where agents must learn from high-dimensional visual inputs. Traditional sampling often relies on random or suboptimal experience selection, leading to redundant updates and slow learning. Improving efficiency requires mechanisms that prioritize informative experiences while also encouraging effective exploration. Prioritized Experience Replay (PER) addresses part of this challenge by reusing high-value transitions, while intrinsic rewards promote the exploration of novel or uncertain states. However, their integration has not been extensively studied. This paper introduces Novelty and Surprise Prioritized Experience Replay (NSPER), which uses novelty to capture underrepresented states and surprise to expose gaps in the agent's understanding of the environment. We further extend this with NSPER+R, integrating these signals as intrinsic rewards to jointly improve replay quality and exploration. Experiments on DeepMind Control Suite tasks show that NSPER and NSPER+R improve training efficiency and convergence speed compared to existing methods in image-based RL. △ Less

Submitted 18 August, 2026; originally announced August 2026.

Comments: 9 pages, 4 figures. Published in International Journal of Computer and Systems Engineering, 2026. Code: https://github.com/UoA-CARES/NSPER

Journal ref: International Journal of Computer and Systems Engineering, Vol. 20, No. 4, pp. 439-447, 2026

Repetition as Reinforcement: Enhancing Sample Efficiency via Instant Episode Repetition in Reinforcement Learning

Authors: Hoda Yamani, Yuning Xing, Koen van Rijnsoever, Bruce A. MacDonald, Henry Williams

Abstract: Repetition is a fundamental mechanism in human learning, where revisiting successful experiences strengthens memory, consolidates skills, and improves future performance. Motivated by this biological principle, we introduce Instant Episode Repetition (IER), a simple and novel mechanism that improves sample efficiency by immediately repeating action sequences from successful episodes during environ… ▽ More Repetition is a fundamental mechanism in human learning, where revisiting successful experiences strengthens memory, consolidates skills, and improves future performance. Motivated by this biological principle, we introduce Instant Episode Repetition (IER), a simple and novel mechanism that improves sample efficiency by immediately repeating action sequences from successful episodes during environment interaction. Unlike conventional approaches such as Experience Replay and Self-Imitation Learning (SIL), which passively reuse past experience during training updates, IER directly influences the data collection process. Upon identifying a high-reward episode, the agent repeats its action sequence for a fixed number of subsequent episodes, reinforcing valuable behaviors through renewed interaction with the environment. We integrate IER into state-of-the-art SAC and TD3 algorithms and evaluate its effectiveness on continuous-control benchmarks, including MuJoCo, the DeepMind Control Suite, and a real-world dynamic object translation task with a robotic manipulator. Experimental results demonstrate that this simple mechanism improves learning performance over standard and self-imitation-based baselines. △ Less

Submitted 18 August, 2026; originally announced August 2026.

Comments: 23 pages, 12 figures. Accepted at RLC 2026; to appear in Reinforcement Learning Journal (RLJ) 2026. Code: https://github.com/UoA-CARES/instant-episode-repetition

Bits per Spike as a Betting Game: An Interpretable Unit for Held-Out Log-Likelihood in Neural Data Analysis

Authors: Alex H. Williams

Abstract: Held-out log-likelihood is the standard currency for comparing statistical models of neural spike trains, and is often reported as bits per spike relative to a homogeneous Poisson baseline. The units of this metric are difficult to reason about: it is rarely obvious whether an improvement of, say, $0.34$ bits per spike is a large effect or a negligible one. This note develops an interpretation of… ▽ More Held-out log-likelihood is the standard currency for comparing statistical models of neural spike trains, and is often reported as bits per spike relative to a homogeneous Poisson baseline. The units of this metric are difficult to reason about: it is rarely obvious whether an improvement of, say, $0.34$ bits per spike is a large effect or a negligible one. This note develops an interpretation of held-out log-likelihood borrowed from game-theoretic statistics. A fitted model $Q$ is treated as a player who bets on each upcoming observation at prices set by a baseline model $B$. Under the optimal (Kelly) betting strategy the player's contract function is exactly the likelihood ratio $q/b$, and the expected log-likelihood ratio $L$ is the exponential growth rate of the player's wealth. Because the wealth process is a nonnegative martingale under the null hypothesis that $B$ generated the data, Ville's inequality turns it into an anytime-valid test: the baseline may be rejected at level $α$ as soon as wealth exceeds $1/α$. This yields a simple summary statistic, the time to significance $τΔ= -Δ\log(α) / L$, which is the amount of held-out recording needed on average to reject the baseline at level $α$. Since $τ$ is a strictly decreasing function of $L$, it ranks models identically to bits per spike; it is not a new statistic but a more interpretable unit for an existing one, expressed in seconds of recording rather than in bits. We illustrate the construction on head-direction cells recorded in mouse anterior thalamus, where a generalized linear model reaches significance against a homogeneous Poisson baseline in roughly $120$ ms of held-out data for a strongly tuned cell and roughly $11$ s for a moderately tuned cell. △ Less

Submitted 30 July, 2026; originally announced July 2026.

Comments: 13 pages, 4 figures. Companion blog post: https://neurostatsblog.github.io/2026/05/27/model-comparison-by-betting/

VENUS: an ultra-faint galaxy hosting the metal-poor type II supernova at $z=5.13$ Witnessing the initial metal enrichment with extremely frequent core-collapse supernovae?

Authors: Yoshihisa Asada, Seiji Fujimoto, Joseph F. V. Allingham, David A. Coulter, Conor Larison, Matthew R. Siebert, Gabriel Brammer, Dan Coe, Pratika Dayal, Qinyue Fei, Lukas J. Furtak, Vasily Kokorev, Keiichi Maeda, Richard Pan, Johan Richard, Fengwu Sun, Abdurro'uf, Jacqueline Antwi-Danso, Franz E. Bauer, Marusa Bradac, Larry D. Bradley, Volker Bromm, John Chisholm, Christopher J. Conselice, Christa DeCoursey , et al. (28 additional authors not shown)

Abstract: We present the first characterization of the host galaxy of a recently discovered type IIP SN at $z=5.13$ (SN Eos). SN Eos and its host galaxy are gravitationally lensed and multiply imaged. The total magnification $μ\sim53$ enables spatially resolving the system, allowing us to localize the core-collapse supernova (CCSN) position and to characterize its local environment within an early galaxy. O… ▽ More We present the first characterization of the host galaxy of a recently discovered type IIP SN at $z=5.13$ (SN Eos). SN Eos and its host galaxy are gravitationally lensed and multiply imaged. The total magnification $μ\sim53$ enables spatially resolving the system, allowing us to localize the core-collapse supernova (CCSN) position and to characterize its local environment within an early galaxy. Our observation reveals that the host is an ultra-faint ($M_{\rm UV}=-14.4\pm0.3$ mag) Lyman-$α$ emitter with a very high equivalent width. The host galaxy also shows very weak [O iii]4959,5007 lines despite an H$α$ line detection ([O iii]5007/H$β<0.7$ with case B recombination). Assuming that the weak [O iii] is due to low gas-phase metallicity given the low-metallicity of SN Eos itself, SN Eos plausibly marks the formation and explosion of a metal-poor star in an extremely metal-poor environment ($<1\ \%\ Z_\odot$), facilitating the initial stages of the chemical enrichment of the host. Finding the CCSN in such an ultra-faint galaxy at $z=5.13$ also indicates that the SN rate could be considerably higher in high-$z$, metal-poor environments, potentially implying e.g., a $Z$-dependent IMF, $Z$-dependent massive star explodability, or runaway stellar collisions in dense star clusters. Without lensing, only SN Eos would be detectable and the host would be below the detection limit in any NIRCam surveys ever performed. The Eos host galaxy can thus be representative of the origin of {\it hostless} supernovae frequently found in JWST blank field surveys. △ Less

Submitted 15 July, 2026; originally announced July 2026.

Comments: Submitted to ApJ. Comments are welcome!

Observations and empirical functions for the ocean surface wave spectrum

Authors: Hannah Hata Williams, Michael E. Mueller, Luc Deike

Abstract: Accurate parameterizations of ocean wave spectra are necessary in a wide array of disciplines including coastal, ocean, and naval engineering as well as in the study of wave interactions and ocean-atmosphere momentum flux. Many such applications use spectrum parameterizations based on temporal data collected well over a half century ago. The development of spatial wave measurement techniques that… ▽ More Accurate parameterizations of ocean wave spectra are necessary in a wide array of disciplines including coastal, ocean, and naval engineering as well as in the study of wave interactions and ocean-atmosphere momentum flux. Many such applications use spectrum parameterizations based on temporal data collected well over a half century ago. The development of spatial wave measurement techniques that can accurately capture a larger range of scales allows us to revisit the question of how best to represent an ocean wave spectrum in a variety of ocean wave conditions. We discuss two commonly used wave spectrum parameterizations through a comparison to data collected in field campaigns studying fetch-limited, fully-developed, and mixed sea conditions. We discuss a spectrum parameterization for fully-developed seas that has a $k^{-2.5}$ (or $ω^{-4}$) dependence on the wavenumber (or angular frequency) in the tail as opposed to the $k^{-3}$ (or $ω^{-5}$) dependence seen in other frequently-used parameterizations. With knowledge of the peak wavenumber $k_p$ and significant wave height $H_s$, alongside the wind speed, fully-developed conditions can be well-represented. We then compare the impact of using different wave spectrum parameterizations through a Large Eddy Simulation (LES) study of Marine Atmospheric Boundary Layers (MABLs) over the sea surface and find that changing the parameterization used results in variations in the equivalent roughness akin to significant changes in wave conditions. △ Less

Submitted 26 June, 2026; originally announced June 2026.

Comments: 19 pages, 7 figures

Sea Surface Roughness Dependence on Ocean Wave Parameters through Large Eddy Simulation with Local Subfilter Wave Drag

Authors: Hannah Hata Williams, Aditya K. Aiyer, Luc Deike, Michael E. Mueller

Abstract: Characterizing the Marine Atmospheric Boundary Layer (MABL) requires understanding the coupling between ocean waves and the turbulent atmospheric boundary layer above them. This coupling controls momentum exchange between the atmosphere and the ocean; it is of practical importance in the global climate, flow of ocean currents, ocean engineering, and offshore wind energy. Computational study of the… ▽ More Characterizing the Marine Atmospheric Boundary Layer (MABL) requires understanding the coupling between ocean waves and the turbulent atmospheric boundary layer above them. This coupling controls momentum exchange between the atmosphere and the ocean; it is of practical importance in the global climate, flow of ocean currents, ocean engineering, and offshore wind energy. Computational study of the MABL is complex because it must resolve the coupled physics of waves and turbulence over a wide range of spatial and temporal scales. This study expands on approaches for representing dynamic, local waves in Large Eddy Simulations (LES) of the MABL by developing a subfilter wave drag model to be local and scale-invariant. It explores the effects of different wave parameters (significant wave height and peak frequency of the wave energy spectrum) on the resulting momentum flux beyond monotonic relationships between surface stress through friction velocity $u_\ast$ and wind velocity above the surface $U_{10}$. Results are compared to field data and in a discussion on how representation of the MABL and associated momentum flux need to account for both wind and wave effects. △ Less

Submitted 13 June, 2026; originally announced June 2026.

