Search

In image-based reinforcement learning (RL), policies usually operate in two steps: first extracting lower-dimensional features from raw images (the "recognition" step), and then taking actions based on the extracted features (the "decision" step). Extracting features that are spuriously correlated with performance or irrelevant for decision-making can lead to poor generalization performance, known as observational overfitting in image-based RL. In such cases, it can be hard to quantify how much of the error can be attributed to poor feature extraction vs. poor decision-making. In order to disentangle the two sources of error, we introduce the notions of recognition regret and decision regret. Using these notions, we characterize and disambiguate the two distinct causes behind observational overfitting: over-specific representations, which include features that are not needed for optimal decision-making (leading to high decision regret), vs. under-specific representations, which only include a limited set of features that were spuriously correlated with performance during training (leading to high recognition regret). Finally, we provide illustrative examples of observational overfitting due to both over-specific and under-specific representations in maze environments as well as the Atari game Pong.

The ribbon number of a knot is the minimum number of ribbon singularities among all ribbon disks bounded by that knot. In this paper, we build on the systematic treatment of this knot invariant initiated in recent work of Friedl, Misev, and Zupan. We show that the set of Alexander polynomials of knots with ribbon number at most four contains 56 polynomials, and we use this set to compute the ribbon numbers for many 12-crossing knots. We also study higher-genus ribbon numbers of knots, presenting some examples that exhibit interesting behavior and establishing that the success of the Alexander polynomial at controlling genus-0 ribbon numbers does not extend to higher genera., Comment: 32 pages, 12 figures, 7 tables

Flavor development in the food industry is increasingly challenged by the need for rapid innovation and precise flavor profile creation. Traditional flavor research methods typically rely on iterative, subjective testing, which lacks the efficiency and scalability required for modern demands. This paper presents three contributions to address the challenges. Firstly, we define a new problem domain for scientific agents in flavor science, conceptualized as the generation of hypotheses for flavor profile sourcing and understanding. To facilitate research in this area, we introduce the FoodPuzzle, a challenging benchmark consisting of 978 food items and 1,766 flavor molecules profiles. We propose a novel Scientific Agent approach, integrating in-context learning and retrieval augmented techniques to generate grounded hypotheses in the domain of food science. Experimental results indicate that our model significantly surpasses traditional methods in flavor profile prediction tasks, demonstrating its potential to transform flavor development practices.

In this work, we significantly enhance masked particle modeling (MPM), a self-supervised learning scheme for constructing highly expressive representations of unordered sets relevant to developing foundation models for high-energy physics. In MPM, a model is trained to recover the missing elements of a set, a learning objective that requires no labels and can be applied directly to experimental data. We achieve significant performance improvements over previous work on MPM by addressing inefficiencies in the implementation and incorporating a more powerful decoder. We compare several pre-training tasks and introduce new reconstruction methods that utilize conditional generative models without data tokenization or discretization. We show that these new methods outperform the tokenized learning objective from the original MPM on a new test bed for foundation models for jets, which includes using a wide variety of downstream tasks relevant to jet physics, such as classification, secondary vertex finding, and track identification.

Quantum memories could benefit many devices in quantum information processing. For a quantum to be useful in real-world applications, the quantum memory must have a high recall efficiency. Here we demonstrate an efficient (up to 80%) spin-storage quantum memory in Er:YSO, using the 4-level rephased amplified spontaneous emission protocol. We show non-classical correlations between the ASE and RASE fields produced by the quantum memory. Also, we demonstrate the storage of 70 temporal modes, with a write time of 150 us, and a storage time of 25 us. Finally, a clear pathway is presented to improve the efficiency, storage time, and mode capacity. Such a device would have applications in quantum networking and measurement-based quantum computing., Comment: 9 pages, 10 figures

A De Bruijn cycle is a cyclic sequence in which every word of length $n$ over an alphabet $\mathcal{A}$ appears exactly once. De Bruijn tori are a two-dimensional analogue. Motivated by recent progress on universal partial cycles and words, which shorten De Bruijn cycles using a wildcard character, we introduce universal partial tori and matrices. We find them computationally and construct infinitely many of them using one-dimensional variants of universal cycles, including a new variant called a universal partial family., Comment: 22 pages, 1 figure

The slow microstructural evolution of materials often plays a key role in determining material properties. When the unit steps of the evolution process are slow, direct simulation approaches such as molecular dynamics become prohibitive and Kinetic Monte-Carlo (kMC) algorithms, where the state-to-state evolution of the system is represented in terms of a continuous-time Markov chain, are instead frequently relied upon to efficiently predict long-time evolution. The accuracy of kMC simulations however relies on the complete and accurate knowledge of reaction pathways and corresponding kinetics. This requirement becomes extremely stringent in complex systems such as concentrated alloys where the astronomical number of local atomic configurations makes the a priori tabulation of all possible transitions impractical. Machine learning models of transition kinetics have been used to mitigate this problem by enabling the efficient on-the-fly prediction of kinetic parameters. In this study, we show how physics-informed ML architectures can exactly enforce the detailed balance condition, by construction. Using the diffusion of a vacancy in a concentrated alloy as an example, we show that such ML architectures also exhibit superior performance in terms of prediction accuracy, demonstrating that the imposition of physical constraints can facilitate the accurate learning of barriers at no increase in computational cost., Comment: 20 pages, 11 figures

Studying the orbital motion of stars around Sagittarius A* in the Galactic Center provides a unique opportunity to probe the gravitational potential near the supermassive black hole at the heart of our Galaxy. Interferometric data obtained with the GRAVITY instrument at the Very Large Telescope Interferometer (VLTI) since 2016 has allowed us to achieve unprecedented precision in tracking the orbits of these stars. GRAVITY data have been key to detecting the in-plane, prograde Schwarzschild precession of the orbit of the star S2, as predicted by General Relativity. By combining astrometric and spectroscopic data from multiple stars, including S2, S29, S38, and S55 - for which we have data around their time of pericenter passage with GRAVITY - we can now strengthen the significance of this detection to an approximately $10 \sigma$ confidence level. The prograde precession of S2's orbit provides valuable insights into the potential presence of an extended mass distribution surrounding Sagittarius A*, which could consist of a dynamically relaxed stellar cusp comprised of old stars and stellar remnants, along with a possible dark matter spike. Our analysis, based on two plausible density profiles - a power-law and a Plummer profile - constrains the enclosed mass within the orbit of S2 to be consistent with zero, establishing an upper limit of approximately $1200 \, M_\odot$ with a $1 \sigma$ confidence level. This significantly improves our constraints on the mass distribution in the Galactic Center. Our upper limit is very close to the expected value from numerical simulations for a stellar cusp in the Galactic Center, leaving little room for a significant enhancement of dark matter density near Sagittarius A*., Comment: Submitted to A&A on September 17, 2024

