Search

Modern machine learning often relies on optimizing a neural network's parameters using a loss function to learn complex features. Beyond training, examining the loss function with respect to a network's parameters (i.e., as a loss landscape) can reveal insights into the architecture and learning process. While the local structure of the loss landscape surrounding an individual solution can be characterized using a variety of approaches, the global structure of a loss landscape, which includes potentially many local minima corresponding to different solutions, remains far more difficult to conceptualize and visualize. To address this difficulty, we introduce LossLens, a visual analytics framework that explores loss landscapes at multiple scales. LossLens integrates metrics from global and local scales into a comprehensive visual representation, enhancing model diagnostics. We demonstrate LossLens through two case studies: visualizing how residual connections influence a ResNet-20, and visualizing how physical parameters influence a physics-informed neural network (PINN) solving a simple convection problem.

The Hubbard-Holstein model is one of the central models that describe the competition between electron-electron and electron-phonon interactions. In one dimension and at half-filling, the interplay between an electronic spin-density wave and a phonon-driven charge-density wave is considered to stabilize an intermediate Luther-Emery liquid. Here we show that, once the extended metallic regime disappears, a narrow bond-order-wave phase emerges. We use an exact directed-loop quantum Monte Carlo method for retarded interactions to simulate system sizes of several hundred sites, necessary to observe the weak signatures of this novel phase. Our results suggest a second-order quantum phase transition between the two charge orders that only turns first-order at strong coupling. Our findings are reminiscent of the extended Hubbard model, but they are driven by competing frequency dependencies of the same dynamically-screened Hubbard interaction., Comment: 7 pages, 4 figures (plus supplemental material: 5 pages, 5 figures)

The question of what is the total entropy of the universe, how it compares to the maximal entropy of de Sitter space, and how it is distributed across the universe's components, bears considerable importance for a number of reasons. Here, we first update the computation of the entropy associated with various sectors of the observed universe, including in the diffuse cosmic and late-time gamma-ray and neutrino backgrounds, in baryonic matter both in diffuse components, in stars and stellar remnants, and in cosmic rays; we then update, crucially, the estimate of entropy in stellar-mass and super-massive black holes, whose abundance and mass function has come into increasingly sharp definition with recent observations and with the rapidly growing statistics of black-hole-black-hole mergers observed with gravity wave detectors. We also provide a new, corrected estimate of the potential entropy associated with a stochastic gravitational wave background, with dark sector radiations, and with several dark matter models. Finally, we utilize the similarly recently updated constraints on the abundance of hypothetical primordial black holes - black holes, that is, of non-stellar origin - to assess the maximal amount entropy they could store. We find that if supermassive primordial black holes exist, they can dominate the entropy budget of the universe consistently with current constraints on their abundance and mass function, to a level potentially not distant from the posited entropy associated with the cosmic event horizon of de Sitter spacetime. The same conclusion holds for certain dark sector models featuring a large number of dark degrees of freedom., Comment: 43 pages, 6 figures

Plastic scintillators are widely used for the detection of elementary particles, and 3D reconstruction of particle tracks is achieved by segmenting the detector into 3D granular structures. In this study, we present a novel prototype fabricated by additive manufacturing, consisting of a 5 x 5 x 5 array of 1 cm3 plastic scintillator cubes, each optically isolated. This innovative approach eliminates the need to construct complex monolithic geometries in a single operation and gets rid of the traditional time-consuming manufacturing and assembling processes. The prototype underwent performance characterization during a beam test at CERN's Proton-Synchrotron facility. Light yield, optical crosstalk, and light response uniformity, were evaluated. The prototype demonstrated a consistent light yield of approximately 27 photoelectrons (p.e.) per channel, similar to traditional cast scintillator detectors. Crosstalk between adjacent cubes averaged 4-5%, and light yield uniformity within individual cubes exhibited about 7% variation, indicating stability and reproducibility. These results underscore the potential of the novel additive manufacturing technique, for efficient and reliable production of high-granularity scintillator detectors.

Background: The rapid development and use of generative AI (GenAI) tools in academia presents complex and multifaceted ethical challenges for its users. Earlier research primarily focused on academic integrity concerns related to students' use of AI tools. However, limited information is available on the impact of GenAI on academic research. This study aims to examine the ethical concerns arising from the use of GenAI across different phases of research and explores potential strategies to encourage its ethical use for research purposes. Methods: We selected one or more GenAI platforms applicable to various research phases (e.g. developing research questions, conducting literature reviews, processing data, and academic writing) and analysed them to identify potential ethical concerns relevant for that stage. Results: The analysis revealed several ethical concerns, including a lack of transparency, bias, censorship, fabrication (e.g. hallucinations and false data generation), copyright violations, and privacy issues. These findings underscore the need for cautious and mindful use of GenAI. Conclusions: The advancement and use of GenAI are continuously evolving, necessitating an ongoing in-depth evaluation. We propose a set of practical recommendations to support researchers in effectively integrating these tools while adhering to the fundamental principles of ethical research practices.

The influence of naps on tinnitus was systematically assessed by exploring the frequency, clinical and demographic characteristics of this phenomenon. 9,724 data from two different tinnitus databases (Tinnitus Hub: $n = 6115$; Tinnitus Research Initiative (TRI): $n = 3627$) were included. After separate analysis of the databases, these results were then compared with each other. In the Tinnitus Hub survey database, a total of 31.1% reported an influence on tinnitus by taking a nap (26.9% in the TRI database), with much more frequent worsening after a nap than improvement (23.0% a little or a lot worse; TRI: 17.7% worse; 8.1% a little or a lot better; TRI: 9.2% better). The influence of napping on tinnitus was associated in both databases with other clinical features, such as the dependence of tinnitus on night quality, stress and somatosensory maneuvers. The present study confirms the clinical observation that more tinnitus sufferers report worsening after a nap than tinnitus sufferers reporting an improvement. It was consistently shown that tinnitus sufferers reporting nap-induced modulation of tinnitus also report more frequently an influence of night sleep on their tinnitus. Further clinical and polysomnographic research is warranted to better understand the interaction between sleep and tinnitus.