Comments: 25 pages, 9 figures

Speed-Weighted Flocking for Sailing Swarms under Dynamic Environmental Forcing

Authors: Pranav Kedia, Aaron Gan, Hannah J. Williams, Andreagiovanni Reina, Heiko Hamann

Abstract: Collective behavior models, such as aggregation and flocking, usually assume self-propelled robots that can directly execute their desired speed and direction of motion without fundamental constraints. However, autonomous sailing robots violate this assumption. Their motion is shaped by wind-dependent propulsion, restricted headings, and spatially varying wind conditions. In particular, maneuverab… ▽ More Collective behavior models, such as aggregation and flocking, usually assume self-propelled robots that can directly execute their desired speed and direction of motion without fundamental constraints. However, autonomous sailing robots violate this assumption. Their motion is shaped by wind-dependent propulsion, restricted headings, and spatially varying wind conditions. In particular, maneuverability is coupled to wind speed: in weak wind, sailboats may turn only slowly or not at all, whereas stronger wind enables faster turns. This introduces transient heterogeneity in speed and maneuverability across the flock. We focus on this fast-slow coordination problem in sailing robot flocks. To study this problem, we introduce SailSwarmSwIM, a reduced-order simulator for autonomous sailing robot swarms that captures wind-dependent speed and maneuverability, no-go zones, tacking behavior, and steady or gusty wind fields. To design our novel flocking technique, we start from the Couzin model and introduce a speed-weighted social interaction rule that accounts for each robot's transient motion constraints. A key result is that increasing the social influence of slower robots improves polarization and reduces close encounters. This effect arises from a balance between attraction to fast neighbors, which helps maintain movement, and cohesion around slow neighbors, which prevents the flock from fragmenting. Together, our simulator, SailSwarmSwIM, and the speed-weighted interaction rule provide a modeling framework for studying adaptive collective behavior in robotic fleets whose motion capabilities are continuously shaped by wind. △ Less

Submitted 21 July, 2026; v1 submitted 19 May, 2026; originally announced May 2026.

Comments: Submitted at 18th International Conference on the Simulation of Adaptive Behavior (SAB 2026)

How Data Augmentation Shapes Neural Representations

Authors: Tianxiao He, Alex H. Williams, Sarah E. Harvey

Abstract: Data augmentation is widely recognized for improving generalization in deep networks, yet its impact on the geometry of learned representations remains poorly understood. In this work, we characterize how different data augmentation strategies reshape internal representations in neural networks. Using tools from shape analysis, we embed network hidden representations into a metric space where dist… ▽ More Data augmentation is widely recognized for improving generalization in deep networks, yet its impact on the geometry of learned representations remains poorly understood. In this work, we characterize how different data augmentation strategies reshape internal representations in neural networks. Using tools from shape analysis, we embed network hidden representations into a metric space where distance is invariant to scaling, translation, rotation and reflection. We show that increasing augmentation strength leads to well-behaved trajectories in this space, and that different augmentation types steer representations in distinct directions. Moreover, we investigate how neural representation shapes are distorted along data augmentation trajectories, and show that insights from neural geometry can predict which representations provide the most improvement when ensembling models. Our results reveal shared geometric patterns across architectures and seeds, and suggest that analyzing shape-space trajectories offers a principled tool for understanding and comparing data augmentation methods. △ Less

Submitted 14 May, 2026; originally announced May 2026.

Prediction of Rectal Cancer Regrowth from Longitudinal Endoscopy

Authors: Jorge Tapias Gomez, Despoina Kanata, Aneesh Rangnekar, Christina Lee, Hannah Williams, Hannah Thompson, J. Joshua Smith, Francisco Sanchez-Vega, Mert R. Sabuncu, Julio Garcia-Aguilar, Harini Veeraraghavan

Abstract: Clinical trial studies indicate benefit of watch-and-wait (WW) surveillance for patients with rectal cancer showing a complete or near clinical response (CR) directly after treatment (restaging). However, there are no objectively accurate methods to early detect local tumor regrowth (LR) in patients undergoing WW from follow-up exams. Hence, we developed Temporal Rectal Endoscopy Cross-attention (… ▽ More Clinical trial studies indicate benefit of watch-and-wait (WW) surveillance for patients with rectal cancer showing a complete or near clinical response (CR) directly after treatment (restaging). However, there are no objectively accurate methods to early detect local tumor regrowth (LR) in patients undergoing WW from follow-up exams. Hence, we developed Temporal Rectal Endoscopy Cross-attention (TREX), a longitudinal deep learning approach that combines pairs of images acquired at restaging and follow-up to distinguish CR from LR. TREX uses pretrained Swin Transformers in a siamese setting to extract features from longitudinal images and dual cross-attention to combine the features without spatial co-registration between image pairs. TREX and Swin-based baselines were trained under two settings: (a) detecting LR or CR at the last available follow-up and (b) early detection of LR at 3--6, 6--12, and 12--24 months before clinical confirmation. TREX achieved the highest accuracy in detecting LR with a high sensitivity of 97% $\pm$ 6% and a balanced accuracy of 90% $\pm$ 3%, and outperformed all baselines in early detection at both 3--6 (74% $\pm$ 1%) and 6--12 months (62% $\pm$ 4%) prior to clinical detection. Clinical validation via a surgeon survey showed that TREX matched attending-level overall accuracy (TREX: 86.21% vs.\ Clinicians: 87.84% $\pm$ 1.28%). Finally, we explored TREX's ability to predict treatment response by combining pre-treatment (pre-TNT) and restaging endoscopies, achieving a balanced accuracy of 73% $\pm$ 12%. These results show that longitudinal deep learning analysis of endoscopy may improve surveillance and enable earlier identification of rectal cancer regrowth. △ Less

Submitted 12 May, 2026; originally announced May 2026.

Comments: 14 Pages, 9 figures, 2 tables

Gravity-induced entanglement under constrained dynamics

Authors: Hollis Williams

Abstract: Tests of gravity-induced entanglement have been proposed as a route to probing the quantum nature of gravity, but existing schemes rely on free-fall interferometry of massive spatial superpositions, imposing severe experimental constraints. We show that systems exhibiting effectively inertial dynamics in the short-time regime reproduce the same gravitational phase accumulation responsible for enta… ▽ More Tests of gravity-induced entanglement have been proposed as a route to probing the quantum nature of gravity, but existing schemes rely on free-fall interferometry of massive spatial superpositions, imposing severe experimental constraints. We show that systems exhibiting effectively inertial dynamics in the short-time regime reproduce the same gravitational phase accumulation responsible for entanglement generation. Deviations from the free-fall phase enter at order $(τ/T)^2$, where $τ$ is the interaction timescale and $T$ is the characteristic period of the constrained motion. We analyse a representative mechanically constrained implementation using carbon nanotube pendula and show that the resulting correction to the entangling phase remains small in experimentally relevant regimes, leading to a negligible modification of the interference visibility used to certify entanglement. These results demonstrate that gravity-induced entanglement protocols extend beyond free-fall implementations to a broader class of constrained dynamical systems, complementing existing proposals for experimental realisations of the Bose-Marletto-Vedral protocol. △ Less

Submitted 16 August, 2026; v1 submitted 1 May, 2026; originally announced May 2026.

First Statistical Study of Over 100 Magnified Stellar Events at Redshift $z \approx 0.725$ with JWST

Authors: J. M. Palencia, Fengwu Sun, J. M. Diego, Yoshinobu Fudamoto, Anton M. Koekemoer, Christopher N. A. Willmer, Eduardo Iani, Xiaojing Lin, Justin D. R. Pierel, Alfred Amruth, Tom Broadhurst, W. Chen, Liang Dai, Daniel Espada, Alexei V. Filippenko, Seiji Fujimoto, Patrick L. Kelly, Mingyu Li, Sung Kei Li, Ashish Kumar Meena, Jordi Miralda-Escudé, P. Morilla, Mitchell F. Struble, Hayley Williams, Rogier A. Windhorst , et al. (3 additional authors not shown)

Abstract: Highly magnified stars at cosmological distances ($z \gtrsim 0.7$) become detectable thanks to microlensing by intracluster stars near the critical curves of galaxy clusters. Multi-epoch photometric campaigns targeting caustic crossing galaxies magnified by massive galaxy clusters enable the detection of these objects as transient events. Such stars provide unique opportunities to study stellar po… ▽ More Highly magnified stars at cosmological distances ($z \gtrsim 0.7$) become detectable thanks to microlensing by intracluster stars near the critical curves of galaxy clusters. Multi-epoch photometric campaigns targeting caustic crossing galaxies magnified by massive galaxy clusters enable the detection of these objects as transient events. Such stars provide unique opportunities to study stellar populations at early cosmic times, probe the nature of dark matter, reveal small-scale structure in the cluster, and improve lens models. To date, only a few dozen high-redshift stars have been reported, with a single lensed galaxy, the Dragon, holding the current record of 44 detections. These numbers, however, remain insufficient to exploit their full potential. In this paper, owing to the inclusion of new observations, we report the identification of more than 100 magnified stellar events in the Dragon, behind the massive galaxy cluster Abell 370. The relatively low redshift of the Dragon ($z\approx0.725$) facilitates the detection of its most massive stars. Using imaging data from three different cycles (2022--2024) with the James Webb Space Telescope, we apply a time-domain technique to identify flux variations associated with caustic-crossing events. From the spatial distribution of stellar events we constrain the high-end slope of the stellar luminosity function, finding $β=2.18^{+0.20}_{-0.30}$. Alternatively, assuming a fixed slope, we constrain the microlens surface mass density. In addition, we examine the parity asymmetry of the detected caustic-crossing events, a proposed probe of wave dark matter, and find that it remains present. We also use the events to trace the regions of highest magnification, offering an alternative way to map the system critical curves. △ Less

Submitted 4 May, 2026; v1 submitted 24 April, 2026; originally announced April 2026.

Comments: 18 pages, 9 figures

Near-deterministic loading of optical tweezer arrays via repulsive barricade potentials

Authors: Archie C. Baldock, Alex J. Matthies, Luke Caldwell, Hannah J. Williams

Abstract: Optical tweezers are a powerful tool for creating defect-free arrays of atoms and molecules, enabling advances in quantum simulation, computation, and precision metrology. However, the achievable array size is limited by the initial loading fraction, typically $50\,\%$ for atoms and $35\,\%$ for molecules. Here, we propose a general scheme for enabling multiple loading cycles by protecting trapped… ▽ More Optical tweezers are a powerful tool for creating defect-free arrays of atoms and molecules, enabling advances in quantum simulation, computation, and precision metrology. However, the achievable array size is limited by the initial loading fraction, typically $50\,\%$ for atoms and $35\,\%$ for molecules. Here, we propose a general scheme for enabling multiple loading cycles by protecting trapped particles using a repulsive barrier. We show that collision-limited lifetimes of particles in protected tweezers can exceed one second, leading to filling fractions of over $80\%$ after four loading cycles. Combined with existing rearrangement techniques, this approach enables efficient unity filling of tweezer arrays and provides a scalable pathway towards larger quantum technology platforms. △ Less

Submitted 27 May, 2026; v1 submitted 24 April, 2026; originally announced April 2026.