We present one of the largest multiwavelength studies of simultaneous optical-to-X-ray spectral energy distributions (SEDs) of unobscured active galactic nuclei (AGN) in the local Universe. Using a representative sample of hard-X-ray-selected AGN from the 70-month Swift/BAT catalog, with optical/UV photometric data from Swift/UVOT and X-ray spectral data from Swift/XRT, we constructed broadband SEDs of 236 nearby AGN (0.001

We present a velocity-resolved reverberation mapping analysis of the hypervariable quasar RM160 (SDSS J141041.25+531849.0) at z = 0.359 with 153 spectroscopic epochs of data representing a ten-year baseline (2013-2023). We split the baseline into two regimes based on the 3x flux increase in the light curve: a 'low state' phase during the years 2013-2019 and a 'high state' phase during the years 2022-2023. The velocity-resolved lag profiles (VRLP) indicate that gas with different kinematics dominates the line emission in different states. The H\b{eta} VRLP begins with a signature of inflow onto the BLR in the 'low state', while in the 'high state' it is flatter with less signature of inflow. The H{\alpha} VRLP begins consistent with a virialized BLR in the 'low state', while in the 'high state' shows a signature of inflow. The differences in the kinematics between the Balmer lines and between the 'low state' and the 'high state' suggests complex BLR dynamics. We find that the BLR radius and velocity (both FWHM and {\sigma}) do not obey a constant virial product throughout the monitoring period. We find that BLR lags and continuum luminosity are correlated, consistent with rapid response of the BLR gas to the illuminating continuum. The BLR kinematic profile changes in unpredictable ways that are not related to continuum changes and reverberation lag. Our observations indicate that non-virial kinematics can significantly contribute to observed line profiles, suggesting caution for black-hole mass estimation in luminous and highly varying quasars like RM160., Comment: 23 pages, 17 figures

One of the fundamental questions of cosmology is the origin and mechanism(s) responsible for the reionization of the Universe beyond $z\sim6$. To address this question, many studies over the past decade have focused on local ($z\sim0.3$) galaxies which leak ionizing radiation (Lyman continuum or LyC). However, line-of-sight effects and data quality have prohibited deeper insight into the nature of LyC escape. To circumvent these limitations, we analyze stacks of a consolidated sample of {\it HST}/COS observations of the LyC in 89 galaxies at $z\sim0.3$. From fitting of the continuum, we obtain information about the underlying stellar populations and neutral ISM geometry. We find that most LyC non-detections are not leaking appreciable LyC ($f_{esc}^{\rm LyC}<1$\%) but also that exceptional cases point to spatial variations in the LyC escape fraction $f_{esc}^{\rm LyC}$. Stellar populations younger than 3 Myr lead to an increase in ionizing feedback, which in turn increases the isotropy of LyC escape. Moreover, mechanical feedback from supernovae in 8-10 Myr stellar populations is important for anisotropic gas distributions needed for LyC escape. While mechanical feedback is necessary for any LyC escape, high $f_{esc}^{\rm LyC}$ ($>5$\%) also requires a confluence of young stars and ionizing feedback. A two-stage burst of star formation could facilitate this optimal LyC escape scenario., Comment: Submitted to AAS publications, 42 pages, 25 figures

Spectral distortions of the cosmic microwave background (CMB) provide stringent constraints on energy and entropy production in the post-BBN (Big Bang Nucleosynthesis) era. This has been used to constrain dark photon models with COBE/FIRAS and forecast the potential gains with future CMB spectrometers. Here, we revisit these constraints by carefully considering the photon to dark photon conversion process and evolution of the distortion signal. Previous works only included the effect of CMB energy density changes but neglected the change to the photon number density. We clearly define the dark photon distortion signal and show that in contrast to previous analytic estimates the distortion has an opposite sign and a $\simeq 1.5$ times larger amplitude. We furthermore extend the treatment into the large distortion regime to also cover the redshift range $\simeq 2\times 10^6-4\times 10^7$ between the $\mu$-era and the end of BBN using CosmoTherm. This shows that the CMB distortion constraints for dark photon masses in the range $10^{-4}\,{\rm eV}\lesssim m_{\rm dp}\lesssim 10^{-3}\,{\rm eV}$ were significantly underestimated. We demonstrate that in the small distortion regime the distortion caused by photon to dark photon conversion is extremely close to a $\mu$-type distortion independent of the conversion redshift. This opens the possibility to study dark photon models using CMB distortion anisotropies and the correlations with CMB temperature anisotropies as we highlight here., Comment: 13 pages, 10+2 figures, to be submitted to MNRAS, comments welcome

Water quality mapping for critical parameters such as temperature, salinity, and turbidity is crucial for assessing an aquaculture farm's health and yield capacity. Traditional approaches involve using boats or human divers, which are time-constrained and lack depth variability. This work presents an innovative approach to 3D water quality mapping in shallow water environments using a BlueROV2 equipped with GPS and a water quality sensor. This system allows for accurate location correction by resurfacing when errors occur. This study is being conducted at an oyster farm in the Chesapeake Bay, USA, providing a more comprehensive and precise water quality analysis in aquaculture settings., Comment: IEEE IROS workshop on Autonomous Robotic Systems in Aquaculture: Research Challenges and Industry Needs

This work presents a novel rehabilitation framework designed for a therapist, wearing an inertial measurement unit (IMU) suit, to virtually interact with a lower-limb exoskeleton worn by a patient with motor impairments. This framework aims to harmonize the skills and knowledge of the therapist with the capabilities of the exoskeleton. The therapist can guide the patient's movements by moving their own joints and making real-time adjustments to meet the patient's needs, while reducing the physical effort of the therapist. This eliminates the need for a predefined trajectory for the patient to follow, as in conventional robotic gait training. For the virtual interaction medium between the therapist and patient, we propose an impedance profile that is stiff at low frequencies and less stiff at high frequencies, that can be tailored to individual patient needs and different stages of rehabilitation. The desired interaction torque from this medium is commanded to a whole-exoskeleton closed-loop compensation controller. The proposed virtual interaction framework was evaluated with a pair of unimpaired individuals in different teacher-student gait training exercises. Results show the proposed interaction control effectively transmits haptic cues, informing future applications in rehabilitation scenarios.