The Self-Optimization (SO) model can be considered as the third operational mode of the classical Hopfield Network (HN), leveraging the power of associative memory to enhance optimization performance. Moreover, is has been argued to express characteristics of minimal agency which, together with its biological plausibility, renders it useful for the study of artificial life. In this article, we draw attention to another facet of the SO model: its capacity for creativity. Drawing on the creativity studies literature, we argue that the model satisfies the necessary and sufficient conditions of a creative process. Moreover, we explore the dependency of different creative outcomes based on learning parameters, specifically the learning and reset rates. We conclude that the SO model allows for simulating and understanding the emergence of creative behaviors in artificial systems that learn., Comment: 46 pages, 9 figures

The global Biochar Industry has witnessed a surge in biochar production, with a total of 350k mt/year production in 2023. With the pressing climate goals set and the potential of Biochar Carbon Removal (BCR) as a climate-relevant technology, scaling up the number of new plants to over 1000 facilities per year by 2030 becomes imperative. However, such a massive scale-up presents not only technical challenges but also control and regulation issues, ensuring maximal output of plants while conforming to regulatory requirements. In this paper, we present a novel method of optimizing the process of a biochar plant based on machine learning methods. We show how a standard Random Forest Regressor can be used to model the states of the pyrolysis machine, the physics of which remains highly complex. This model then serves as a surrogate of the machine -- reproducing several key outcomes of the machine -- in a numerical optimization. This, in turn, could enable us to reduce NOx emissions -- a key regulatory goal in that industry -- while achieving maximal output still. In a preliminary test our approach shows remarkable results, proves to be applicable on two different machines from different manufacturers, and can be implemented on standard Internet of Things (IoT) devices more generally.

We use symmetry analysis and density functional theory to show that changes in magnetic order at a surface with respect to magnetic order in the bulk can be generically determined by considering local magnetoelectric responses of the crystal. Specifically, analysis of the atomic-site magnetoelectric responses, or equivalently the corresponding local magnetic multipoles, can be used to predict all surface magnetic modifications arising purely from symmetry lowering via termination of the bulk magnetic order. This analysis applies even in materials with no bulk magnetoelectric response or surface magnetization. We then demonstrate our arguments for two example antiferromagnets, metallic $\mathrm{CuMnAs}$ and rock-salt $\mathrm{NiO}$. We find that the $(010)$ and $(1\bar{1}0)$ surfaces of $\mathrm{CuMnAs}$ and $\mathrm{NiO}$ respectively exhibit a series of antiferroically, as well as roughness-sensitive, ferroically ordered, modifications of the surface magnetic dipole moments, via canting or changes in sublattice magnitude, consistent with the bulk ordering of the magnetic multipoles. Our findings demonstrate a universal bulk-boundary correspondance allowing the general prediction of minimal possible surface and interface magnetic modifications, even in non-magnetoelectric materials. Furthermore, it paves the way for more accurate interpretations of a wide variety of surface-sensitive measurements.

STARR (STAnford Research Repository) is a clinical research support ecosystem that supports basic science research, population health research and translational research at Stanford University. STARR consists of raw and analysis ready multi-modal data, and tools for cohort analysis and self service data access. STARR data is accessible on secure shared computing systems for ad hoc analysis. Also present is a suite of services on top of STARR, that allow researchers access to complex purpose built data cuts, common data models and software solutions. This manuscript is a research resource description and describes the evolution of STARR Tools that are used to offer self-service access to detailed clinical data for research purposes to researchers at Stanford Medicine, along with a framework used to ensure that data acquired via the self-service tools is handled in compliance with all applicable regulations and rules.

Borrowing historical data for use in clinical trials has increased in recent years. This is accomplished in the Bayesian framework by specification of informative prior distributions. One such approach is the robust mixture prior arising as a weighted mixture of an informative prior and a robust prior inducing dynamic borrowing that allows to borrow most when the current and external data are observed to be similar. The robust mixture prior requires the choice of three additional quantities: the mixture weight, and the mean and dispersion of the robust component. Some general guidance is available, but a case-by-case study of the impact of these quantities on specific operating characteristics seems lacking. We focus on evaluating the impact of parameter choices for the robust component of the mixture prior in one-arm and hybrid-control trials. The results show that all three quantities can strongly impact the operating characteristics. In particular, as already known, variance of the robust component is linked to robustness. Less known, however, is that its location can have a strong impact on Type I error rate and MSE which can even become unbounded. Further, the impact of the weight choice is strongly linked with the robust component's location and variance. Recommendations are provided for the choice of the robust component parameters, prior weight, alternative functional form for this component as well as considerations to keep in mind when evaluating operating characteristics.

Efficient deployment of neural networks on resource-constrained hardware demands optimal use of on-chip memory. In event-based processors, this is particularly critical for routing architectures, where substantial memory is dedicated to managing network connectivity. While prior work has focused on optimizing event routing during hardware design, optimizing memory utilization for routing during network training remains underexplored. Key challenges include: (i) integrating routing into the loss function, which often introduces non-differentiability, and (ii) computational expense in evaluating network mappability to hardware. We propose a hardware-algorithm co-design approach to train routing-aware neural networks. To address challenge (i), we extend the DeepR training algorithm, leveraging dynamic pruning and random re-assignment to optimize memory use. For challenge (ii), we introduce a proxy-based approximation of the mapping function to incorporate placement and routing constraints efficiently. We demonstrate our approach by optimizing a network for the Spiking Heidelberg Digits (SHD) dataset using a small-world connectivity-based hardware architecture as a case study. The resulting network, trained with our routing-aware methodology, is fully mappable to the hardware, achieving 5% more accuracy using the same number of parameters, and iso-accuracy with 10x less memory usage, compared to non-routing-aware training methods. This work highlights the critical role of co-optimizing algorithms and hardware to enable efficient and scalable solutions for constrained environments., Comment: 4 pages, 3 figures

The Lyman alpha (Lya) forest in the spectra of z>5 quasars provides a powerful probe of the late stages of the Epoch of Reionization (EoR). With the recent advent of exquisite datasets such as XQR-30, many models have struggled to reproduce the observed large-scale fluctuations in the Lya opacity. Here we introduce a Bayesian analysis framework that forward-models large-scale lightcones of IGM properties, and accounts for unresolved sub-structure in the Lya opacity by calibrating to higher-resolution hydrodynamic simulations. Our models directly connect physically-intuitive galaxy properties with the corresponding IGM evolution, without having to tune "effective" parameters or calibrate out the mean transmission. The forest data, in combination with UV luminosity functions and the CMB optical depth, are able to constrain global IGM properties at percent level precision in our fiducial model. Unlike many other works, we recover the forest observations without evoking a rapid drop in the ionizing emissivity from z~7 to 5.5, which we attribute to our sub-grid model for recombinations. In this fiducial model, reionization ends at $z=5.44\pm0.02$ and the EoR mid-point is at $z=7.7\pm0.1$. The ionizing escape fraction increases towards faint galaxies, showing a mild redshift evolution at fixed UV magnitude, Muv. Half of the ionizing photons are provided by galaxies fainter than Muv~-12, well below direct detection limits of optical/NIR instruments including JWST. We also show results from an alternative galaxy model that does not allow for a redshift evolution in the ionizing escape fraction. Despite being decisively disfavored by the Bayesian evidence, the posterior of this model is in qualitative agreement with that from our fiducial model. We caution however that our conclusions regarding the early stages of the EoR and which sources reionized the Universe are more model-dependent., Comment: 21 pages, 12 figures, 3 tables (submitted to PASA)