Comments: 7 pages, 3 figures

Modulating Cross-Modal Convergence with Single-Stimulus, Intra-Modal Dispersion

Authors: Eghbal A. Hosseini, Brian Cheung, Evelina Fedorenko, Alex H. Williams

Abstract: Neural networks exhibit a remarkable degree of representational convergence across diverse architectures, training objectives, and even data modalities. This convergence is predictive of alignment with brain representation. A recent hypothesis suggests this arises from learning the underlying structure in the environment in similar ways. However, it is unclear how individual stimuli elicit converg… ▽ More Neural networks exhibit a remarkable degree of representational convergence across diverse architectures, training objectives, and even data modalities. This convergence is predictive of alignment with brain representation. A recent hypothesis suggests this arises from learning the underlying structure in the environment in similar ways. However, it is unclear how individual stimuli elicit convergent representations across networks. An image can be perceived in multiple ways and expressed differently using words. Here, we introduce a methodology based on the Generalized Procrustes Algorithm to measure intra-modal representational convergence at the single-stimulus level. We applied this to vision models with distinct training objectives, selecting stimuli based on their degree of alignment (intra-modal dispersion). Crucially, we found that this intra-modal dispersion strongly modulates alignment between vision and language models (cross-modal convergence). Specifically, stimuli with low intra-modal dispersion (high agreement among vision models) elicited significantly higher cross-modal alignment than those with high dispersion, by up to a factor of two (e.g., in pairings of DINOv2 with language models). This effect was robust to stimulus selection criteria and generalized across different pairings of vision and language models. Measuring convergence at the single-stimulus level provides a path toward understanding the sources of convergence and divergence across modalities, and between neural networks and human neural representations. △ Less

Submitted 23 April, 2026; originally announced April 2026.

Journal ref: ICLR 2026 Workshop on Representational Alignment (Re-Align)

SN 2022riv in RX J2129: Discovery, Spectroscopic Classification, and Microlensing of a Strongly Lensed Type Ia Supernova from JWST and HST Observations

Authors: Birendra Dhanasingham, Patrick L. Kelly, Wenlei Chen, Justin Pierel, Masamune Oguri, Derek Perera, Jose M. Diego, Adi Zitrin, Ashish K. Meena, Mathilde Jauzac, Guillaume Mahler, Elias Mamuzic, Liliya L. R. Williams, Yoon Chan Taak, Anton M. Koekemoer, Thomas J. Broadhurst, Lukas J. Furtak, David Lagattuta, Hayley Williams, Kyle Dalrymple, Alexei V. Filippenko, Christa Gall, Daniel Gilman, Jens Hjorth, Saurabh W. Jha , et al. (5 additional authors not shown)

Abstract: The multiply imaged SN 2022riv was discovered through a search of galaxy cluster fields as part of a Hubble Space Telescope (HST) SNAP program to find highly magnified stars. The supernova (SN) was detected in the image corresponding to the longest time delay of a galaxy at redshift $z=1.522$ strongly lensed by the foreground galaxy cluster RX J2129.7+0005. Follow up James Webb Space Telescope (JW… ▽ More The multiply imaged SN 2022riv was discovered through a search of galaxy cluster fields as part of a Hubble Space Telescope (HST) SNAP program to find highly magnified stars. The supernova (SN) was detected in the image corresponding to the longest time delay of a galaxy at redshift $z=1.522$ strongly lensed by the foreground galaxy cluster RX J2129.7+0005. Follow up James Webb Space Telescope (JWST) NIRSpec G140M and PRISM spectroscopy yields a Type Ia SN classification. Using the SALT3-NIR light-curve fitter, we obtain a cosmology-independent measurement of the magnification of $5.35\pm1.01$ for the last-to-arrive image of the SN, with multiple SALT SN spectral time-series models yielding consistent constraints. The last-to-arrive image of SN 2022riv we detect appeared adjacent to the brightest cluster galaxy (BCG) at a location with an exceptionally high stellar mass density ($\sim 1-2$ dex higher than that of SN Refsdal), where microlensing is expected to introduce a 20-50% modulation of the magnification. Analyzing six independent lens models of the cluster, we find that four predict the magnification with much greater precision ($p < 0.05$) than would be expected by random chance, given the large effect anticipated from microlensing. Five models yield magnifications of roughly $4-7$ (within $1σ$) prior to accounting for microlensing, whereas HoliGRALE favors a significantly higher value of $15.39 \pm 0.85$. After incorporating nominal microlensing, the HoliGRALE prediction is within $1σ$ tension with our measurement. A companion paper (Dalrymple et al.) will present constraints on the relative time delay of the image that arrived earlier. △ Less

Submitted 20 August, 2026; v1 submitted 13 April, 2026; originally announced April 2026.

Comments: 27 pages, 16 figures + appendices. Consistent with the published version in APJ. Lens models are available for download from Zenodo at https://zenodo.org/records/21460864

JWST's PEARLS: A clumpy ring galaxy at $z = 4.0148$

Authors: David Vizgan, Ming-Yang Zhuang, Ian Smail, Rogier Windhorst, Gibson Bowling, Cheng Cheng, Seth Cohen, Christopher Conselice, Jose Diego, Brenda Frye, Norman Grogin, Rolf Jansen, Patrick Kamieneski, Anton Koekemoer, Rafael Ortiz III, Massimo Ricotti, Bangzheng Sun, Hayley Williams, S. P. Willner, Haojing Yan, Aadya Agrawal, Manuel Solimano, Zachary Stone, Joaquin Vieira, Chentao Yang

Abstract: Ring galaxies are an uncommon class of galaxies whose morphology is closely related to dynamical processes that govern galaxy evolution. Some ring galaxies, known as "collisional ring galaxies", are thought to form as a consequence of head-on collisions between galaxies, and a number of high-redshift collisional ring galaxies have been discovered and/or studied in the era of the James Webb Space T… ▽ More Ring galaxies are an uncommon class of galaxies whose morphology is closely related to dynamical processes that govern galaxy evolution. Some ring galaxies, known as "collisional ring galaxies", are thought to form as a consequence of head-on collisions between galaxies, and a number of high-redshift collisional ring galaxies have been discovered and/or studied in the era of the James Webb Space Telescope (JWST). In this paper, we present HST/ACS, JWST/NIRCam, and JWST/NIRSpec observations of a candidate ring galaxy at $z_{\rm spec} = 4.0148$, previously identified as a potential gravitational lens. The galaxy exhibits a complex morphology, including three bright clumps along an apparent ring with radius $\approx 0.25$" $\simeq 1.8$ kpc. It has a total SFR $= 140^{+20}_{-30}$ ${\rm M}_{\rm \odot}$ yr$^{-1}$ and $\log(M_\ast/{\rm M}_\odot) = 10.41^{+0.11}_{-0.13}$, making it similar to other high-redshift collisional ring galaxies. Although we argue strongly in favor of the collisional ring explanation, we cannot entirely rule out a galaxy-galaxy strong lensing explanation for the system's morphology, in which a foreground galaxy at $z \simeq 1.7$ lenses a galaxy at $z \simeq 4.0$ into an Einstein ring-like configuration; to confirm the nature of this source, we require kinematic information via high spectral resolution observations. We suggest that current and future gravitational lens surveys should consider high-redshift ring galaxies as possible but significant contaminants. △ Less

Submitted 12 May, 2026; v1 submitted 12 March, 2026; originally announced March 2026.

Comments: 14 pages, 7 figures, 2 tables. Accepted to ApJ after minor revision

Partial Soft-Matching Distance for Neural Representational Comparison with Partial Unit Correspondence

Authors: Chaitanya Kapoor, Alex H. Williams, Meenakshi Khosla

Abstract: Representational similarity metrics typically force all units to be matched, making them susceptible to noise and outliers common in neural representations. We extend the soft-matching distance to a partial optimal transport setting that allows some neurons to remain unmatched, yielding rotation-sensitive but robust correspondences. This partial soft-matching distance provides theoretical advantag… ▽ More Representational similarity metrics typically force all units to be matched, making them susceptible to noise and outliers common in neural representations. We extend the soft-matching distance to a partial optimal transport setting that allows some neurons to remain unmatched, yielding rotation-sensitive but robust correspondences. This partial soft-matching distance provides theoretical advantages -- relaxing strict mass conservation while maintaining interpretable transport costs -- and practical benefits through efficient neuron ranking in terms of cross-network alignment without costly iterative recomputation. In simulations, it preserves correct matches under outliers and reliably selects the correct model in noise-corrupted identification tasks. On fMRI data, it automatically excludes low-reliability voxels and produces voxel rankings by alignment quality that closely match computationally expensive brute-force approaches. It achieves higher alignment precision across homologous brain areas than standard soft-matching, which is forced to match all units regardless of quality. In deep networks, highly matched units exhibit similar maximally exciting images, while unmatched units show divergent patterns. This ability to partition by match quality enables focused analyses, e.g., testing whether networks have privileged axes even within their most aligned subpopulations. Overall, partial soft-matching provides a principled and practical method for representational comparison under partial correspondence. △ Less

Submitted 22 February, 2026; originally announced February 2026.

Entanglement Before Spacetime in Quantum-Gravity-Induced Interactions

Authors: Hollis Williams

Abstract: Quantum-gravity-induced entanglement of massive systems (QGEM) is commonly approximated in the nonrelativistic static limit by a Newtonian interaction between spatially separated masses. In this work, we reformulate the gravitationally mediated interaction phase in a conformally invariant twistor framework in which no notion of spacetime distance is assumed. We show that the bilocal phase responsi… ▽ More Quantum-gravity-induced entanglement of massive systems (QGEM) is commonly approximated in the nonrelativistic static limit by a Newtonian interaction between spatially separated masses. In this work, we reformulate the gravitationally mediated interaction phase in a conformally invariant twistor framework in which no notion of spacetime distance is assumed. We show that the bilocal phase responsible for entanglement generation remains well-defined and non-factorizable even in the absence of spacetime geometry. The familiar Newtonian $1/r$ phase, relevant for QGEM protocols, arises only after the conformal invariance is broken by introducing the infinity twistor, which selects a particular spacetime representation of the underlying bilocal quantum interaction. Our results isolate the genuinely quantum content of QGEM protocols and clarify the contingent role played by spacetime geometry in mediating entanglement. △ Less

Submitted 5 February, 2026; originally announced February 2026.

Automated Marine Biofouling Assessment: Benchmarking Computer Vision and Multimodal LLMs on the Level of Fouling Scale

Authors: Brayden Hamilton, Tim Cashmore, Peter Driscoll, Trevor Gee, Henry Williams

Abstract: Marine biofouling on vessel hulls poses major ecological, economic, and biosecurity risks. Traditional survey methods rely on diver inspections, which are hazardous and limited in scalability. This work investigates automated classification of biofouling severity on the Level of Fouling (LoF) scale using both custom computer vision models and large multimodal language models (LLMs). Convolutional… ▽ More Marine biofouling on vessel hulls poses major ecological, economic, and biosecurity risks. Traditional survey methods rely on diver inspections, which are hazardous and limited in scalability. This work investigates automated classification of biofouling severity on the Level of Fouling (LoF) scale using both custom computer vision models and large multimodal language models (LLMs). Convolutional neural networks, transformer-based segmentation, and zero-shot LLMs were evaluated on an expert-labelled dataset from the New Zealand Ministry for Primary Industries. Computer vision models showed high accuracy at extreme LoF categories but struggled with intermediate levels due to dataset imbalance and image framing. LLMs, guided by structured prompts and retrieval, achieved competitive performance without training and provided interpretable outputs. The results demonstrate complementary strengths across approaches and suggest that hybrid methods integrating segmentation coverage with LLM reasoning offer a promising pathway toward scalable and interpretable biofouling assessment. △ Less

Submitted 27 January, 2026; originally announced January 2026.