We study spherical accretion of magnetized plasma with low angular momentum onto a supermassive black hole, utilizing global General Relativistic Magnetohydrodynamic simulations. Black hole-driven feedback in the form of magnetic eruptions and jets triggers magnetized turbulence in the surrounding medium. We find that when the Bondi radius exceeds a certain value relative to the black hole's gravitational radius, this turbulence restricts the subsequent inflow of magnetic flux, strongly suppressing the strength of the jet. Consequently, magnetically arrested disks and powerful jets are not a generic outcome of accretion of magnetized plasma, even if there is an abundance of magnetic flux available in the system. However, if there is significant angular momentum in the inflowing gas, the eruption-driven turbulence is suppressed (sheared out), allowing for the presence of a powerful jet. Both the initially rotating and non-rotating flows go through periods of low and high gas angular momentum, showing that the angular momentum content of the inflowing gas is not just a feature of the ambient medium, but is strongly modified by the eruption and jet-driven black hole feedback. In the lower angular momentum states, our results predict that there should be dynamically strong magnetic fields on horizon scales, but no powerful jet; this state may be consistent with Sgr A* in the Galactic Center., Comment: 17 pages, 13 figures, submitted to ApJ

Background: Advanced methods for causal inference, such as targeted maximum likelihood estimation (TMLE), require certain conditions for statistical inference. However, in situations where there is not differentiability due to data sparsity or near-positivity violations, the Donsker class condition is violated. In such situations, TMLE variance can suffer from inflation of the type I error and poor coverage, leading to conservative confidence intervals. Cross-validation of the TMLE algorithm (CVTMLE) has been suggested to improve on performance compared to TMLE in settings of positivity or Donsker class violations. We aim to investigate the performance of CVTMLE compared to TMLE in various settings. Methods: We utilised the data-generating mechanism as described in Leger et al. (2022) to run a Monte Carlo experiment under different Donsker class violations. Then, we evaluated the respective statistical performances of TMLE and CVTMLE with different super learner libraries, with and without regression tree methods. Results: We found that CVTMLE vastly improves confidence interval coverage without adversely affecting bias, particularly in settings with small sample sizes and near-positivity violations. Furthermore, incorporating regression trees using standard TMLE with ensemble super learner-based initial estimates increases bias and variance leading to invalid statistical inference. Conclusions: It has been shown that when using CVTMLE the Donsker class condition is no longer necessary to obtain valid statistical inference when using regression trees and under either data sparsity or near-positivity violations. We show through simulations that CVTMLE is much less sensitive to the choice of the super learner library and thereby provides better estimation and inference in cases where the super learner library uses more flexible candidates and is prone to overfitting., Comment: 20 pages, 3 figures, 1 table

The development of a system vulnerability analysis tool (SVAT) for complex mission critical systems (CMCS) produced the software for operation and network attack results review (SONARR). This software builds upon the Blackboard Architecture and uses its a rule-fact logic to assess model networks to identify potential pathways that an attacker might take through them via the exploitation of vulnerabilities within the network. The SONARR objects and algorithm were developed previously; however, performance was insufficient for analyzing large networks. This paper describes and analyzes the performance of a multi-threaded SONARR algorithm and other enhancements which were developed to increase SONARR's performance and facilitate the analysis of large networks., Comment: The U.S. federal sponsor has requested that we not include funding acknowledgement for this publication

A system vulnerability analysis technique (SVAT) for the analysis of complex mission critical systems (CMCS) that cannot be taken offline or subjected to the risks posed by traditional penetration testing was previously developed. This system uses path-based analysis of vulnerabilities to identify potential threats to system security. Generalization logic building on the Blackboard Architecture's rule-fact paradigm was implemented in this system, the software for operation and network attack results review (SONARR). This paper presents an overview of additional functionality that has been added to this tool and the experimentation that was conducted to analyze their efficacy and the performance benefits of the new in-memory processing capabilities of the SONARR algorithm. The results of the performance tests and their relation to networks' architecture are discussed. The paper concludes with a discussion of avenues of future work, including the implementation of multithreading, additional analysis metrics like confidentiality, integrity, and availability, and improved heuristic development., Comment: The U.S. federal sponsor has requested that we not include funding acknowledgement for this publication

Tunable biphoton quantum entanglement generated from nonlinear processes is highly desirable for cutting-edge quantum technologies, yet its tunability is substantially constrained by the symmetry of material nonlinear tensors. Here, we overcome this constraint by introducing symmetry-breaking in nonlinear polarization to generate optically tunable biphoton entanglement at picosecond speeds. Asymmetric optical responses have made breakthroughs in classical applications like non-reciprocal light transmission. We now experimentally demonstrate the nonlinear asymmetry response for biphoton entanglement using a semiconductor metasurface incorporating [110] InGaP nano-resonators with structural asymmetry. We realize continuous tuning of polarization entanglement from near-unentangled states to a Bell state. This tunability can also extend to produce tailored hyperentanglement. Furthermore, our nanoscale entanglement source features an ultra-high coincidence-to-accidental ratio of $\approx7\times10^4$, outperforming existing semiconductor flat optics by two orders of magnitude. Introducing asymmetric nonlinear response in quantum metasurfaces opens new directions for tailoring on-demand quantum states and beyond., Comment: 27 pages, 4 figures

HPC systems used for research run a wide variety of software and workflows. This software is often written or modified by users to meet the needs of their research projects, and rarely is built with security in mind. In this paper we explore several of the key techniques that MIT Lincoln Laboratory Supercomputing Center has deployed on its systems to manage the security implications of these workflows by providing enforced separation for processes, filesystem access, network traffic, and accelerators to make every user feel like they are running on a personal HPC.