Context. The LISA space observatory will explore the sub-Hz spectrum of gravitational wave emission from the Universe. The space environment, where will be immersed in, is responsible for charge accumulation on its free falling test masses (TMs) due to the galactic cosmic rays (GCRs) and solar energetic particles (SEP) impinging on the spacecraft. Primary and secondary particles produced in the spacecraft material eventually reach the TMs by depositing a net positive charge fluctuating in time. This work is relevant for any present and future space missions that, like LISA, host free-falling TMs as inertial reference. Aims. The coupling of the TM charge with native stray electrostatic field produces noise forces on the TMs, which can limit the performance of the LISA mission. A precise knowledge of the charging process allows us to predict the intensity of these charge-induced disturbances and to design specific counter-measures. Methods. We present a comprehensive toolkit that allows us to calculate the TM charging time-series in a geometry representative of LISA mission, and the associated induced forces under different conditions of the space environment by considering the effects of short, long GCR flux modulations and SEPs. Results. We study, for each of the previously mentioned conditions, the impact of spurious forces associated with the TM charging process on the mission sensitivity for gravitational wave detection.

We show a priori bounds for the dynamic fractional $\Phi^4$ model on $\mathbb{T}^3$ in the full subcritical regime using the framework of Hairer's regularity structures theory. Assuming the model bounds our estimates imply global existence of solutions and existence of an invariant measure. We extend the method developed for the usual heat operator by Chandra, Moinat and Weber [CMW23] to the fractional heat operator, thereby treating a more physically relevant model. A key ingredient in this work is the development of localised multilevel Schauder estimates for the fractional heat operator which is not covered by Hairer's original work. Furthermore, the algebraic arguments from [CMW23] are streamlined significantly., Comment: Minor revision with updated references

Solid oxide fuel cells (SOFCs) are becoming increasingly important due to their high electrical efficiency, the flexible choice of fuels and relatively low emissions of pollutants. However, the increasingly growing demands for electrochemical devices require further performance improvements. Since it is well known that the 3D morphology of the electrodes, which is significantly influenced by the underlying manufacturing process, has a profound impact on the resulting performance, a deeper understanding for the structural changes caused by modifications of the manufacturing process or degradation phenomena is desirable. In the present paper, we investigate the influence of the operating time and the operating temperature on the 3D morphology of SOFC anodes using 3D image data obtained by focused-ion beam scanning electron microscopy, which is segmented into gadolinium-doped ceria, nickel and pore space. In addition, structural differences caused by manufacturing the anode via infiltration or powder technology, respectively, are analyzed quantitatively by means of various geometrical descriptors such as specific surface area, length of triple phase boundary per unit volume, mean geodesic tortuosity, and constrictivity. The computation of these descriptors from 3D image data is carried out both globally as well as locally to quantify the heterogeneity of the anode structure.

The Laser Interferometer Space Antenna (LISA) mission is being developed by ESA with NASA participation. As it has recently passed the Mission Adoption milestone, models of the instruments and noise performance are becoming more detailed, and likewise prototype data analyses must as well. Assumptions such as Gaussianity, stationarity, and data continuity are unrealistic, and must be replaced with physically motivated data simulations, and data analysis methods adapted to accommodate such likely imperfections. To this end, the LISA Data Challenges have produced datasets featuring time-varying and unequal constellation armlength, and measurement artifacts including data interruptions and instrumental transients. In this work, we assess the impact of these data artifacts on the inference of Galactic Binary and Massive Black Hole properties. Our analysis shows that the treatment of noise transients and gaps is necessary for effective parameter estimation, as they substantially corrupt the analysis if unmitigated. We find that straightforward mitigation techniques can significantly if imperfectly suppress artifacts. For the Galactic Binaries, mitigation of glitches was essentially total, while mitigations of the data gaps increased parameter uncertainty by approximately 10%. For the Massive Black Hole binaries the particularly pernicious glitches resulted in a 30% uncertainty increase after mitigations, while the data gaps can increase parameter uncertainty by up to several times. Critically, this underlines the importance of early detection of transient gravitational waves to ensure they are protected from planned data interruptions., Comment: 28 pages, 15 figures

The quest for a global quantum internet is based on the realization of a scalable network which requires quantum hardware with exceptional performance. Among them are quantum light sources providing deterministic, high brightness, high-fidelity entangled photons and quantum memories with coherence times in the millisecond range and above. To operate the network on a global scale, the quantum light source should emit at telecommunication wavelengths with minimum propagation losses. A cornerstone for the operation of such a quantum network is the demonstration of quantum teleportation. Here we realize full-photonic quantum teleportation employing one of the most promising platforms, i.e. semiconductor quantum dots, which can fulfill all the aforementioned requirements. Two remote quantum dots are used, one as a source of entangled photon pairs and the other as a single-photon source. The frequency mismatch between the triggered sources is erased using two polarization-preserving quantum frequency converters, enabling a Bell state measurement at telecommunication wavelengths. A post-selected teleportation fidelity of up to 0.721(33) is achieved, significantly above the classical limit, demonstrating successful quantum teleportation between light generated by distinct sources. These results mark a major advance for the semiconductor platform as a source of quantum light fulfilling a key requirement for a scalable quantum network. This becomes particularly relevant after the seminal breakthrough of addressing a nuclear spin in semiconductor quantum dots demonstrating long coherence times, thus fulfilling another crucial step towards a scalable quantum network.