Comments: Australasian Conference on Robotics and Automation, ACRA2025 13 Pages, 8 Figures

Quasi Monte Carlo methods enable extremely low-dimensional deep generative models

Authors: Miles Martinez, Alex H. Williams

Abstract: This paper introduces quasi-Monte Carlo latent variable models (QLVMs): a class of deep generative models that are specialized for finding extremely low-dimensional and interpretable embeddings of high-dimensional datasets. Unlike standard approaches, which rely on a learned encoder and variational lower bounds, QLVMs directly approximate the marginal likelihood by randomized quasi-Monte Carlo int… ▽ More This paper introduces quasi-Monte Carlo latent variable models (QLVMs): a class of deep generative models that are specialized for finding extremely low-dimensional and interpretable embeddings of high-dimensional datasets. Unlike standard approaches, which rely on a learned encoder and variational lower bounds, QLVMs directly approximate the marginal likelihood by randomized quasi-Monte Carlo integration. While this brute force approach has drawbacks in higher-dimensional spaces, we find that it excels in fitting one, two, and three dimensional deep latent variable models. Empirical results on a range of datasets show that QLVMs consistently outperform conventional variational autoencoders (VAEs) and importance weighted autoencoders (IWAEs) with matched latent dimensionality. The resulting embeddings enable transparent visualization and post hoc analyses such as nonparametric density estimation, clustering, and geodesic path computation, which are nontrivial to validate in higher-dimensional spaces. While our approach is compute-intensive and struggles to generate fine-scale details in complex datasets, it offers a compelling solution for applications prioritizing interpretability and latent space analysis. △ Less

Submitted 26 January, 2026; originally announced January 2026.

Numerical investigation of the generalized Jang equation coupled to conformal flow of metrics

Authors: Hollis Williams

Abstract: A recent result of Jaracz has established nonexistence of global solutions to the coupled generalized Jang equation and zero divergence system which satisfy the asymptotic conditions needed to prove the Penrose conjecture by identifying a breakdown mechanism for the Jang slope at finite radius. In this work, we investigate whether a similar obstruction arises when the generalized Jang equation is… ▽ More A recent result of Jaracz has established nonexistence of global solutions to the coupled generalized Jang equation and zero divergence system which satisfy the asymptotic conditions needed to prove the Penrose conjecture by identifying a breakdown mechanism for the Jang slope at finite radius. In this work, we investigate whether a similar obstruction arises when the generalized Jang equation is instead coupled to the conformal flow of metrics. Restricting to spherical symmetry and time-symmetric initial data, we formulate a numerically tractable version of the Jang/conformal flow system. Our numerical results show no evidence of a finite radius breakdown analogous to that observed by Jaracz. Instead, the Jang slope remains regular and approaches its limiting value asymptotically. This behavior persists under controlled perturbations of the warping factor, indicating robustness of the observed phenomenon. These findings suggest that coupling to conformal flow of metrics alters the obstruction mechanism present in the Jang/zero divergence system, and hence that this system may still be viable for proving the Penrose conjecture. △ Less

Submitted 20 April, 2026; v1 submitted 21 January, 2026; originally announced January 2026.

Evolution of Hawking mass under perturbative spacetime uniformly expanding flows

Authors: Hollis Williams

Abstract: We present a numerical investigation of the evolution of the Hawking mass for perturbed surfaces evolving under hypersurface-restricted uniformly expanding flows in Minkowski spacetime. Although monotonicity of the Hawking mass under inverse mean curvature flow is well understood, much less is known about the behaviour of such flows in genuine spacetime contexts. To move beyond the totally geodesi… ▽ More We present a numerical investigation of the evolution of the Hawking mass for perturbed surfaces evolving under hypersurface-restricted uniformly expanding flows in Minkowski spacetime. Although monotonicity of the Hawking mass under inverse mean curvature flow is well understood, much less is known about the behaviour of such flows in genuine spacetime contexts. To move beyond the totally geodesic setting where uniformly expanding flows reduce to Euclidean inverse mean curvature flow, we introduce controlled perturbations of the ambient extrinsic curvature. This yields a hypersurface-restricted realization of a spacetime uniformly expanding flow with a nontrivial mean curvature vector. Our results indicate that monotonicity of the Hawking mass remains stable under a range of perturbation amplitudes, angular modes, and spacetime deformations. These results provide evidence for robustness of monotonicity and establish a computational framework for future investigations of uniformly expanding flows in more general spacetime geometries. △ Less

Submitted 23 May, 2026; v1 submitted 17 January, 2026; originally announced January 2026.

Finite-Inertia Corrections and Breakdown of Gor'kov Theory in Acoustic Levitation of Droplets

Authors: Hollis Williams

Abstract: Acoustic levitation is widely used for contactless droplet manipulation, yet the standard Gor'kov description obtained by time averaging the acoustic field lacks a quantitative validity criterion. In this work, we derive Gor'kov theory as the leading-order slow time limit of the instantaneous radiation force, compute the first finite-inertia correction, and obtain a simple breakdown parameter. The… ▽ More Acoustic levitation is widely used for contactless droplet manipulation, yet the standard Gor'kov description obtained by time averaging the acoustic field lacks a quantitative validity criterion. In this work, we derive Gor'kov theory as the leading-order slow time limit of the instantaneous radiation force, compute the first finite-inertia correction, and obtain a simple breakdown parameter. The correction reduces the effective trapping drift and predicts fast time oscillations of amplitude $x_1^{\mathrm{max}}\simλ/8$, corresponding to hundreds of micron for typical ultrasonic levitation experiments. This sets a measurable criterion for experiments using phased transducer arrays. Our results provide a universal rule of thumb for acoustic trap design and clarify where time-averaged radiation force models fail. △ Less

Submitted 31 December, 2025; originally announced January 2026.

Geometry-controlled Onset of Inertial Drag in Granular Impact

Authors: Hollis Williams

Abstract: The impact of solid intruders into granular media is commonly described by a combination of quasi-static resistance and an inertial drag force proportional to the square of the impact speed. While intruder geometry is known to influence force magnitudes, its role in controlling the onset of inertial drag has remained largely unexplored. Here we present systematic impact experiments using conical i… ▽ More The impact of solid intruders into granular media is commonly described by a combination of quasi-static resistance and an inertial drag force proportional to the square of the impact speed. While intruder geometry is known to influence force magnitudes, its role in controlling the onset of inertial drag has remained largely unexplored. Here we present systematic impact experiments using conical intruders spanning a wide range of apex angles. By measuring the peak acceleration during impact, we show that the emergence of a well-defined inertial response depends sensitively on cone geometry. Blunt cones exhibit quadratic scaling with impact speed over the full range of velocities studied, whereas sharper cones display a delayed transition to inertial behavior at higher speeds. We define a geometry-dependent crossover speed marking the onset of the inertial regime and find that it scales approximately linearly with the cone angle through $\tanφ$. Once the inertial regime is established, the peak force collapses when rescaled by $\tanφ$, indicating that cone geometry controls the effective momentum transfer to the grains. These results demonstrate that intruder geometry governs not only the magnitude of inertial drag, but also the impact speed at which it becomes dominant. △ Less

Submitted 27 December, 2025; originally announced December 2025.

Post-Newtonian Constraints on Semiclassical Gravity with Quantum Superpositions

Authors: Hollis Williams

Abstract: Semiclassical gravity, in which a classical spacetime is sourced by the quantum expectation value of the stress-energy tensor, is a standard framework for describing the gravitational interaction of quantum matter. In the nonrelativistic limit this approach leads to the Schrödinger-Newton equation, which is often assumed to be consistent at least in the weak-field regime. In this work, we reexamin… ▽ More Semiclassical gravity, in which a classical spacetime is sourced by the quantum expectation value of the stress-energy tensor, is a standard framework for describing the gravitational interaction of quantum matter. In the nonrelativistic limit this approach leads to the Schrödinger-Newton equation, which is often assumed to be consistent at least in the weak-field regime. In this work, we reexamine this assumption for spatial quantum superpositions of massive particles. We show that, when the quantum state is properly normalized, no modification of the Newtonian gravitational potential arises at leading order. However, at first post-Newtonian order the semiclassical coupling generically produces state-dependent contributions involving the mass density and the mass current of the superposition. These terms have a parametric scaling which is different from that of the corresponding relativistic corrections and which does not have Planck mass suppression. Our results therefore impose a strong post-Newtonian consistency constraint on deterministic semiclassical gravity, indicating that sourcing the metric solely by expectation values is insufficient to recover a consistent relativistic weak-field expansion. △ Less

Submitted 21 December, 2025; originally announced December 2025.

Nonabelian multiplicative integration and curvature obstructions for surface holonomy

Authors: Hollis Williams

Abstract: Surface holonomy plays a central role in higher gauge theory, bundle gerbes and the geometric formulation of Wess--Zumino terms in string theory. In this work, we consider the relation between surface holonomy and nonabelian multiplicative integration on surfaces. In this framework, we interpret the local Stokes law as a curvature obstruction law for higher holonomy and investigate its consequence… ▽ More Surface holonomy plays a central role in higher gauge theory, bundle gerbes and the geometric formulation of Wess--Zumino terms in string theory. In this work, we consider the relation between surface holonomy and nonabelian multiplicative integration on surfaces. In this framework, we interpret the local Stokes law as a curvature obstruction law for higher holonomy and investigate its consequences in the abelian setting. We derive a global three-dimensional Stokes relation and show that it reproduces the familiar Wess-Zumino phase formula. In particular, the phase difference between two surfaces with common boundary is governed by the integral of the corresponding $3$-form curvature over an interpolating three-manifold. These results provide a geometric interpretation of multiplicative integration on surfaces in terms of surface holonomy and clarify its relationship with the classical theory of bundle gerbes and Wess-Zumino terms. We conclude by discussing possible extensions to nonabelian higher gauge theories and their relation to Wilson surface operators and generalized symmetries. △ Less

Submitted 31 May, 2026; v1 submitted 3 December, 2025; originally announced December 2025.

A portable LED-based diamond magnetometer for outreach and teaching labs

Authors: Hollis Williams, Alex Newman, Stuart Graham, Colin Stephen, Gavin Morley

Abstract: We present a compact, low-cost version of an NV center diamond magnetometer which replaces the standard green laser with a high-power LED. This modification improves safety, reduces cost, and allows the green excitation and red photoluminescence to be viewed directly during demonstrations. The device is simple to assemble and suitable for outreach activities and undergraduate laboratories. We show… ▽ More We present a compact, low-cost version of an NV center diamond magnetometer which replaces the standard green laser with a high-power LED. This modification improves safety, reduces cost, and allows the green excitation and red photoluminescence to be viewed directly during demonstrations. The device is simple to assemble and suitable for outreach activities and undergraduate laboratories. We show that it can produce ODMR spectra and respond to nearby magnetic objects, with a sensitivity on the order of 1 $μ$T/$\sqrt{\text{Hz}}$. Supplementary material provides details of the construction and suggestions for student investigations to support use in teaching laboratories. △ Less

Submitted 6 February, 2026; v1 submitted 1 December, 2025; originally announced December 2025.