This work introduces a framework to diagnose the strengths and shortcomings of Autonomous Vehicle (AV) collision avoidance technology with synthetic yet realistic potential collision scenarios adapted from real-world, collision-free data. Our framework generates counterfactual collisions with diverse crash properties, e.g., crash angle and velocity, between an adversary and a target vehicle by adding perturbations to the adversary's predicted trajectory from a learned AV behavior model. Our main contribution is to ground these adversarial perturbations in realistic behavior as defined through the lens of data-alignment in the behavior model's parameter space. Then, we cluster these synthetic counterfactuals to identify plausible and representative collision scenarios to form the basis of a test suite for downstream AV system evaluation. We demonstrate our framework using two state-of-the-art behavior prediction models as sources of realistic adversarial perturbations, and show that our scenario clustering evokes interpretable failure modes from a baseline AV policy under evaluation.

The hydrogen Lyman-alpha (Lya) emission line, the brightest spectral feature of a photoionized gas, is considered an indirect tracer of the escape of Lyman continuum (LyC) photons, particularly when the intergalactic medium is too opaque for direct detection. However, resonant scattering complicates interpreting the empirical properties of Lya photons, necessitating radiative transfer simulations to capture their strong coupling with underlying gas kinematics. In this study, we leverage the exceptional spatial resolution from strong gravitational lensing to investigate the connection between Lya line profiles and LyC leakage on scales of a few 100 pc in the Sunburst Arc galaxy at $z\sim2.37$. New optical echelle spectra obtained using Magellan MIKE show that both the LyC leaking and non-leaking regions exhibit a classic double-peak Lya feature with an enhanced red peak, indicating outflows at multiple locations in the galaxy. Both regions also show a central Gaussian peak atop the double peaks, indicating directly escaped Lya photons independent of LyC leakage. We introduce a machine learning-based method for emulating Lya simulations to quantify intrinsic dynamics ($\sigma_{\mathrm{int}}$), neutral hydrogen column density ($N_{\mathrm{HI}}$), outflow velocity ($v_{\mathrm{exp}}$), and effective temperature ($T$) across continuous parameter spaces. By comparing the spatially and spectrally resolved Lya lines in Sunburst, we argue that the directly escaped Lya photons originate in a volume-filling, warm ionized medium spanning $\sim1$ kpc, while the LyC leakage is confined to regions of $\lesssim200$ pc. These sub-kpc variations in Lya profiles highlight the complexity of interpreting integrated properties in the presence of inhomogeneous mixtures of gas and young stars, emphasizing the need for spatially and spectrally resolved observations of distant galaxies., Comment: 19 pages, 7 figures, 8 tables; added full names of authors

The resonant conversion of cosmic microwave background (CMB) photons into axions within large-scale structure induces an anisotropic spectral distortion in CMB temperature maps. Applying state-of-the-art foreground cleaning techniques to $\textit{Planck}$ CMB observations, we construct maps of axion-induced "patchy screening" of the CMB. We cross-correlate these maps with data from the $\textit{unWISE}$ galaxy survey and find no evidence of axions. We constrain the axion-photon coupling, $g_{a\gamma\gamma} \lesssim 2 \times 10^{-12}~{\rm GeV}^{-1}$, at the 95% confidence level for axion masses in the range $10^{-13}~{\rm eV} \lesssim m_a \lesssim 10^{-12}~{\rm eV}$. These constraints are competitive with the tightest astrophysical axion limits in this mass range and are inferred from robust population-level statistics, which makes them complementary to existing searches that rely on modeling of individual systems., Comment: 5+15 pages; 3+15 figures

We establish a log-Sobolev inequality for the stationary distribution of mean-field Langevin dynamics with a constant that is independent of the number of particles $N$. Our proof proceeds by establishing the existence of a Lipschitz transport map from the standard Gaussian measure via the reverse heat flow of Kim and Milman., Comment: 14 pages

This paper presents uniGasFoam, an open-source particle-based solver for multiscale rarefied gas flow simulations, which has been developed within the well-established OpenFOAM framework, and is an extension of the direct simulation Monte Carlo (DSMC) solver dsmcFoam+. The developed solver addresses the coupling challenges inherent in hybrid continuum-particle methods, originating from the disparate nature of finite-volume (FV) solvers found in computational fluid dynamics (CFD) software and DSMC particle solvers. This is achieved by employing alternative stochastic particle methods, resembling DSMC, to tackle the continuum limit. The uniGasFoam particle-particle coupling produces a numerical implementation that is simpler and more robust, faster in many steady-state flows, and more scalable for transient flows compared to conventional continuum-particle coupling. The presented framework is unified and generic, and can couple DSMC with stochastic particle (SP) and unified stochastic particle (USP) methods, or be employed for pure DSMC, SP, and USP gas simulations. To enhance user experience, optimise computational resources and minimise user error, advanced adaptive algorithms such as transient adaptive sub-cells, non-uniform cell weighting, and adaptive global time stepping have been integrated into uniGasFoam. In this paper, the hybrid USP-DSMC module of uniGasFoam is rigorously validated through multiple benchmark cases, consistently showing excellent agreement with pure DSMC, hybrid CFD-DSMC, and literature results. Notably, uniGasFoam achieves significant computational gains compared to pure dsmcFoam+ simulations, rendering it a robust computational tool well-suited for addressing multiscale rarefied gas flows of engineering importance.