We show that the surface of a centrosymmetric, collinear, compensated antiferromagnet, which hosts bulk ferroically ordered magnetic octupoles, exhibits a linear magnetoelectric effect, a net magnetization, and a net electric dipole moment. Thus, the surface satisfies all the conditions of a multiferroic, in striking contrast to the bulk, which is neither polar nor exhibits any net magnetization or linear magnetoelectric response. Of particular interest is the case of non-relativistic $d$-wave spin split antiferromagnets, in which the bulk magnetic octupoles and consequently the surface multiferroicity exist even without spin-orbit interaction. We illustrate our findings using first-principles calculations, taking FeF$_2$ as an example material. Our work underscores the bulk-boundary correspondence in these unconventional antiferromagnets., Comment: 8 pages and 6 figures

Large language models are increasingly becoming a cornerstone technology in artificial intelligence, the sciences, and society as a whole, yet the optimal strategies for dataset composition and filtering remain largely elusive. Many of the top-performing models lack transparency in their dataset curation and model development processes, posing an obstacle to the development of fully open language models. In this paper, we identify three core data-related challenges that must be addressed to advance open-source language models. These include (1) transparency in model development, including the data curation process, (2) access to large quantities of high-quality data, and (3) availability of artifacts and metadata for dataset curation and analysis. To address these challenges, we release RedPajama-V1, an open reproduction of the LLaMA training dataset. In addition, we release RedPajama-V2, a massive web-only dataset consisting of raw, unfiltered text data together with quality signals and metadata. Together, the RedPajama datasets comprise over 100 trillion tokens spanning multiple domains and with their quality signals facilitate the filtering of data, aiming to inspire the development of numerous new datasets. To date, these datasets have already been used in the training of strong language models used in production, such as Snowflake Arctic, Salesforce's XGen and AI2's OLMo. To provide insight into the quality of RedPajama, we present a series of analyses and ablation studies with decoder-only language models with up to 1.6B parameters. Our findings demonstrate how quality signals for web data can be effectively leveraged to curate high-quality subsets of the dataset, underscoring the potential of RedPajama to advance the development of transparent and high-performing language models at scale., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks

Nuclear magnetic resonance instruments are becoming available to the do-it-yourself community. The challenges encountered in the endeavor to build a magnetic resonance imaging instrument from scratch were confronted in a four-day hackathon at Singapore University of Technology and Design in spring 2024. One day was devoted to educational lectures and three days to system construction and testing. Seventy young researchers from all parts of the world formed six teams focusing on magnet, gradient coil, RF coil, console, system integration, and design, which together produced a working MRI instrument in three days. The different steps, encountered challenges, and their solutions are reported., Comment: 49 pages, 23 figures, comments welcome (this paper is meant to be useful to people constructing their own MRI systems)

In machine learning, a loss function measures the difference between model predictions and ground-truth (or target) values. For neural network models, visualizing how this loss changes as model parameters are varied can provide insights into the local structure of the so-called loss landscape (e.g., smoothness) as well as global properties of the underlying model (e.g., generalization performance). While various methods for visualizing the loss landscape have been proposed, many approaches limit sampling to just one or two directions, ignoring potentially relevant information in this extremely high-dimensional space. This paper introduces a new representation based on topological data analysis that enables the visualization of higher-dimensional loss landscapes. After describing this new topological landscape profile representation, we show how the shape of loss landscapes can reveal new details about model performance and learning dynamics, highlighting several use cases, including image segmentation (e.g., UNet) and scientific machine learning (e.g., physics-informed neural networks). Through these examples, we provide new insights into how loss landscapes vary across distinct hyperparameter spaces: we find that the topology of the loss landscape is simpler for better-performing models; and we observe greater variation in the shape of loss landscapes near transitions from low to high model performance.

Efforts are needed to identify and measure both communities' exposure to climate hazards and the social vulnerabilities that interact with these hazards, but the science of validating hazard vulnerability indicators is still in its infancy. Progress is needed to improve: 1) the selection of variables that are used as proxies to represent hazard vulnerability; 2) the applicability and scale for which these indicators are intended, including their transnational applicability. We administered an international urban survey in Buenos Aires, Argentina; Johannesburg, South Africa; London, United Kingdom; New York City, United States; and Seoul, South Korea in order to collect data on exposure to various types of extreme weather events, socioeconomic characteristics commonly used as proxies for vulnerability (i.e., income, education level, gender, and age), and additional characteristics not often included in existing composite indices (i.e., queer identity, disability identity, non-dominant primary language, and self-perceptions of both discrimination and vulnerability to flood risk). We then use feature importance analysis with gradient-boosted decision trees to measure the importance that these variables have in predicting exposure to various types of extreme weather events. Our results show that non-traditional variables were more relevant to self-reported exposure to extreme weather events than traditionally employed variables such as income or age. Furthermore, differences in variable relevance across different types of hazards and across urban contexts suggest that vulnerability indicators need to be fit to context and should not be used in a one-size-fits-all fashion., Comment: 15 pages, 2 figures

Millimeter-wave refracting telescopes targeting the degree-scale structure of the cosmic microwave background (CMB) have recently grown to diffraction-limited apertures of over 0.5 meters. These instruments are entirely housed in vacuum cryostats to support their sub-kelvin bolometric detectors and to minimize radiative loading from thermal emission due to absorption loss in their transmissive optical elements. The large vacuum window is the only optical element in the system at ambient temperature, and therefore minimizing loss in the window is crucial for maximizing detector sensitivity. This motivates the use of low-loss polymer materials and a window as thin as practicable. However, the window must simultaneously meet the requirement to keep sufficient vacuum, and therefore must limit gas permeation and remain mechanically robust against catastrophic failure under pressure. We report on the development of extremely thin composite polyethylene window technology that meets these goals. Two windows have been deployed for two full observing seasons on the BICEP3 and BA150 CMB telescopes at the South Pole. On BICEP3, the window has demonstrated a 6% improvement in detector sensitivity., Comment: 20 pages, 12 figures, 4 tables

Collimator detection remains a challenging task in X-ray systems with unreliable or non-available information about the detectors position relative to the source. This paper presents a physically motivated image processing pipeline for simulating the characteristics of collimator shadows in X-ray images. By generating randomized labels for collimator shapes and locations, incorporating scattered radiation simulation, and including Poisson noise, the pipeline enables the expansion of limited datasets for training deep neural networks. We validate the proposed pipeline by a qualitative and quantitative comparison against real collimator shadows. Furthermore, it is demonstrated that utilizing simulated data within our deep learning framework not only serves as a suitable substitute for actual collimators but also enhances the generalization performance when applied to real-world data.