Accelerating Real-World Overtaking in F1TENTH Racing Employing Reinforcement Learning Methods

Authors: Emily Steiner, Daniel van der Spuy, Futian Zhou, Afereti Pama, Minas Liarokapis, Henry Williams

Abstract: While autonomous racing performance in Time-Trial scenarios has seen significant progress and development, autonomous wheel-to-wheel racing and overtaking are still severely limited. These limitations are particularly apparent in real-life driving scenarios where state-of-the-art algorithms struggle to safely or reliably complete overtaking manoeuvres. This is important, as reliable navigation aro… ▽ More While autonomous racing performance in Time-Trial scenarios has seen significant progress and development, autonomous wheel-to-wheel racing and overtaking are still severely limited. These limitations are particularly apparent in real-life driving scenarios where state-of-the-art algorithms struggle to safely or reliably complete overtaking manoeuvres. This is important, as reliable navigation around other vehicles is vital for safe autonomous wheel-to-wheel racing. The F1Tenth Competition provides a useful opportunity for developing wheel-to-wheel racing algorithms on a standardised physical platform. The competition format makes it possible to evaluate overtaking and wheel-to-wheel racing algorithms against the state-of-the-art. This research presents a novel racing and overtaking agent capable of learning to reliably navigate a track and overtake opponents in both simulation and reality. The agent was deployed on an F1Tenth vehicle and competed against opponents running varying competitive algorithms in the real world. The results demonstrate that the agent's training against opponents enables deliberate overtaking behaviours with an overtaking rate of 87% compared 56% for an agent trained just to race. △ Less

Submitted 29 October, 2025; originally announced October 2025.

Scalable inference of functional neural connectivity at submillisecond timescales

Authors: Arina Medvedeva, Edoardo Balzani, Alex H Williams, Stephen L Keeley

Abstract: The Poisson Generalized Linear Model (GLM) is a foundational tool for analyzing neural spike train data. However, standard implementations rely on discretizing spike times into binned count data, limiting temporal resolution and scalability. Here, we develop Monte Carlo (MC) methods and polynomial approximations (PA) to the continuous-time analog of these models, and show them to be advantageous o… ▽ More The Poisson Generalized Linear Model (GLM) is a foundational tool for analyzing neural spike train data. However, standard implementations rely on discretizing spike times into binned count data, limiting temporal resolution and scalability. Here, we develop Monte Carlo (MC) methods and polynomial approximations (PA) to the continuous-time analog of these models, and show them to be advantageous over their discrete-time counterparts. Further, we propose using a set of exponentially scaled Laguerre polynomials as an orthogonal temporal basis, which improves filter identification and yields closed-form integral solutions under the polynomial approximation. Applied to both synthetic and real spike-time data from rodent hippocampus, our methods demonstrate superior accuracy and scalability compared to traditional binned GLMs, enabling functional connectivity inference in large-scale neural recordings that are temporally precise on the order of synaptic dynamical timescales and in agreement with known anatomical properties of hippocampal subregions. We provide open-source implementations of both MC and PA estimators, optimized for GPU acceleration, to facilitate adoption in the neuroscience community. △ Less

Submitted 23 October, 2025; originally announced October 2025.

Comments: Accepted at NeurIPS 2025

OpenFLAME: Federated Visual Positioning System to Enable Large-Scale Augmented Reality Applications

Authors: Sagar Bharadwaj, Harrison Williams, Luke Wang, Michael Liang, Tao Jin, Srinivasan Seshan, Anthony Rowe

Abstract: World-scale augmented reality (AR) applications need a ubiquitous 6DoF localization backend to anchor content to the real world consistently across devices. Large organizations such as Google and Niantic are 3D scanning outdoor public spaces in order to build their own Visual Positioning Systems (VPS). These centralized VPS solutions fail to meet the needs of many future AR applications -- they do… ▽ More World-scale augmented reality (AR) applications need a ubiquitous 6DoF localization backend to anchor content to the real world consistently across devices. Large organizations such as Google and Niantic are 3D scanning outdoor public spaces in order to build their own Visual Positioning Systems (VPS). These centralized VPS solutions fail to meet the needs of many future AR applications -- they do not cover private indoor spaces because of privacy concerns, regulations, and the labor bottleneck of updating and maintaining 3D scans. In this paper, we present OpenFLAME, a federated VPS backend that allows independent organizations to 3D scan and maintain a separate VPS service for their own spaces. This enables access control of indoor 3D scans, distributed maintenance of the VPS backend, and encourages larger coverage. Sharding of VPS services introduces several unique challenges -- coherency of localization results across spaces, quality control of VPS services, selection of the right VPS service for a location, and many others. We introduce the concept of federated image-based localization and provide reference solutions for managing and merging data across maps without sharing private data. △ Less

Submitted 4 October, 2025; originally announced October 2025.

Aristotle: IMO-level Automated Theorem Proving

Authors: Tudor Achim, Alex Best, Alberto Bietti, Kevin Der, Mathïs Fédérico, Sergei Gukov, Daniel Halpern-Leistner, Kirsten Henningsgard, Yury Kudryashov, Alexander Meiburg, Martin Michelsen, Riley Patterson, Eric Rodriguez, Laura Scharff, Vikram Shanker, Vladmir Sicca, Hari Sowrirajan, Aidan Swope, Matyas Tamas, Vlad Tenev, Jonathan Thomm, Harold Williams, Lawrence Wu

Abstract: We introduce Aristotle, an AI system that combines formal verification with informal reasoning, achieving gold-medal-equivalent performance on the 2025 International Mathematical Olympiad problems. Aristotle integrates three main components: a Lean proof search system, an informal reasoning system that generates and formalizes lemmas, and a dedicated geometry solver. Our system demonstrates state-… ▽ More We introduce Aristotle, an AI system that combines formal verification with informal reasoning, achieving gold-medal-equivalent performance on the 2025 International Mathematical Olympiad problems. Aristotle integrates three main components: a Lean proof search system, an informal reasoning system that generates and formalizes lemmas, and a dedicated geometry solver. Our system demonstrates state-of-the-art performance with favorable scaling properties for automated theorem proving. △ Less

Submitted 10 October, 2025; v1 submitted 1 October, 2025; originally announced October 2025.

An operator-based bound on information and disturbance in quantum measurements

Authors: Hollis Williams, Holger F. Hofmann

Abstract: Quantum measurements can be described by operators that assign conditional probabilities to different outcomes while also describing unavoidable physical changes to the system. Here, we point out that operators describing information gain at minimal disturbance can be expanded into a set of unitary operators representing experimentally distinguishable patterns of disturbance. The observable statis… ▽ More Quantum measurements can be described by operators that assign conditional probabilities to different outcomes while also describing unavoidable physical changes to the system. Here, we point out that operators describing information gain at minimal disturbance can be expanded into a set of unitary operators representing experimentally distinguishable patterns of disturbance. The observable statistics of disturbance defines a tight upper bound on the information gain of the measurement. △ Less

Submitted 15 May, 2026; v1 submitted 29 September, 2025; originally announced October 2025.

A Sentinel-3 foundation model for ocean colour

Authors: Geoffrey Dawson, Remy Vandaele, Andrew Taylor, David Moffat, Helen Tamura-Wicks, Sarah Jackson, Rosie Lickorish, Paolo Fraccaro, Hywel Williams, Chunbo Luo, Anne Jones

Abstract: Artificial Intelligence (AI) Foundation models (FMs), pre-trained on massive unlabelled datasets, have the potential to drastically change AI applications in ocean science, where labelled data are often sparse and expensive to collect. In this work, we describe a new foundation model using the Prithvi-EO Vision Transformer architecture which has been pre-trained to reconstruct data from the Sentin… ▽ More Artificial Intelligence (AI) Foundation models (FMs), pre-trained on massive unlabelled datasets, have the potential to drastically change AI applications in ocean science, where labelled data are often sparse and expensive to collect. In this work, we describe a new foundation model using the Prithvi-EO Vision Transformer architecture which has been pre-trained to reconstruct data from the Sentinel-3 Ocean and Land Colour Instrument (OLCI). We evaluate the model by fine-tuning on two downstream marine earth observation tasks. We first assess model performance compared to current baseline models used to quantify chlorophyll concentration. We then evaluate the FMs ability to refine remote sensing-based estimates of ocean primary production. Our results demonstrate the utility of self-trained FMs for marine monitoring, in particular for making use of small amounts of high quality labelled data and in capturing detailed spatial patterns of ocean colour whilst matching point observations. We conclude that this new generation of geospatial AI models has the potential to provide more robust, data-driven insights into ocean ecosystems and their role in global climate processes. △ Less

Submitted 25 September, 2025; originally announced September 2025.

Comments: 15 pages, 8 figures

Automated classification of natural habitats using ground-level imagery

Authors: Mahdis Tourian, Sareh Rowlands, Remy Vandaele, Max Fancourt, Rebecca Mein, Hywel T. P. Williams

Abstract: Accurate classification of terrestrial habitats is critical for biodiversity conservation, ecological monitoring, and land-use planning. Several habitat classification schemes are in use, typically based on analysis of satellite imagery with validation by field ecologists. Here we present a methodology for classification of habitats based solely on ground-level imagery (photographs), offering impr… ▽ More Accurate classification of terrestrial habitats is critical for biodiversity conservation, ecological monitoring, and land-use planning. Several habitat classification schemes are in use, typically based on analysis of satellite imagery with validation by field ecologists. Here we present a methodology for classification of habitats based solely on ground-level imagery (photographs), offering improved validation and the ability to classify habitats at scale (for example using citizen-science imagery). In collaboration with Natural England, a public sector organisation responsible for nature conservation in England, this study develops a classification system that applies deep learning to ground-level habitat photographs, categorising each image into one of 18 classes defined by the 'Living England' framework. Images were pre-processed using resizing, normalisation, and augmentation; re-sampling was used to balance classes in the training data and enhance model robustness. We developed and fine-tuned a DeepLabV3-ResNet101 classifier to assign a habitat class label to each photograph. Using five-fold cross-validation, the model demonstrated strong overall performance across 18 habitat classes, with accuracy and F1-scores varying between classes. Across all folds, the model achieved a mean F1-score of 0.61, with visually distinct habitats such as Bare Soil, Silt and Peat (BSSP) and Bare Sand (BS) reaching values above 0.90, and mixed or ambiguous classes scoring lower. These findings demonstrate the potential of this approach for ecological monitoring. Ground-level imagery is readily obtained, and accurate computational methods for habitat classification based on such data have many potential applications. To support use by practitioners, we also provide a simple web application that classifies uploaded images using our model. △ Less

Submitted 26 August, 2025; originally announced August 2025.