Although existing medical image segmentation methods provide impressive pixel-wise accuracy, they often neglect topological correctness, making their segmentations unusable for many downstream tasks. One option is to retrain such models whilst including a topology-driven loss component. However, this is computationally expensive and often impractical. A better solution would be to have a versatile plug-and-play topology refinement method that is compatible with any domain-specific segmentation pipeline. Directly training a post-processing model to mitigate topological errors often fails as such models tend to be biased towards the topological errors of a target segmentation network. The diversity of these errors is confined to the information provided by a labelled training set, which is especially problematic for small datasets. Our method solves this problem by training a model-agnostic topology refinement network with synthetic segmentations that cover a wide variety of topological errors. Inspired by the Stone-Weierstrass theorem, we synthesize topology-perturbation masks with randomly sampled coefficients of orthogonal polynomial bases, which ensures a complete and unbiased representation. Practically, we verified the efficiency and effectiveness of our methods as being compatible with multiple families of polynomial bases, and show evidence that our universal plug-and-play topology refinement network outperforms both existing topology-driven learning-based and post-processing methods. We also show that combining our method with learning-based models provides an effortless add-on, which can further improve the performance of existing approaches., Comment: Accepted by the 27th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2024)

The mechanism of X-ray outbursts in Be X-ray binaries remains a mystery, and understanding their circumstellar disks is crucial for a solution of the mass-transfer problem. In particular, it is important to identify the Be star activities (e.g., pulsations) that cause mass ejection and, hence, disk formation. Therefore, we investigated the relationship between optical flux oscillations and the infrared (IR) excess in a sample of five Be X-ray binaries. Applying the Lomb-Scargle technique to high-cadence optical light curves from the Transiting Exoplanet Survey Satellite (TESS), we detected several significant oscillation modes in the 3 to 24 hour period range for each source. We also measured the IR excess (a proxy for disk growth) of those five sources, using J-band light curves from Palomar Gattini-IR. In four of the five sources, we found anti-correlations between the IR excess and the amplitude of the main flux oscillation modes. This result is inconsistent with the conventional idea that non-radial pulsations drive mass ejections. We propose an alternative scenario where internal temperature variations in the Be star cause transitions between pulsation-active and mass-ejection-active states., Comment: 17 pages, 27 figures, 6 tables, accepted for publication in MNRAS

We propose a novel approach to enable the use of large, single speaker ASR models, such as Whisper, for target speaker ASR. The key insight of this method is that it is much easier to model relative differences among speakers by learning to condition on frame-level diarization outputs, than to learn the space of all speaker embeddings. We find that adding even a single bias term per diarization output type before the first transformer block can transform single speaker ASR models, into target speaker ASR models. Our target-speaker ASR model can be used for speaker attributed ASR by producing, in sequence, a transcript for each hypothesized speaker in a diarization output. This simplified model for speaker attributed ASR using only a single microphone outperforms cascades of speech separation and diarization by 11% absolute ORC-WER on the NOTSOFAR-1 dataset., Comment: Submitted to ICASSP 2025

To maximize the amount of information extracted from cosmological datasets, simulations that accurately represent these observations are necessary. However, traditional simulations that evolve particles under gravity by estimating particle-particle interactions (N-body simulations) are computationally expensive and prohibitive to scale to the large volumes and resolutions necessary for the upcoming datasets. Moreover, modeling the distribution of galaxies typically involves identifying virialized dark matter halos, which is also a time- and memory-consuming process for large N-body simulations, further exacerbating the computational cost. In this study, we introduce CHARM, a novel method for creating mock halo catalogs by matching the spatial, mass, and velocity statistics of halos directly from the large-scale distribution of the dark matter density field. We develop multi-stage neural spline flow-based networks to learn this mapping at redshift z=0.5 directly with computationally cheaper low-resolution particle mesh simulations instead of relying on the high-resolution N-body simulations. We show that the mock halo catalogs and painted galaxy catalogs have the same statistical properties as obtained from $N$-body simulations in both real space and redshift space. Finally, we use these mock catalogs for cosmological inference using redshift-space galaxy power spectrum, bispectrum, and wavelet-based statistics using simulation-based inference, performing the first inference with accelerated forward model simulations and finding unbiased cosmological constraints with well-calibrated posteriors. The code was developed as part of the Simons Collaboration on Learning the Universe and is publicly available at \url{https://github.com/shivampcosmo/CHARM}., Comment: 12 pages and 8 figures. This is a Learning the Universe Publication

Traditional star formation subgrid models implemented in cosmological galaxy formation simulations, such as that of Springel & Hernquist (2003, hereafter SH03), employ adjustable parameters to satisfy constraints measured in the local Universe. In recent years, however, theory and spatially-resolved simulations of the turbulent, multiphase, star-forming ISM have begun to produce new first-principles models, which when fully developed can replace traditional subgrid prescriptions. This approach has advantages of being physically motivated and predictive rather than empirically tuned, and allowing for varying environmental conditions rather than being tied to local Universe conditions. As a prototype of this new approach, by combining calibrations from the TIGRESS numerical framework with the Pressure-Regulated Feedback-Modulated (PRFM) theory, simple formulae can be obtained for both the gas depletion time and an effective equation of state. Considering galaxies in TNG50, we compare the "native" simulation outputs with post-processed predictions from PRFM. At TNG50 resolution, the total midplane pressure is nearly equal to the total ISM weight, indicating that galaxies in TNG50 are close to satisfying vertical equilibrium. The measured gas scale height is also close to theoretical equilibrium predictions. The slopes of the effective equations of states are similar, but with effective velocity dispersion normalization from SH03 slightly larger than that from current TIGRESS simulations. Because of this and the decrease in PRFM feedback yield at high pressure, the PRFM model predicts shorter gas depletion times than the SH03 model at high densities and redshift. Our results represent a first step towards implementing new, numerically calibrated subgrid algorithms in cosmological galaxy formation simulations., Comment: 33 pages, 12 figures, Accepted for publication in ApJ. This is a Learning the Universe Publication. All codes and data used to produce this work can be found at the following $\href{https://github.com/sultan-hassan/tng50-post-processing-prfm}{GitHub \,Link.}$