Characterizing the loss of a neural network with respect to model parameters, i.e., the loss landscape, can provide valuable insights into properties of that model. Various methods for visualizing loss landscapes have been proposed, but less emphasis has been placed on quantifying and extracting actionable and reproducible insights from these complex representations. Inspired by powerful tools from topological data analysis (TDA) for summarizing the structure of high-dimensional data, here we characterize the underlying shape (or topology) of loss landscapes, quantifying the topology to reveal new insights about neural networks. To relate our findings to the machine learning (ML) literature, we compute simple performance metrics (e.g., accuracy, error), and we characterize the local structure of loss landscapes using Hessian-based metrics (e.g., largest eigenvalue, trace, eigenvalue spectral density). Following this approach, we study established models from image pattern recognition (e.g., ResNets) and scientific ML (e.g., physics-informed neural networks), and we show how quantifying the shape of loss landscapes can provide new insights into model performance and learning dynamics.

We propose a novel extension to the recently developed non-commutative Riemannian foliated branch-cut quantum gravity (BCQG). Based on an extended Faddeev-Jackiw symplectic deformation of the conventional Poisson algebra, we investigate non-commutativity effects on a symplectic topological manifold that provides a natural isomorphic setting composed by a triad of canonically conjugate scalar complex fields which comprise quantum complementary dualities. Based on a complementary analytically continued Friedmann-type equation, combined with a quantum approach based on the Ho\v{r}awa-Lifshitz quantum gravity, we describe the dynamic evolution of the universe's wave function, unfolding unprecedented predictions for the cosmic evolution and inflation. The non-commutative foliated quantum gravity approach offers a new perspective on explaining the accelerated cosmic expansion of the universe, strongly suggesting that non-commutative algebra induces the late accelerated growth of both the universe's wave function and the corresponding scale factor, along with their quantum counterparts. In contrast to the conventional inflationary model, where inflation requires a remarkably fine-tuned set of initial conditions in a patch of the universe, non-commutative foliated quantum gravity, analytically continued to the complex plane, captures short and long scales of spacetime, leading to an evolutionary cosmic dynamic through a topological reconfiguration of the primordial cosmic matter and energy content. This result introduces new speculative framework elements regarding the reconfiguration of matter and energy due to an underlying non-commutative spatio-temporal structure as a driver of spacetime cosmic acceleration.

Altermagnets constitute a new class of magnetic materials that combine properties previously thought to be exclusive to either antiferromagnets or ferromagnets, and have unique properties of their own. In particular, a combination of symmetries connecting magnetic sublattices gives rise to a band spin splitting exhibiting unconventional d, g, or i-wave character. Their unique electronic properties have already led to new spin-dependent transport effects. Here, we consider their spin and charge dynamics on ultrafast timescales. We use a minimal tight binding model that captures the main features of the altermagnetic candidate material KRu$_4$O$_8$. In the framework of this model, we compute the spin-dependent electronic scattering dynamics after ultrashort-pulse excitation and show through these microscopic calculations how electron-electron and electron-phonon scattering processes redistribute optically excited carriers in a 2D slice of the Brillouin zone. We find that the optically excited spin polarization is long lived (~1ps) compared to the electron-electron momentum scattering lifetime of roughly 10fs. This contrasts remarkably with the much shorter spin lifetimes observed in typical ultrafast electronic spin dynamics in conventional ferromagnets and antiferromagnets, making these pulse-driven spin excitation experiments a key probe of altermagnetism.

We show the late time limit of the spectral form factor (SFF) in unorientable JT gravity agrees with universal random matrix theory up to genus one in the topological expansion, establishing a key signature of quantum chaos for the time-reversal symmetric case. The loop equations for an orthogonal matrix model with spectral curve $y(z) \propto \sin(2\pi z)$ are used to compute the moduli space volume of unorientable surfaces. The divergences of the unorientable volumes are regularized by first regularizing the resolvents of the orthogonal matrix model. Using properties of the volumes, we streamline the loop equations to allow computation of the volumes that were previously inaccessible. The method can efficiently extract the part of the volume that contributes in the late time limit of the SFF. In this limit, the SFF becomes finite and independent of regularization.

Studies aiming to determine the astrophysical origins of nuclei produced by the rapid neutron capture process ($r$ process) rely on nuclear properties as inputs for simulations. The solar abundances can be used as a benchmark for such calculations, with the $r$-process rare-earth peak (REP) around mass number ($A$) 164 being of special interest due to its presently unknown origin. With the advancement of rare isotope beam production over the last decade and improvement in experimental sensitivities, many of these REP nuclides have become accessible for measurement. Masses are one of the most critical inputs as they impact multiple nuclear properties, namely the neutron-separation energies, neutron capture rates, $\beta$-decay rates, and $\beta$-delayed neutron emission probabilities. In this work, we report masses of 20 neutron-rich nuclides (along the Ba, La, Ce, Pr, Nd, Pm, Gd, Dy and Ho isotopic chains) produced at the CAlifornium Rare Isotope Breeder Upgrade (CARIBU) facility at Argonne National Laboratory. The masses were measured with the Canadian Penning trap (CPT) mass spectrometer using the Phase-Imaging Ion-Cyclotron-Resonance (PI-ICR) technique. We then use these new masses along with previously published CPT masses to inform predictions for a Markov Chain Monte Carlo (MCMC) procedure aiming to identify the astrophysical conditions consistent with both solar data and mass measurements. We show that the MCMC responds to this updated mass information, producing refined results for both mass predictions and REP abundances., Comment: 11 pages, 8 figures

Metasurfaces offer a powerful platform for effective light manipulation, which is crucial for advanced optical technologies. While designs of polarization-independent structures have reduced the need for polarized illumination, they are often limited by either low Q factors or low resonance modulation. Here, we design and experimentally demonstrate a metasurface with polarization-independent quasi-bound state in the continuum (quasi-BIC), where the unit cell consists of four silicon squares arranged in a two-dimensional array and the resonance properties can be controlled by adjusting the edge length difference between different squares. Our metasurface experimentally achieves a Q factor of approximately 100 and a resonance modulation of around 50%. This work addresses a common limitation in previous designs, which either achieved high Q factors exceeding 200 with a resonance modulation of less than 10%, leading to challenging signal-to-noise ratio requirements, or achieved strong resonance modulation with Q factors of only around 10, limiting light confinement and fine-tuning capabilities. In contrast, our metasurface ensures that the polarization-independent signal is sharp and distinct within the system, reducing the demands on signal-to-noise ratio and improving robustness. Experiments show the consistent performance across different polarization angles. This work contributes to the development of versatile optical devices, enhancing the potential for the practical application of BIC-based designs in areas such as optical filtering and sensing.