Comments: 15 pages, 6 figures, 2 tables

JWST's PEARLS: A Candidate Massive Binary Star System in a Lensed Galaxy at Redshift 0.94

Authors: Hayley Williams, Patrick L. Kelly, Emmanouil Zapartas, Rogier A. Windhorst, Christopher J. Conselice, Seth H. Cohen, Birendra Dhanasingham, Jose M. Diego, Alexei V. Filippenko, Brenda L. Frye, Benne W. Holwerda, Terry J. Jones, Anton M. Koekemoer, Ashish Kumar Meena, Massimo Ricotti, Clayton D. Robertson, Payaswini Saikia, Bangzheng Sun, S. P. Willner, Haojing Yan, Adi Zitrin

Abstract: Massive stars at cosmological distances can be individually detected during transient microlensing events, when gravitational lensing magnifications may exceed ~1000. Nine such sources were identified in JWST NIRCam imaging of a single galaxy at redshift z=0.94 known as the "Warhol arc,'' which is mirror-imaged by the galaxy cluster MACSJ0416.1-2403. Here we present the discovery of two coincident… ▽ More Massive stars at cosmological distances can be individually detected during transient microlensing events, when gravitational lensing magnifications may exceed ~1000. Nine such sources were identified in JWST NIRCam imaging of a single galaxy at redshift z=0.94 known as the "Warhol arc,'' which is mirror-imaged by the galaxy cluster MACSJ0416.1-2403. Here we present the discovery of two coincident and well-characterized microlensing events at the same location followed by a third event observed in a single filter approximately 18 months later. The events can be explained by microlensing of a binary star system consisting of a red supergiant (T ~ 4000 K) and a B-type (T ~ 13,000 K) companion star. The timescale of the coincident microlensing events constrains the estimated projected source-plane size to tens of AU. The most likely binary configurations consistent with the observational constraints on the temperature and luminosity of each star are stars with initial masses M1=23.6+5.3-4.3 Msol and an initial mass ratio between the two stars close to unity. A kinematic model that reproduces the observed light curve in all filters gives a relatively small transverse velocity of 50 km/s. This requires the dominant velocity component of several hundreds of km/s to be roughly parallel to the microcaustic. An alternative possibility would be that the three microlensing events correspond to unrelated stars crossing distinct microcaustics, but this would imply a highly elevated rate of events at their common position, even though no underlying knot is present at the location. △ Less

Submitted 18 November, 2025; v1 submitted 3 July, 2025; originally announced July 2025.

JWST's PEARLS: Temperatures of Nine Highly Magnified Stars in a Galaxy at Redshift 0.94 and Simulated Stellar Population Dependence on Stellar Metallicity and the Initial Mass Function

Authors: Hayley Williams, Patrick L. Kelly, Rogier A. Windhorst, Alexei V. Filippenko, Amruth Alfred, Tom Broadhurst, Wenlei Chen, Christopher J. Conselice, Seth H. Cohen, Jose M. Diego, Benne W. Holwerda, Anton M. Koekemoer, Sung Kei Li, Ashish Kumar Meena, Jose M. Palencia, Massimo Ricotti, Clayton D. Robertson, Bangzheng Sun, Haojing Yan, Adi Zitrin

Abstract: We present stellar atmosphere modeling of JWST NIRCam photometry of nine highly magnified individual stars in a single galaxy at redshift z=0.94 known as the Warhol arc, which is strongly lensed by the galaxy cluster MACSJ0416. Seven of these transients were identified by Yan et al. (2023). The nine sources are all likely red supergiants with temperatures of T~4000K. We present new longslit spectr… ▽ More We present stellar atmosphere modeling of JWST NIRCam photometry of nine highly magnified individual stars in a single galaxy at redshift z=0.94 known as the Warhol arc, which is strongly lensed by the galaxy cluster MACSJ0416. Seven of these transients were identified by Yan et al. (2023). The nine sources are all likely red supergiants with temperatures of T~4000K. We present new longslit spectroscopy of the Warhol arc acquired with Keck-I and the Large Binocular Telescope, and use these data to constrain the arc's oxygen abundance to be 12+log(O/H)=8.45+-0.08. We perform a microlensing simulation on synthetic stellar populations using a range of stellar metallicities and initial mass function slopes. The temperature distribution of the simulated detectable stars is sensitive to the choice of stellar metallicity, and setting the stellar metallicity equal to the arc's nebular metallicity (log(Z*/Zsun)=-0.24) produces a simulated temperature distribution that is consistent with the observations, while lower stellar metallicities (log(Z*/Zsun)<-0.75) produce simulated temperatures that are inconsistent with the observations. The expected detection rate is strongly anticorrelated with the IMF slope for alpha>1.2. For the canonical IMF slope alpha=2.35, the simulation yields expected transient detection rates that agree with the observed detection rates in the HST Flashlights filters, but over predicts the detection rate by a factor of ~3-12 (<2sigma tension) in the JWST filters. The simulated detection rate is sensitive to the choice of stellar metallicity, with lower metallicities (log(Z*/Zsun)<-0.75) yielding a significantly lower simulated detection rate that further reduces the modest tension with the observations. △ Less

Submitted 18 November, 2025; v1 submitted 3 July, 2025; originally announced July 2025.

Transient star B/R ratio and star formation in $z\gtrsim 1$ lensed galaxies

Authors: Sung Kei Li, Jose M. Palencia, Jose M. Diego, Jeremy Lim, Patrick L. Kelly, Ashish K. Meena, James Nianias, Hayley Williams, Liliya L. R. Williams, Adi Zitrin, Thomas J. Broadhurst

Abstract: The extreme magnification from galaxy clusters and microlenses therein allows the detection of individual, luminous stars in lensed galaxies as transient events, and hence provides a valuable window into the high mass stellar population in $z>1$ galaxies. As these bright stars can only be formed at specific ages, the relative abundance of transient events at blue (B) and red (R) optical wavelength… ▽ More The extreme magnification from galaxy clusters and microlenses therein allows the detection of individual, luminous stars in lensed galaxies as transient events, and hence provides a valuable window into the high mass stellar population in $z>1$ galaxies. As these bright stars can only be formed at specific ages, the relative abundance of transient events at blue (B) and red (R) optical wavelengths ($B/R$ ratio) can provide insights into the recent star formation history of galaxies that are not well constrained by their spectral energy distributions (SEDs). Here, we forward model the transient detection rates in an idealized mock scenario to find that the $B/R$ ratio of strongly lensed $z>1$ galaxies decreases quickly with increasing age. This ratio has moderate sensitivity to metallicity and comparatively low sensitivity to dust attenuation, with no significant dependency on the stellar initial mass function. Fitting model stellar populations to either the SED or $B/R$ ratio alone of ``Warhol'' arc ($z = 0.94$), we find that neither a simple single starburst nor a more complex star formation can simultaneously reproduce both constraints. We then demonstrate that a best-fit model constrained by both the B/R ratio and SED requires a star-formation rate that has varied quite dramatically over the past $\sim$50 Myr, for which the total stellar mass formed over this time is a factor of 10 (with $2-3σ$ significance) different from the best-fit models to the SED alone. Our work shows that the transient $B/R$ ratio can be used as an additional powerful constraint on the recent star formation history of higher-redshift galaxies in future works that are strongly lensed by galaxy clusters. △ Less

Submitted 7 October, 2025; v1 submitted 20 June, 2025; originally announced June 2025.

Comments: 12 pages, 14 figures, accepted for publication in Astronomy & Astrophysics

Journal ref: A&A 704, A309 (2025)

Contact isotopies in the coherent-constructible correspondence

Authors: Jishnu Bose, Harold Williams

Abstract: The coherent-constructible correspondence is a realization of toric mirror symmetry in which the A-side is modeled by constructible sheaves on $T^n$. This paper provides a geometric realization of the mirror Picard group action in this correspondence, characterizing it in terms of quantized contact isotopies and providing a sheaf-theoretic counterpart to work of Hanlon in the Fukaya-Seidel setting… ▽ More The coherent-constructible correspondence is a realization of toric mirror symmetry in which the A-side is modeled by constructible sheaves on $T^n$. This paper provides a geometric realization of the mirror Picard group action in this correspondence, characterizing it in terms of quantized contact isotopies and providing a sheaf-theoretic counterpart to work of Hanlon in the Fukaya-Seidel setting. Given a toric Cartier divisor $D$, we consider a family of homogeneous Hamiltonians $H_\varepsilon$ on $T^* T^n$. Their flows act on sheaves via a family of kernels $K_\varepsilon$ on $T^n \times T^n$. The nearby cycles kernel $K_0$ corresponds heuristically to the Hamiltonian flow of the non-differentiable function $\lim_{\varepsilon \to 0} H_\varepsilon$, which is the pullback of the support function of $D$ along the cofiber projection. We show that the action of $K_0$ coincides with the convolution action of the associated twisted polytope sheaf, hence mirrors the action of $\mathcal{O}(D)$ on coherent sheaves. △ Less

Submitted 8 May, 2025; originally announced May 2025.

Comments: 26 pages

Future Circular Collider Feasibility Study Report: Volume 2, Accelerators, Technical Infrastructure and Safety

Authors: M. Benedikt, F. Zimmermann, B. Auchmann, W. Bartmann, J. P. Burnet, C. Carli, A. Chancé, P. Craievich, M. Giovannozzi, C. Grojean, J. Gutleber, K. Hanke, A. Henriques, P. Janot, C. Lourenço, M. Mangano, T. Otto, J. Poole, S. Rajagopalan, T. Raubenheimer, E. Todesco, L. Ulrici, T. Watson, G. Wilkinson, A. Abada , et al. (1439 additional authors not shown)

Abstract: In response to the 2020 Update of the European Strategy for Particle Physics, the Future Circular Collider (FCC) Feasibility Study was launched as an international collaboration hosted by CERN. This report describes the FCC integrated programme, which consists of two stages: an electron-positron collider (FCC-ee) in the first phase, serving as a high-luminosity Higgs, top, and electroweak factory;… ▽ More In response to the 2020 Update of the European Strategy for Particle Physics, the Future Circular Collider (FCC) Feasibility Study was launched as an international collaboration hosted by CERN. This report describes the FCC integrated programme, which consists of two stages: an electron-positron collider (FCC-ee) in the first phase, serving as a high-luminosity Higgs, top, and electroweak factory; followed by a proton-proton collider (FCC-hh) at the energy frontier in the second phase. FCC-ee is designed to operate at four key centre-of-mass energies: the Z pole, the WW production threshold, the ZH production peak, and the top/anti-top production threshold - delivering the highest possible luminosities to four experiments. Over 15 years of operation, FCC-ee will produce more than 6 trillion Z bosons, 200 million WW pairs, nearly 3 million Higgs bosons, and 2 million top anti-top pairs. Precise energy calibration at the Z pole and WW threshold will be achieved through frequent resonant depolarisation of pilot bunches. The sequence of operation modes remains flexible. FCC-hh will operate at a centre-of-mass energy of approximately 85 TeV - nearly an order of magnitude higher than the LHC - and is designed to deliver 5 to 10 times the integrated luminosity of the HL-LHC. Its mass reach for direct discovery extends to several tens of TeV. In addition to proton-proton collisions, FCC-hh is capable of supporting ion-ion, ion-proton, and lepton-hadron collision modes. This second volume of the Feasibility Study Report presents the complete design of the FCC-ee collider, its operation and staging strategy, the full-energy booster and injector complex, required accelerator technologies, safety concepts, and technical infrastructure. It also includes the design of the FCC-hh hadron collider, development of high-field magnets, hadron injector options, and key technical systems for FCC-hh. △ Less

Submitted 25 April, 2025; originally announced May 2025.