Upcoming wide field surveys such as the Rubin Observatory's Legacy Survey of Space and Time (LSST) will monitor thousands of strongly lensed quasars over a 10-year period. Many of these monitored quasars will undergo high magnification events (HMEs) through microlensing as the accretion disk crosses a caustic, places of infinite magnification. Microlensing allows us to map the inner regions of the accretion disk as it crosses a caustic, even at large cosmological distances. The observational cadences of LSST are not ideal for probing the inner regions of the accretion disk, so there is a need to predict HMEs as early as possible to trigger high-cadence multi-band or spectroscopic follow-up observations. Here we simulate a diverse and realistic sample of 10-year quasar microlensing light curves to train a recurrent neural network (RNN) to predict HMEs before they occur by classifying the location of the peaks at each time step. This is the first deep learning approach to predict HMEs. We give estimates at how well we expect to predict HME peaks during LSST and benchmark how our metrics change with different cadence strategies. With LSST-like observations, we can predict approximately 55% of HME peaks corresponding to tens to hundreds per year and a false positive rate of around 20% compared to the number of HMEs. Our network can be continuously applied throughout the LSST survey, providing crucial alerts to optimize follow-up resources., Comment: 16 pages, 9 figures, submitted to ApJ

We present a number of systems for the Voice Privacy Challenge, including voice conversion based systems such as the kNN-VC method and the WavLM voice Conversion method, and text-to-speech (TTS) based systems including Whisper-VITS. We found that while voice conversion systems better preserve emotional content, they struggle to conceal speaker identity in semi-white-box attack scenarios; conversely, TTS methods perform better at anonymization and worse at emotion preservation. Finally, we propose a random admixture system which seeks to balance out the strengths and weaknesses of the two category of systems, achieving a strong EER of over 40% while maintaining UAR at a respectable 47%., Comment: Submission to the Voice Privacy Challenge 2024. Accepted and presented at the 4th Symposium on Security and Privacy in Speech Communication

We present Benchpress, a benchmarking suite for evaluating the performance and range of functionality of multiple quantum computing software development kits. This suite consists of a collection of over $1000$ tests measuring key performance metrics for a wide variety of operations on quantum circuits comprised of up to $930$ qubits and $\mathcal{O}(10^{6})$ two-qubit gates, as well as an execution framework for running the tests over multiple quantum software packages in a unified manner. We give a detailed overview of the benchmark suite, its methodology, and generate representative results over seven different quantum software packages. The flexibility of the Benchpress framework allows for benchmarking that not only keeps pace with quantum hardware improvements but can preemptively gauge the quantum circuit processing costs of future device architectures. Being open-source, Benchpress ensures the transparency and verification of performance claims., Comment: 11 pages, 4 figures, 2 tables

The growing demand for machine learning in healthcare requires processing increasingly large electronic health record (EHR) datasets, but existing pipelines are not computationally efficient or scalable. In this paper, we introduce meds_reader, an optimized Python package for efficient EHR data processing that is designed to take advantage of many intrinsic properties of EHR data for improved speed. We then demonstrate the benefits of meds_reader by reimplementing key components of two major EHR processing pipelines, achieving 10-100x improvements in memory, speed, and disk usage. The code for meds_reader can be found at https://github.com/som-shahlab/meds_reader.

The tuning of the Fermi level in tin telluride, a topological crystalline insulator, is essential for accessing its unique surface states and optimizing its electronic properties for applications such as spintronics and quantum computing. In this study, we demonstrate that the Fermi level in tin telluride can be effectively modulated by controlling the tin concentration during chemical vapor deposition synthesis. By introducing tin-rich conditions, we observed a blue shift in the X-ray photoelectron spectroscopy core-level peaks of both tin and tellurium, indicating an upward shift in the Fermi level. This shift is corroborated by a decrease in work function values measured via ultraviolet photoelectron spectroscopy, confirming the suppression of Sn vacancies. Our findings provide a low-cost, scalable method to achieve tunable Fermi levels in tin telluride, offering a significant advancement in the development of materials with tailored electronic properties for next-generation technological applications.

Anomaly detection is an important problem in many application areas, such as network security. Many deep learning methods for unsupervised anomaly detection produce good empirical performance but lack theoretical guarantees. By casting anomaly detection into a binary classification problem, we establish non-asymptotic upper bounds and a convergence rate on the excess risk on rectified linear unit (ReLU) neural networks trained on synthetic anomalies. Our convergence rate on the excess risk matches the minimax optimal rate in the literature. Furthermore, we provide lower and upper bounds on the number of synthetic anomalies that can attain this optimality. For practical implementation, we relax some conditions to improve the search for the empirical risk minimizer, which leads to competitive performance to other classification-based methods for anomaly detection. Overall, our work provides the first theoretical guarantees of unsupervised neural network-based anomaly detectors and empirical insights on how to design them well.

We conduct a systematic search for astrophysical transients using data from the Atacama Cosmology Telescope (ACT). The data were taken from 2017 to 2022 in three frequency bands spanning 77 GHz to 277 GHz. In this paper we present a pipeline for transient detection using single observation maps where each pixel of a map contains one observation with an integration time of approximately four minutes. We find 34 transient events at 27 unique locations. All but two of the transients are associated with Galactic stars and exhibit a wide range of properties. We also detect an event coincident with the classical nova, YZ Ret and one event consistent with a flaring active galactic nucleus. We notably do not detect any reverse shock emission from gamma ray bursts, a non-detection which is in tension with current models., Comment: 23 pages, 10 figures, 10 tables. First and second author share equal contributions. Article and accompanying data submitted to ApJ. Data tables will be made available upon publication

We introduce LightSABRE, a significant enhancement of the SABRE algorithm that advances both runtime efficiency and circuit quality. LightSABRE addresses the increasing demands of modern quantum hardware, which can now accommodate complex scenarios, and circuits with millions of gates. Through iterative development within Qiskit, primarily using the Rust programming language, we have achieved a version of the algorithm in Qiskit 1.2.0 that is approximately 200 times faster than the implementation in Qiskit 0.20.1, which already introduced key improvements like the release valve mechanism. Additionally, when compared to the SABRE algorithm presented in Li et al., LightSABRE delivers an average decrease of 18.9\% in SWAP gate count across the same benchmark circuits. Unlike SABRE, which struggles with scalability and convergence on large circuits, LightSABRE delivers consistently high-quality routing solutions, enabling the efficient execution of large quantum circuits on near-term and future quantum devices. LightSABRE's improvements in speed, scalability, and quality position it as a critical tool for optimizing quantum circuits in the context of evolving quantum hardware and error correction techniques., Comment: 10 pages, 8 figures