The increasing importance of supply chain security for digital devices -- from consumer electronics to critical infrastructure -- has created a high demand for skilled cybersecurity experts. These experts use Hardware Reverse Engineering (HRE) as a crucial technique to ensure trust in digital semiconductors. Recently, the US and EU have provided substantial funding to educate this cybersecurity-ready semiconductor workforce, but success depends on the widespread availability of academic training programs. In this paper, we investigate the current state of education in hardware security and HRE to identify efficient approaches for establishing effective HRE training programs. Through a systematic literature review, we uncover 13 relevant courses, including eight with accompanying academic publications. We identify common topics, threat models, key pedagogical features, and course evaluation methods. We find that most hardware security courses do not prioritize HRE, making HRE training scarce. While the predominant course structure of lectures paired with hands-on projects appears to be largely effective, we observe a lack of standardized evaluation methods and limited reliability of student self-assessment surveys. Our results suggest several possible improvements to HRE education and offer recommendations for developing new training courses. We advocate for the integration of HRE education into curriculum guidelines to meet the growing societal and industry demand for HRE experts., Comment: 7 pages, 1 figure, 2 tables. To be published in Proceedings of the 56th ACM Technical Symposium on Computer Science Education V. 1. For supplementary materials, see https://osf.io/dt8ne/

An isolated attosecond vortex $\gamma$-ray pulse is generated by using a relativistic spatiotemporal optical vortex (STOV) laser in particle-in-cell simulations. A $\sim$ 300-attosecond electron slice with transverse orbital angular momentum (TOAM) is initially selected and accelerated by the central spatiotemporal singularity of the STOV laser. This slice then collides with the laser's reflected Gaussian-like front from a planar target, initiating nonlinear Compton scattering and resulting in an isolated, attosecond ($\sim$ 300 as), highly collimated ($\sim$ 4$\degree$), ultra-brilliant ($\sim 5\times 10^{24}$ photons/s/mm$^2$/mrad$^2$/0.1\%BW at 1 MeV) $\gamma$-ray pulse. This STOV-driven approach overcomes the significant beam divergence and complex two-laser requirements of prior Gaussian-based methods while introducting TOAM to the attosecond $\gamma$-ray pulse, which opens avenues for ultrafast imaging, nuclear excitation, and detection applications.

Neutrino-nucleus cross section measurements are needed to improve interaction modeling to meet the precision needs of neutrino experiments in efforts to measure oscillation parameters and search for physics beyond the Standard Model. We review the difficulties associated with modeling neutrino-nucleus interactions that lead to a dependence on event generators in oscillation analyses and cross section measurements alike. We then describe data-driven model validation techniques intended to address this model dependence. The method relies on utilizing various goodness-of-fit tests and the correlations between different observables and channels to probe the model for defects in the phase space relevant for the desired analysis. These techniques shed light on relevant mis-modeling, allowing it to be detected before it begins to bias the cross section results. We compare more commonly used model validation methods which directly validate the model against alternative ones to these data-driven techniques and show their efficacy with fake data studies. These studies demonstrate that employing data-driven model validation in cross section measurements represents a reliable strategy to produce robust results that will stimulate the desired improvements to interaction modeling.

CONTEXT: Assurance Cases (ACs) are prepared to argue that the system's desired quality attributes (e.g., safety or security) are satisfied. While there is strong adoption of ACs, practitioners are often left asking an important question: are we confident that the claims made by the case are true? While many confidence assessment methods (CAMs) exist, little is known about the use of these methods in practice OBJECTIVE: Develop an understanding of the current state of practice for AC confidence assessment: what methods are used in practice and what barriers exist for their use? METHOD: Structured interviews were performed with practitioners with experience contributing to real-world ACs. Open-coding was performed on transcripts. A description of the current state of AC practice and future considerations for researchers was synthesized from the results. RESULTS: A total of n = 19 practitioners were interviewed. The most common CAMs were (peer-)review of ACs, dialectic reasoning ("defeaters"), and comparing against checklists. Participants preferred qualitative methods and expressed concerns about quantitative CAMs. Barriers to using CAMs included additional work, inadequate guidance, subjectivity and interpretation of results, and trustworthiness of methods. CONCLUSION: While many CAMs are described in the literature there is a gap between the proposed methods and needs of practitioners. Researchers working in this area should consider the need to: connect CAMs to established practices, use CAMs to communicate with interest holders, crystallize the details of CAM application, curate accessible guidance, and confirm that methods are trustworthy.

Common challenges in fault diagnosis include the lack of labeled data and the need to build models for each domain, resulting in many models that require supervision. Transfer learning can help tackle these challenges by learning cross-domain knowledge. Many approaches still require at least some labeled data in the target domain, and often provide unexplainable results. To this end, we propose a supervised transfer learning framework for fault diagnosis in wind turbines that operates in an Anomaly-Space. This space was created using SCADA data and vibration data and was built and provided to us by our research partner. Data within the Anomaly-Space can be interpreted as anomaly scores for each component in the wind turbine, making each value intuitive to understand. We conducted cross-domain evaluation on the train set using popular supervised classifiers like Random Forest, Light-Gradient-Boosting-Machines and Multilayer Perceptron as metamodels for the diagnosis of bearing and sensor faults. The Multilayer Perceptron achieved the highest classification performance. This model was then used for a final evaluation in our test set. The results show, that the proposed framework is able to detect cross-domain faults in the test set with a high degree of accuracy by using one single classifier, which is a significant asset to the diagnostic team., Comment: 9 pages, 4 figures, to be published in: Upper Rhine AI Symposium (URAI) 2024

Chemical vapour deposition enables large-domain growth of ideal graphene, yet many applications of graphene require the controlled inclusion of specific defects. We present a one-step chemical vapour deposition procedure aimed at retaining the precursor topology when incorporated into the grown carbonaceous film. When azupyrene, the molecular analogue of the Stone-Wales defect in graphene, is used as a precursor, carbonaceous monolayers with a range of morphologies are produced as a function of the copper substrate growth temperature. The higher the substrate temperature during deposition, the closer the resulting monolayer is to ideal graphene. Analysis, with a set of complementary materials characterisation techniques, reveals morphological changes closely correlated with changes in the atomic adsorption heights, network topology, and concentration of 5-/7-membered carbon rings. The engineered defective carbon monolayers can be transferred to different substrates, potentially enabling applications in nanoelectronics, sensorics, and catalysis., Comment: manuscript and SI provided