Comments: 627 pages. Please address any comment or request to fcc.secretariat@cern.ch

Report number: CERN-FCC-ACC-2025-0004

Future Circular Collider Feasibility Study Report: Volume 3, Civil Engineering, Implementation and Sustainability

Authors: M. Benedikt, F. Zimmermann, B. Auchmann, W. Bartmann, J. P. Burnet, C. Carli, A. Chancé, P. Craievich, M. Giovannozzi, C. Grojean, J. Gutleber, K. Hanke, A. Henriques, P. Janot, C. Lourenço, M. Mangano, T. Otto, J. Poole, S. Rajagopalan, T. Raubenheimer, E. Todesco, L. Ulrici, T. Watson, G. Wilkinson, P. Azzi , et al. (1439 additional authors not shown)

Abstract: Volume 3 of the FCC Feasibility Report presents studies related to civil engineering, the development of a project implementation scenario, and environmental and sustainability aspects. The report details the iterative improvements made to the civil engineering concepts since 2018, taking into account subsurface conditions, accelerator and experiment requirements, and territorial considerations. I… ▽ More Volume 3 of the FCC Feasibility Report presents studies related to civil engineering, the development of a project implementation scenario, and environmental and sustainability aspects. The report details the iterative improvements made to the civil engineering concepts since 2018, taking into account subsurface conditions, accelerator and experiment requirements, and territorial considerations. It outlines a technically feasible and economically viable civil engineering configuration that serves as the baseline for detailed subsurface investigations, construction design, cost estimation, and project implementation planning. Additionally, the report highlights ongoing subsurface investigations in key areas to support the development of an improved 3D subsurface model of the region. The report describes development of the project scenario based on the 'avoid-reduce-compensate' iterative optimisation approach. The reference scenario balances optimal physics performance with territorial compatibility, implementation risks, and costs. Environmental field investigations covering almost 600 hectares of terrain - including numerous urban, economic, social, and technical aspects - confirmed the project's technical feasibility and contributed to the preparation of essential input documents for the formal project authorisation phase. The summary also highlights the initiation of public dialogue as part of the authorisation process. The results of a comprehensive socio-economic impact assessment, which included significant environmental effects, are presented. Even under the most conservative and stringent conditions, a positive benefit-cost ratio for the FCC-ee is obtained. Finally, the report provides a concise summary of the studies conducted to document the current state of the environment. △ Less

Submitted 25 April, 2025; originally announced May 2025.

Comments: 357 pages. Please address any comment or request to fcc.secretariat@cern.ch

Report number: CERN-FCC-ACC-2025-0003

Future Circular Collider Feasibility Study Report: Volume 1, Physics, Experiments, Detectors

Authors: M. Benedikt, F. Zimmermann, B. Auchmann, W. Bartmann, J. P. Burnet, C. Carli, A. Chancé, P. Craievich, M. Giovannozzi, C. Grojean, J. Gutleber, K. Hanke, A. Henriques, P. Janot, C. Lourenço, M. Mangano, T. Otto, J. Poole, S. Rajagopalan, T. Raubenheimer, E. Todesco, L. Ulrici, T. Watson, G. Wilkinson, P. Azzi , et al. (1439 additional authors not shown)

Abstract: Volume 1 of the FCC Feasibility Report presents an overview of the physics case, experimental programme, and detector concepts for the Future Circular Collider (FCC). This volume outlines how FCC would address some of the most profound open questions in particle physics, from precision studies of the Higgs and EW bosons and of the top quark, to the exploration of physics beyond the Standard Model.… ▽ More Volume 1 of the FCC Feasibility Report presents an overview of the physics case, experimental programme, and detector concepts for the Future Circular Collider (FCC). This volume outlines how FCC would address some of the most profound open questions in particle physics, from precision studies of the Higgs and EW bosons and of the top quark, to the exploration of physics beyond the Standard Model. The report reviews the experimental opportunities offered by the staged implementation of FCC, beginning with an electron-positron collider (FCC-ee), operating at several centre-of-mass energies, followed by a hadron collider (FCC-hh). Benchmark examples are given of the expected physics performance, in terms of precision and sensitivity to new phenomena, of each collider stage. Detector requirements and conceptual designs for FCC-ee experiments are discussed, as are the specific demands that the physics programme imposes on the accelerator in the domains of the calibration of the collision energy, and the interface region between the accelerator and the detector. The report also highlights advances in detector, software and computing technologies, as well as the theoretical tools /reconstruction techniques that will enable the precision measurements and discovery potential of the FCC experimental programme. This volume reflects the outcome of a global collaborative effort involving hundreds of scientists and institutions, aided by a dedicated community-building coordination, and provides a targeted assessment of the scientific opportunities and experimental foundations of the FCC programme. △ Less

Submitted 25 April, 2025; originally announced May 2025.

Comments: 290 pages. Please address any comment or request to fcc.secretariat@cern.ch

Report number: CERN-FCC-PHYS-2025-0002

Microlensing at Cosmological Distances: Event Rate Predictions in the Warhol Arc of MACS 0416

Authors: J. M. Palencia, J. M. Diego, L. Dai, M. Pascale, R. Windhorst, A. M. Koekemoer, Sung Kei Li, B. J. Kavanagh, Fengwu Sun, Amruth Alfred, Ashish K. Meena, Thomas J. Broadhurst, Patrick L. Kelly, Derek Perera, Hayley Williams, Adi Zitrin

Abstract: Highly magnified stars ($μ$ $>$ 100) are now outinely identified as transient events at cosmological distances thanks to microlensing by intra-cluster stars near the critical curves of galaxy clusters. Using the {\it James Webb} Space Telescope (JWST) in combination with the {\it Hubble} Space Telescope (HST), we outline here an analytical framework that is applied to the Warhol arc (at $z=0.94$)… ▽ More Highly magnified stars ($μ$ $>$ 100) are now outinely identified as transient events at cosmological distances thanks to microlensing by intra-cluster stars near the critical curves of galaxy clusters. Using the {\it James Webb} Space Telescope (JWST) in combination with the {\it Hubble} Space Telescope (HST), we outline here an analytical framework that is applied to the Warhol arc (at $z=0.94$) in the MACS 0416 galaxy cluster (at $z=0.396)$ where over a dozen microlensed stars have been detected to date. This method is general and can be applied to other lensed arcs. Within this lensed galaxy we fit the spatially resolved SED spanned by eight JWST-NIRCam filters combined with three ACS filters, for accurate lensed star predictions in 2D. With this tool we can generate 2D maps of microlensed stars for well resolved arcs in general, including dependence on wavelength and limiting apparent magnitude, for comparison with with planned cadenced campaigns for JWST and Hubble, for constraining directly the IMF and the level of dark matter substructure. △ Less

Submitted 28 April, 2025; v1 submitted 9 April, 2025; originally announced April 2025.

Comments: 19 pages, 13 figures

Journal ref: A&A 699, A295 (2025)

Constraining the z $\sim$ 1 Initial Mass Function with {\it HST} and {\it JWST} Lensed Stars in MACS J0416.1-2403

Authors: Sung Kei Li, Jose M. Diego, Ashish K. Meena, Jeremy Lim, Leo W. H. Fung, Arsen Levitskiy, James Nianias, Jose M. Palencia, Hayley Williams, Jiashuo Zhang, Alfred Amruth, Thomas J. Broadhurst, Wenlei Chen, Alexei V. Filippenko, Patrick L. Kelly, Anton M. Koekemoer, Derek Perera, Bangzheng Sun, Liliya L. R. Williams, Rogier A. Windhorst, Haojin Yan, Adi Zitrin

Abstract: Our understanding of galaxy properties and evolution is contingent on knowing the initial mass function (IMF), and yet to date, the IMF is constrained only to local galaxies. Individual stars are now becoming routinely detected at cosmological distances, where luminous stars such as supergiants in background galaxies strongly lensed by galaxy clusters are temporarily further magnified by huge fact… ▽ More Our understanding of galaxy properties and evolution is contingent on knowing the initial mass function (IMF), and yet to date, the IMF is constrained only to local galaxies. Individual stars are now becoming routinely detected at cosmological distances, where luminous stars such as supergiants in background galaxies strongly lensed by galaxy clusters are temporarily further magnified by huge factors (up to $10^{4}$) by intracluster stars, thus being detected as transients. The detection rate of these events depends on the abundance of luminous stars in the background galaxy and is thus sensitive to the IMF and the star-formation history (SFH), especially for the blue supergiants detected as transients in the rest-frame ultraviolet/optical filters. As a proof of concept, we use simple SFH and IMF models constrained by spectral energy distributions (SEDs) to see how well we can predict the {\it HST} and {\it JWST} transient detection rate in a lensed arc dubbed ``Spock'' ($z = 1.0054$). We find that demanding a simultaneous fit of the SED and the transient detection rate places constraints on the IMF, independent of the assumed simple SFH model. We conclude our likelihood analysis indicates that the data definitively prefers the ``Spock'' galaxy to have a Salpeter IMF ($α= 2.35$) rather than a Top-heavy IMF ($α= 1$) -- which is thought to be the case in the early universe -- with no clear excess of supergiants above the standard IMF. △ Less

Submitted 12 June, 2025; v1 submitted 9 April, 2025; originally announced April 2025.

Comments: 21 pages, 7 figures, accepted for publication in ApJ

The Linear Collider Facility (LCF) at CERN

Authors: H. Abramowicz, E. Adli, F. Alharthi, M. Almanza-Soto, M. M. Altakach, S. Ampudia Castelazo, D. Angal-Kalinin, J. A. Anguiano, R. B. Appleby, O. Apsimon, A. Arbey, O. Arquero, D. Attié, J. L. Avila-Jimenez, H. Baer, Y. Bai, C. Balazs, P. Bambade, T. Barklow, J. Baudot, P. Bechtle, T. Behnke, A. B. Bellerive, S. Belomestnykh, Y. Benhammou , et al. (386 additional authors not shown)

Abstract: In this paper we outline a proposal for a Linear Collider Facility as the next flagship project for CERN. It offers the opportunity for a timely, cost-effective and staged construction of a new collider that will be able to comprehensively map the Higgs boson's properties, including the Higgs field potential, thanks to a large span in centre-of-mass energies and polarised beams. A comprehensive pr… ▽ More In this paper we outline a proposal for a Linear Collider Facility as the next flagship project for CERN. It offers the opportunity for a timely, cost-effective and staged construction of a new collider that will be able to comprehensively map the Higgs boson's properties, including the Higgs field potential, thanks to a large span in centre-of-mass energies and polarised beams. A comprehensive programme to study the Higgs boson and its closest relatives with high precision requires data at centre-of-mass energies from the Z pole to at least 1 TeV. It should include measurements of the Higgs boson in both major production mechanisms, ee -> ZH and ee -> vvH, precision measurements of gauge boson interactions as well as of the W boson, Higgs boson and top-quark masses, measurement of the top-quark Yukawa coupling through ee ->ttH, measurement of the Higgs boson self-coupling through HH production, and precision measurements of the electroweak couplings of the top quark. In addition, ee collisions offer discovery potential for new particles complementary to HL-LHC. △ Less

Submitted 19 June, 2025; v1 submitted 31 March, 2025; originally announced March 2025.