Convex programming with linear constraints plays an important role in the operation of a number of everyday systems. However, absent any additional protections, revealing or acting on the solutions to such problems may reveal information about their constraints, which can be sensitive. Therefore, in this paper, we introduce a method for solving convex programs while keeping linear constraints private. First, we prove that this method is differentially private and always generates a feasible optimization problem (i.e., one whose solution exists). Then we show that the solution to the privatized problem also satisfies the original, non-private constraints. Next, we bound the expected loss in performance from privacy, which is measured by comparing the cost with privacy to that without privacy. Simulation results apply this framework to constrained policy synthesis in a Markov decision process, and they show that a typical privacy implementation induces only an approximately $9\%$ loss in solution quality., Comment: 2 figures, 6 pages

We study the communication complexity of truthful combinatorial auctions, and in particular the case where valuations are either subadditive or single-minded, which we denote with $\mathsf{SubAdd}\cup\mathsf{SingleM}$. We show that for three bidders with valuations in $\mathsf{SubAdd}\cup\mathsf{SingleM}$, any deterministic truthful mechanism that achieves at least a $0.366$-approximation requires $\exp(m)$ communication. In contrast, a natural extension of [Fei09] yields a non-truthful $\mathrm{poly}(m)$-communication protocol that achieves a $\frac{1}{2}$-approximation, demonstrating a gap between the power of truthful mechanisms and non-truthful protocols for this problem. Our approach follows the taxation complexity framework laid out in [Dob16b], but applies this framework in a setting not encompassed by the techniques used in past work. In particular, the only successful prior application of this framework uses a reduction to simultaneous protocols which only applies for two bidders [AKSW20], whereas our three-player lower bounds are stronger than what can possibly arise from a two-player construction (since a trivial truthful auction guarantees a $\frac{1}{2}$-approximation for two players).

The MIT/IEEE/Amazon GraphChallenge encourages community approaches to developing new solutions for analyzing graphs and sparse data derived from social media, sensor feeds, and scientific data to discover relationships between events as they unfold in the field. The anonymized network sensing Graph Challenge seeks to enable large, open, community-based approaches to protecting networks. Many large-scale networking problems can only be solved with community access to very broad data sets with the highest regard for privacy and strong community buy-in. Such approaches often require community-based data sharing. In the broader networking community (commercial, federal, and academia) anonymized source-to-destination traffic matrices with standard data sharing agreements have emerged as a data product that can meet many of these requirements. This challenge provides an opportunity to highlight novel approaches for optimizing the construction and analysis of anonymized traffic matrices using over 100 billion network packets derived from the largest Internet telescope in the world (CAIDA). This challenge specifies the anonymization, construction, and analysis of these traffic matrices. A GraphBLAS reference implementation is provided, but the use of GraphBLAS is not required in this Graph Challenge. As with prior Graph Challenges the goal is to provide a well-defined context for demonstrating innovation. Graph Challenge participants are free to select (with accompanying explanation) the Graph Challenge elements that are appropriate for highlighting their innovations., Comment: Accepted to IEEE HPEC 2024

This paper builds one-cusped complex hyperbolic $2$-manifolds by an explicit geometric construction. Specifically, for each odd $d \ge 1$ there is a smooth projective surface $Z_d$ with $c_1^2(Z_d) = c_2(Z_d) = 6d$ and a smooth irreducible curve $E_d$ on $Z_d$ of genus one so that $Z_d \smallsetminus E_d$ admits a finite volume uniformization by the unit ball $\mathbb{B}^2$ in $\mathbb{C}^2$. This produces one-cusped complex hyperbolic $2$-manifolds of arbitrarily large volume. As a consequence, the $3$-dimensional nilmanifold of Euler number $12d$ bounds geometrically for all odd $d \ge 1$.

Machine-learned interatomic potentials (MLIPs) are becoming an essential tool in materials modeling. However, optimizing the generation of training data used to parameterize the MLIPs remains a significant challenge. This is because MLIPs can fail when encountering local enviroments too different from those present in the training data. The difficulty of determining \textit{a priori} the environments that will be encountered during molecular dynamics (MD) simulation necessitates diverse, high-quality training data. This study investigates how training data diversity affects the performance of MLIPs using the Ultra-Fast Force Field (UF$^3$) to model amorphous silicon nitride. We employ expert and autonomously generated data to create the training data and fit four force-field variants to subsets of the data. Our findings reveal a critical balance in training data diversity: insufficient diversity hinders generalization, while excessive diversity can exceed the MLIP's learning capacity, reducing simulation accuracy. Specifically, we found that the UF$^3$ variant trained on a subset of the training data, in which nitrogen-rich structures were removed, offered vastly better prediction and simulation accuracy than any other variant. By comparing these UF$^3$ variants, we highlight the nuanced requirements for creating accurate MLIPs, emphasizing the importance of application-specific training data to achieve optimal performance in modeling complex material behaviors., Comment: 6 pages, 4 figures

To date, galaxy image simulations for weak lensing surveys usually approximate the light profiles of all galaxies as a single or double S\'ersic profile, neglecting the influence of galaxy substructures and morphologies deviating from such a simplified parametric characterization. While this approximation may be sufficient for previous data sets, the stringent cosmic shear calibration requirements and the high quality of the data in the upcoming Euclid survey demand a consideration of the effects that realistic galaxy substructures have on shear measurement biases. Here we present a novel deep learning-based method to create such simulated galaxies directly from HST data. We first build and validate a convolutional neural network based on the wavelet scattering transform to learn noise-free representations independent of the point-spread function of HST galaxy images that can be injected into simulations of images from Euclid's optical instrument VIS without introducing noise correlations during PSF convolution or shearing. Then, we demonstrate the generation of new galaxy images by sampling from the model randomly and conditionally. Next, we quantify the cosmic shear bias from complex galaxy shapes in Euclid-like simulations by comparing the shear measurement biases between a sample of model objects and their best-fit double-S\'ersic counterparts. Using the KSB shape measurement algorithm, we find a multiplicative bias difference between these branches with realistic morphologies and parametric profiles on the order of $6.9\times 10^{-3}$ for a realistic magnitude-S\'ersic index distribution. Moreover, we find clear detection bias differences between full image scenes simulated with parametric and realistic galaxies, leading to a bias difference of $4.0\times 10^{-3}$ independent of the shape measurement method. This makes it relevant for stage IV weak lensing surveys such as Euclid., Comment: Submitted to A&A. 29 pages, 20 figures, 2 tables