The eROSITA instrument on board Spectrum-Roentgen-Gamma has completed four scans of the X-ray sky, leading to the detection of almost one million X-ray sources in eRASS1 only, including multiple new X-ray binary candidates. We report on analysis of the X-ray binary 1eRASS J085039.9-421151, using a ~55\,ks long NuSTAR observation, following its detection in each eROSITA scan. Analysis of the eROSITA and NuSTAR X-ray spectra in combination with X-shooter data of the optical counterpart provide evidence of an X-ray binary with a red supergiant (RSG) companion, confirming previous results, although we determine a cooler spectral type of M2-3, owing to the presence of TiO bands in the optical and near infrared spectra. The X-ray spectrum is well-described by an absorbed power law with a high energy cutoff typically applied for accreting high mass X-ray binaries. In addition, we detect a strong fluorescent neutral iron line with an equivalent width of ~700\,eV and an absorption edge, the latter indicating strong absorption by a partial covering component. It is unclear if the partial absorber is ionised. There is no significant evidence of a cyclotron resonant scattering feature. We do not detect any pulsations in the NuSTAR lightcurves, possibly on account of a large spin period that goes undetected due to insufficient statistics at low frequencies or potentially large absorption that causes pulsations to be smeared out. Even so, the low persistent luminosity, the spectral parameters observed (photon index, $\Gamma<1.0$), and the minuscule likelihood of detection of RSG-black hole systems, suggest that the compact object is a neutron star., Comment: 17 pages, 11 figures, accepted by A&A

In molecular systems, the ultrafast motion of electrons initiates the process of chemical change. Tracking this electronic motion across molecules requires coupling attosecond time resolution to atomic-scale spatial sensitivity. In this work, we employ a pair of attosecond x-ray pulses from an x-ray free-electron laser to follow electron motion resulting from the sudden removal of an electron from a prototypical aromatic system, para-aminophenol. X-ray absorption enables tracking this motion with atomic-site specificity. Our measurements are compared with state-of-the-art computational modeling, reproducing the observed response across multiple timescales. Sub-femtosecond dynamics are assigned to states undergoing non-radiative decay, while few-femtosecond oscillatory motion is associated with electronic wavepacket motion in stable cation states, that will eventually couple to nuclear motion. Our work provides insight on the ultrafast charge motion preceding and initiating chemical transformations in moderately complex systems, and provides a powerful benchmark for computational models of ultrafast charge motion in matter.

Ultrafast transmission electron microscopy (UTEM) has emerged as a versatile technique for the time-resolved imaging of nanoscale dynamics on timescales down to few-hundred attoseconds but the temporal and spatial resolutions are still limited by the coherence properties of pulsed electron sources. Here, we report the development of a novel laser-driven linear cold-field electron emitter integrated in a state-of-the-art UTEM system. Illuminating the sharp tungsten emitter tip with a UV light pulse generates ultrashort femtosecond electron pulses of 220 fs pulse duration, with energy widths as low as 360 meV. The photoelectron emitter demonstrates exceptional spatial coherence, achieving focal spot sizes down to 2 $\mathring {\mathrm A}$ and a peak normalized brightness exceeding 6.7 $\times 10^{13}$ A/m$^2$sr. With an order-of-magnitude improvement compared to previously employed laser-driven Schottky field emitters, the present development opens up the field of ultrafast atomic-scale electron probing., Comment: 20 pages, Manuscript + Supporting Information

English, as a very high-resource language, enables the pretraining of high-quality large language models (LLMs). The same cannot be said for most other languages, as leading LLMs still underperform for non-English languages, likely due to a gap in the quality and diversity of the available multilingual pretraining corpora. In this work, we find that machine-translated text from a single high-quality source language can contribute significantly to the pretraining of multilingual LLMs. We translate FineWeb-Edu, a high-quality English web dataset, into French, German, and Spanish, resulting in a final 300B-token dataset, which we call TransWeb-Edu, and pretrain a 1.3B-parameter model, CuatroLLM, from scratch on this dataset. Across five non-English reasoning tasks, we show that CuatroLLM matches or outperforms state-of-the-art multilingual models trained using closed data, such as Llama3.2 and Gemma2, despite using an order of magnitude less data, such as about 6% of the tokens used for Llama3.2's training. We further demonstrate that with additional domain-specific pretraining, amounting to less than 1% of TransWeb-Edu, CuatroLLM surpasses the state of the art in multilingual reasoning. To promote reproducibility, we release our corpus, models, and training pipeline under open licenses at hf.co/britllm/CuatroLLM.

This paper introduces the first generalization and adaptation benchmark using machine learning for evaluating out-of-distribution performance of electromyography (EMG) classification algorithms. The ability of an EMG classifier to handle inputs drawn from a different distribution than the training distribution is critical for real-world deployment as a control interface. By predicting the user's intended gesture using EMG signals, we can create a wearable solution to control assistive technologies, such as computers, prosthetics, and mobile manipulator robots. This new out-of-distribution benchmark consists of two major tasks that have utility for building robust and adaptable control interfaces: 1) intersubject classification and 2) adaptation using train-test splits for time-series. This benchmark spans nine datasets--the largest collection of EMG datasets in a benchmark. Among these, a new dataset is introduced, featuring a novel, easy-to-wear high-density EMG wearable for data collection. The lack of open-source benchmarks has made comparing accuracy results between papers challenging for the EMG research community. This new benchmark provides researchers with a valuable resource for analyzing practical measures of out-of-distribution performance for EMG datasets. Our code and data from our new dataset can be found at emgbench.github.io.

In many modern computer application problems, the classification of image data plays an important role. Among many different supervised machine learning models, convolutional neural networks (CNNs) and linear discriminant analysis (LDA) as well as sophisticated variants thereof are popular techniques. In this work, two different domain decomposed CNN models are experimentally compared for different image classification problems. Both models are loosely inspired by domain decomposition methods and in addition, combined with a transfer learning strategy. The resulting models show improved classification accuracies compared to the corresponding, composed global CNN model without transfer learning and besides, also help to speed up the training process. Moreover, a novel decomposed LDA strategy is proposed which also relies on a localization approach and which is combined with a small neural network model. In comparison with a global LDA applied to the entire input data, the presented decomposed LDA approach shows increased classification accuracies for the considered test problems.