Comments: Submission to the ESPPU, as updated version May 26

Report number: DESY-25-054

DiPolMol-Py: A Python package for calculations for $^2Σ$ ground-state molecules

Authors: Bethan Humphreys, Alex J. Matthies, Hannah J. Williams

Abstract: We present the python package DiPolMol-Py, which can be used to calculate the rotational and hyperfine structure of $^2Σ$ molecules. The calculations can be performed in the presence of dc magnetic fields, dc electric fields and far off-resonant optical fields. We additionally include functions to calculate the polarisability of the molecule and the transition dipole moment between different energ… ▽ More We present the python package DiPolMol-Py, which can be used to calculate the rotational and hyperfine structure of $^2Σ$ molecules. The calculations can be performed in the presence of dc magnetic fields, dc electric fields and far off-resonant optical fields. We additionally include functions to calculate the polarisability of the molecule and the transition dipole moment between different energy eigenstates. The package is applicable to many of the molecules which can be laser cooled, specifically the alkaline earth fluorides. We provide a constants file which includes many of the required literature values for CaF, SrF and BaF. Additional species can easily be added by updating this file. △ Less

Submitted 27 March, 2025; originally announced March 2025.

Comments: 25 pages, 5 figures

A Linear Collider Vision for the Future of Particle Physics

Authors: H. Abramowicz, E. Adli, F. Alharthi, M. Almanza-Soto, M. M. Altakach, W. Altmannshofer, S. Ampudia Castelazo, D. Angal-Kalinin, J. A. Anguiano, R. B. Appleby, O. Apsimon, A. Arbey, F. Arco, O. Arquero, A. Aryshev, S. Asai, D. Attie, J. L. Avila-Jimenez, H. Baer, J. A. Bagger, Y. Bai, I. R. Bailey, C. Balazs, P. Bambade, T. Barklow , et al. (426 additional authors not shown)

Abstract: In this paper we review the physics opportunities at linear $e^+e^-$ colliders with a special focus on high centre-of-mass energies and beam polarisation, take a fresh look at the various accelerator technologies available or under development and, for the first time, discuss how a facility first equipped with a technology mature today could be upgraded with technologies of tomorrow to reach much… ▽ More In this paper we review the physics opportunities at linear $e^+e^-$ colliders with a special focus on high centre-of-mass energies and beam polarisation, take a fresh look at the various accelerator technologies available or under development and, for the first time, discuss how a facility first equipped with a technology mature today could be upgraded with technologies of tomorrow to reach much higher energies and/or luminosities. In addition, we will discuss detectors and alternative collider modes, as well as opportunities for beyond-collider experiments and R\&D facilities as part of a linear collider facility (LCF). The material of this paper will support all plans for $e^+e^-$ linear colliders and additional opportunities they offer, independently of technology choice or proposed site, as well as R\&D for advanced accelerator technologies. This joint perspective on the physics goals, early technologies and upgrade strategies has been developed by the LCVision team based on an initial discussion at LCWS2024 in Tokyo and a follow-up at the LCVision Community Event at CERN in January 2025. It heavily builds on decades of achievements of the global linear collider community, in particular in the context of CLIC and ILC. △ Less

Submitted 23 December, 2025; v1 submitted 25 March, 2025; originally announced March 2025.

Comments: Community document for EPPSU, major updates of several sections, in particular the global interpretation section

Modeling Neural Activity with Conditionally Linear Dynamical Systems

Authors: Victor Geadah, Amin Nejatbakhsh, David Lipshutz, Jonathan W. Pillow, Alex H. Williams

Abstract: Neural population activity exhibits complex, nonlinear dynamics, varying in time, over trials, and across experimental conditions. Here, we develop Conditionally Linear Dynamical System (CLDS) models as a general-purpose method to characterize these dynamics. These models use Gaussian Process (GP) priors to capture the nonlinear dependence of circuit dynamics on task and behavioral variables. Cond… ▽ More Neural population activity exhibits complex, nonlinear dynamics, varying in time, over trials, and across experimental conditions. Here, we develop Conditionally Linear Dynamical System (CLDS) models as a general-purpose method to characterize these dynamics. These models use Gaussian Process (GP) priors to capture the nonlinear dependence of circuit dynamics on task and behavioral variables. Conditioned on these covariates, the data is modeled with linear dynamics. This allows for transparent interpretation and tractable Bayesian inference. We find that CLDS models can perform well even in severely data-limited regimes (e.g. one trial per condition) due to their Bayesian formulation and ability to share statistical power across nearby task conditions. In example applications, we apply CLDS to model thalamic neurons that nonlinearly encode heading direction and to model motor cortical neurons during a cued reaching task. △ Less

Submitted 30 October, 2025; v1 submitted 25 February, 2025; originally announced February 2025.

Comments: 24 pages, 7 figures. Associated code available at: https://github.com/neurostatslab/clds. To appear at the 39th Conference on Neural Information Processing Systems (NeurIPS 2025)

OrchardDepth: Precise Metric Depth Estimation of Orchard Scene from Monocular Camera Images

Authors: Zhichao Zheng, Henry Williams, Bruce A MacDonald

Abstract: Monocular depth estimation is a rudimentary task in robotic perception. Recently, with the development of more accurate and robust neural network models and different types of datasets, monocular depth estimation has significantly improved performance and efficiency. However, most of the research in this area focuses on very concentrated domains. In particular, most of the benchmarks in outdoor sc… ▽ More Monocular depth estimation is a rudimentary task in robotic perception. Recently, with the development of more accurate and robust neural network models and different types of datasets, monocular depth estimation has significantly improved performance and efficiency. However, most of the research in this area focuses on very concentrated domains. In particular, most of the benchmarks in outdoor scenarios belong to urban environments for the improvement of autonomous driving devices, and these benchmarks have a massive disparity with the orchard/vineyard environment, which is hardly helpful for research in the primary industry. Therefore, we propose OrchardDepth, which fills the gap in the estimation of the metric depth of the monocular camera in the orchard/vineyard environment. In addition, we present a new retraining method to improve the training result by monitoring the consistent regularization between dense depth maps and sparse points. Our method improves the RMSE of depth estimation in the orchard environment from 1.5337 to 0.6738, proving our method's validation. △ Less

Submitted 20 February, 2025; originally announced February 2025.

Comments: 10 pages, 5 figures, Australasian Conference on Robotics and Automation, ACRA, 2024

Lorentz-violating pseudovectors in effective field theories for quantum gravity

Authors: Hollis Williams

Abstract: Effective field theories which describe the coupling between gravity and matter fields have recently been extended to include terms with operators of non-minimal mass dimension. These terms preserve the usual gauge symmetries but may violate local Lorentz and diffeomorphism invariance. The number of possible terms in the field theory explodes once one allows for non-minimal operators, with no crit… ▽ More Effective field theories which describe the coupling between gravity and matter fields have recently been extended to include terms with operators of non-minimal mass dimension. These terms preserve the usual gauge symmetries but may violate local Lorentz and diffeomorphism invariance. The number of possible terms in the field theory explodes once one allows for non-minimal operators, with no criterion to choose between them. We suggest as such a criterion to focus on terms which violate Lorentz invariance via a (pseudo)vector background field, leaving a number of possible terms in the Higgs, gauge and gravitational sectors. Further study of these terms is motivated by the proposed correspondence between the general effective theory for Lorentz violation and emergent Lorentz symmetry in condensed-matter systems, which is mostly unexplored for higher mass dimension operators and couplings to gauge fields and gravity. We suggest bounds in the Higgs sector and we show that some of the coefficients in the gauge sector vanish at one loop, whereas others have bounds which are comparable with those suggested by Kostelecký and Li for coefficients in Lorentz-violating QCD and QED coupled to quarks. We also find new bounds in the gravitational sector by considering Robertson-Walker cosmology. Finally, we discuss the special case where only diffeomorphism invariance is spontaneously broken and explain why it does not allow for non-trivial Nambu-Goldstone modes. △ Less

Submitted 2 February, 2025; originally announced February 2025.

Journal ref: Acta Phys. Pol. B 55, 11-A3 (2024)

Reward Prediction Error Prioritisation in Experience Replay: The RPE-PER Method

Authors: Hoda Yamani, Yuning Xing, Lee Violet C. Ong, Bruce A. MacDonald, Henry Williams

Abstract: Reinforcement Learning algorithms aim to learn optimal control strategies through iterative interactions with an environment. A critical element in this process is the experience replay buffer, which stores past experiences, allowing the algorithm to learn from a diverse range of interactions rather than just the most recent ones. This buffer is especially essential in dynamic environments with li… ▽ More Reinforcement Learning algorithms aim to learn optimal control strategies through iterative interactions with an environment. A critical element in this process is the experience replay buffer, which stores past experiences, allowing the algorithm to learn from a diverse range of interactions rather than just the most recent ones. This buffer is especially essential in dynamic environments with limited experiences. However, efficiently selecting high-value experiences to accelerate training remains a challenge. Drawing inspiration from the role of reward prediction errors (RPEs) in biological systems, where they are essential for adaptive behaviour and learning, we introduce Reward Predictive Error Prioritised Experience Replay (RPE-PER). This novel approach prioritises experiences in the buffer based on RPEs. Our method employs a critic network, EMCN, that predicts rewards in addition to the Q-values produced by standard critic networks. The discrepancy between these predicted and actual rewards is computed as RPE and utilised as a signal for experience prioritisation. Experimental evaluations across various continuous control tasks demonstrate RPE-PER's effectiveness in enhancing the learning speed and performance of off-policy actor-critic algorithms compared to baseline approaches. △ Less

Submitted 29 January, 2025; originally announced January 2025.

Comments: This paper was accepted for presentation at the 2024 Australasian Conference on Robotics and Automation (ACRA 2024). It consists of 10 pages, including four figures and two tables

Tropical Lagrangian coamoebae and free resolutions

Authors: Christopher Kuo, Harold Williams

Abstract: We study the coamoebae of Lagrangian submanifolds of $(\mathbb{C}^\times)^n$, specifically how the combinatorics of their degenerations encodes the homological algebra of mirror coherent sheaves. Concretely, to a minimal free resolution $F^\bullet$ of a module $M$ over $\mathbb{C}[z_1^{\pm 1}, \dotsc, z_n^{\pm 1}]$ we associate a simplicial complex $T(F^\bullet) \subset T^n$. We call… ▽ More We study the coamoebae of Lagrangian submanifolds of $(\mathbb{C}^\times)^n$, specifically how the combinatorics of their degenerations encodes the homological algebra of mirror coherent sheaves. Concretely, to a minimal free resolution $F^\bullet$ of a module $M$ over $\mathbb{C}[z_1^{\pm 1}, \dotsc, z_n^{\pm 1}]$ we associate a simplicial complex $T(F^\bullet) \subset T^n$. We call $T(F^\bullet)$ a tropical Lagrangian coamoeba. We show that the discrete information in $F^\bullet$ can often be recovered from $T(F^\bullet)$, and that more generally $M$ is mirror to a certain constructible sheaf supported on $T(F^\bullet)$. The resulting interplay between coherent sheaves on $(\mathbb{C}^\times)^n$ and simplicial complexes in $T^n$ provides a higher-dimensional generalization of the spectral theory of dimer models in $T^2$, as well as a symplectic counterpart to the theory of brane brick models. △ Less

Submitted 23 December, 2024; originally announced December 2024.

Comments: 43 pages