Compact objects observed in gravitational-wave astronomy so far always come in pairs and never individually. Identifying the two components of a binary system is a delicate operation that is often taken for granted. The labeling procedure (i.e. which is object "1" and which is object "2") effectively acts as systematics, or equivalently an unspecified prior, in gravitational-wave data inference. The common approach is to label the objects solely by their masses, on a sample-by-sample basis; while intuitive, this leads to degeneracies when binaries have comparable masses. Instead, we argue that object identification should be tackled using the posterior distribution as a whole. We frame the problem in terms of constrained clustering -- a flavor of semi-supervised machine learning -- and find that unfolding the labeling systematics can significantly impact, and arguably improve, our interpretation of the data. In particular, the precision of black-hole spin measurements improves by up to 50%, spurious multimodalities and tails tend to disappear, posteriors become closer to Gaussian distributions, and the identification of the nature of the object (i.e. black hole vs. neutron star) is facilitated. We estimate that about 10% of the LIGO/Virgo posterior samples are affected by this relabeling, i.e. they might have been attributed to the wrong compact object in the observed binaries., Comment: 5 pages, 3 figures + appendices

A metagame is a collection of knowledge that goes beyond the rules of a game. In competitive, team-based games like Pok\'emon or League of Legends, it refers to the set of current dominant characters and/or strategies within the player base. Developer changes to the balance of the game can have drastic and unforeseen consequences on these sets of meta characters. A framework for predicting the impact of balance changes could aid developers in making more informed balance decisions. In this paper we present such a Meta Discovery framework, leveraging Reinforcement Learning for automated testing of balance changes. Our results demonstrate the ability to predict the outcome of balance changes in Pok\'emon Showdown, a collection of competitive Pok\'emon tiers, with high accuracy., Comment: 11 pages, 1 figure, IEEE Transactions on Games

Computed tomography (CT) reconstruction plays a crucial role in industrial nondestructive testing and medical diagnosis. Sparse view CT reconstruction aims to reconstruct high-quality CT images while only using a small number of projections, which helps to improve the detection speed of industrial assembly lines and is also meaningful for reducing radiation in medical scenarios. Sparse CT reconstruction methods based on implicit neural representations (INRs) have recently shown promising performance, but still produce artifacts because of the difficulty of obtaining useful prior information. In this work, we incorporate a powerful prior: the total number of material categories of objects. To utilize the prior, we design AC-IND, a self-supervised method based on Attenuation Coefficient Estimation and Implicit Neural Distribution. Specifically, our method first transforms the traditional INR from scalar mapping to probability distribution mapping. Then we design a compact attenuation coefficient estimator initialized with values from a rough reconstruction and fast segmentation. Finally, our algorithm finishes the CT reconstruction by jointly optimizing the estimator and the generated distribution. Through experiments, we find that our method not only outperforms the comparative methods in sparse CT reconstruction but also can automatically generate semantic segmentation maps., Comment: 12 pages

Understanding the rapid formation of supermassive black holes (SMBHs) in the early universe requires an understanding of how stellar mass grows in the host galaxies. Here, we perform an analysis of rest-frame optical spectra and imaging from JWST of two quasar host galaxies at z>6 which exhibit Balmer absorption lines. These features in the stellar continuum indicate a lack of young stars, similar to low-redshift post-starburst galaxies whose star formation was recently quenched. We find that the stellar mass (log(M_* / M_sun) > 10.6) of each quasar host grew in a starburst episode at redshift 7 or 8. One of the targets exhibits little ongoing star formation, as evidenced by the photometric signature of the Balmer break and a lack of spatially resolved H-alpha emission, placing it well below the star formation main sequence at z = 6. The other galaxy is transitioning to a quiescent phase; together, the two galaxies represent the most distant massive post-starburst galaxies known. The maturity of these two galaxies is further supported by the stellar velocity dispersions of their host galaxies, placing them slightly above the upper end of the local M_BH - sigma_* relation. The properties of our two post-starburst galaxies, each hosting an active SMBH with log(M_BH / M_sun) > 9, suggests that black holes played a major role in shaping the formation of the first massive galaxies in the Universe., Comment: 24 pages, 7 figures, submitted to a Nature journal

Novel view synthesis of indoor scenes can be achieved by capturing a monocular video sequence of the environment. However, redundant information caused by artificial movements in the input video data reduces the efficiency of scene modeling. In this work, we tackle this challenge from the perspective of camera selection. We begin by constructing a similarity matrix that incorporates both the spatial diversity of the cameras and the semantic variation of the images. Based on this matrix, we use the Intra-List Diversity (ILD) metric to assess camera redundancy, formulating the camera selection task as an optimization problem. Then we apply a diversity-based sampling algorithm to optimize the camera selection. We also develop a new dataset, IndoorTraj, which includes long and complex camera movements captured by humans in virtual indoor environments, closely mimicking real-world scenarios. Experimental results demonstrate that our strategy outperforms other approaches under time and memory constraints. Remarkably, our method achieves performance comparable to models trained on the full dataset, while using only an average of 15% of the frames and 75% of the allotted time.

In this paper we consider the problem of releasing a Gaussian Differentially Private (GDP) 3D human face. The human face is a complex structure with many features and inherently tied to one's identity. Protecting this data, in a formally private way, is important yet challenging given the dimensionality of the problem. We extend approximate DP techniques for functional data to the GDP framework. We further propose a novel representation, face radial curves, of a 3D face as a set of functions and then utilize our proposed GDP functional data mechanism. To preserve the shape of the face while injecting noise we rely on tools from shape analysis for our novel representation of the face. We show that our method preserves the shape of the average face and injects less noise than traditional methods for the same privacy budget. Our mechanism consists of two primary components, the first is generally applicable to function value summaries (as are commonly found in nonparametric statistics or functional data analysis) while the second is general to disk-like surfaces and hence more applicable than just to human faces., Comment: 10 pages, 6 figures