The spin-valley or Kramers qubit promises significantly enhanced spin-valley lifetimes due to strong coupling of the electrons' spin to their momentum (valley) degrees of freedom. In transition metal dichalcogenides (TMDCs) such spin-valley locking is expected to be particularly strong owing to the significant intrinsic spin-orbit coupling strength. Very recently, a small number of experiments on TMDC quantum dots have put forth evidence for spin-valley locking for the first time at the few-electron limit. Employing quantum transport theory, here we numerically simulate their ground- and excited-state transport spectroscopy signatures in a unified theoretical framework. In doing so, we reveal the operating conditions under which spin-valley locking occurs in TMDC quantum dots, thereby weaving the connection between intrinsic material properties and the experimental data under diverse conditions. Our simulations thus provide a predictive modeling tool for TMDC quantum dots at the few-electron limit allowing us to deduce from experiments the degree of spin-valley locking based on the SOC strength, inter-valley mixing, and the spin and valley $g$-factors. Our theoretical analysis provides an important milestone towards the next challenge of experimentally confirming valley-relaxation times using single-shot projective measurements, Comment: 13 pages, 6 figures with Appendix included, comments welcome

This work investigates the political biases and personality traits of ChatGPT, specifically comparing GPT-3.5 to GPT-4. In addition, the ability of the models to emulate political viewpoints (e.g., liberal or conservative positions) is analyzed. The Political Compass Test and the Big Five Personality Test were employed 100 times for each scenario, providing statistically significant results and an insight into the results correlations. The responses were analyzed by computing averages, standard deviations, and performing significance tests to investigate differences between GPT-3.5 and GPT-4. Correlations were found for traits that have been shown to be interdependent in human studies. Both models showed a progressive and libertarian political bias, with GPT-4's biases being slightly, but negligibly, less pronounced. Specifically, on the Political Compass, GPT-3.5 scored -6.59 on the economic axis and -6.07 on the social axis, whereas GPT-4 scored -5.40 and -4.73. In contrast to GPT-3.5, GPT-4 showed a remarkable capacity to emulate assigned political viewpoints, accurately reflecting the assigned quadrant (libertarian-left, libertarian-right, authoritarian-left, authoritarian-right) in all four tested instances. On the Big Five Personality Test, GPT-3.5 showed highly pronounced Openness and Agreeableness traits (O: 85.9%, A: 84.6%). Such pronounced traits correlate with libertarian views in human studies. While GPT-4 overall exhibited less pronounced Big Five personality traits, it did show a notably higher Neuroticism score. Assigned political orientations influenced Openness, Agreeableness, and Conscientiousness, again reflecting interdependencies observed in human studies. Finally, we observed that test sequencing affected ChatGPT's responses and the observed correlations, indicating a form of contextual memory.

Drug discovery typically consists of multiple steps, including identifying a target protein key to a disease's etiology, validating that interacting with this target could prevent symptoms or cure the disease, discovering a small molecule or biologic therapeutic to interact with it, and optimizing the candidate molecule through a complex landscape of required properties. Drug discovery related tasks often involve prediction and generation while considering multiple entities that potentially interact, which poses a challenge for typical AI models. For this purpose we present MAMMAL - Molecular Aligned Multi-Modal Architecture and Language - a method that we applied to create a versatile multi-task multi-align foundation model that learns from large-scale biological datasets (2 billion samples) across diverse modalities, including proteins, small molecules, and genes. We introduce a prompt syntax that supports a wide range of classification, regression, and generation tasks. It allows combining different modalities and entity types as inputs and/or outputs. Our model handles combinations of tokens and scalars and enables the generation of small molecules and proteins, property prediction, and transcriptomic lab test predictions. We evaluated the model on 11 diverse downstream tasks spanning different steps within a typical drug discovery pipeline, where it reaches new SOTA in 9 tasks and is comparable to SOTA in 2 tasks. This performance is achieved while using a unified architecture serving all tasks, in contrast to the original SOTA performance achieved using tailored architectures. The model code and pretrained weights are publicly available at https://github.com/BiomedSciAI/biomed-multi-alignment and https://huggingface.co/ibm/biomed.omics.bl.sm.ma-ted-458m.

In the article we give a self-contained new proof that a normal quasi-ordinary surface germ is analytically isomorphic to a cyclic quotient surface germ.

Research towards energy optimization in buildings heavily relies on building-related data such as measured indoor climate factors. While data collection is a labor- and cost-intensive task, simulations are a cheap alternative to generate datasets of arbitrary sizes, particularly useful for data-intensive deep learning methods. In this paper, we present the tool EnergyPlus Room Simulator, which enables the simulation of indoor climate in a specific room of a building using the simulation software EnergyPlus. It allows to alter room models and simulate various factors such as temperature, humidity, and CO2 concentration. In contrast to manually working with EnergyPlus, this tool enhances the simulation process by offering a convenient interface, including a user-friendly graphical user interface (GUI) as well as a REST API. The tool is intended to support scientific, building-related tasks such as occupancy detection on a room level by facilitating fast access to simulation data that may, for instance, be used for pre-training machine learning models., Comment: Presented at BuildSim Nordic 2024. The conference was held from June 9 to 11, 2024, in Espoo, Finland

We present a generalization of the phaseless auxiliary-field quantum Monte Carlo (AFQMC) method to cavity quantum-electrodynamical (QED) matter systems. The method can be formulated in both the Coulomb and the dipole gauge. We verify its accuracy by benchmarking calculations on a set of small molecules against full configuration interaction and state-of-the-art QED coupled cluster (QED-CCSD) calculations. Our results show that (i) gauge invariance can be achieved within correlation-consistent Gaussian basis sets, (ii) the accuracy of QED-CCSD can be enhanced significantly by adding the standard perturbative triples correction without light-matter coupling, and (iii) there is a straightforward way to evaluate the differential expression for the photon occupation number that works in any gauge. The high accuracy and favorable computational scaling of our AFQMC approach will enable a broad range of applications. Besides polaritonic chemistry, the method opens a way to simulate extended QED matter systems., Comment: 10 pages, 7 figures

In the strong-coupling regime, the interaction between light and matter reaches a hybridization state where the photonic and material components become inseparably linked. Using tailored states of light to break symmetries in such systems can underpin the development of novel non-equilibrium quantum materials. Chiral optical cavities offer a promising way for this, enabling either temporal or spatial symmetry-breaking, both of which are unachievable with conventional mirror cavities. For spatial symmetry-breaking a cavity needs to discriminate the handedness of circularly polarized light, a functionality that can only be achieved with metamaterials. Here, we suggest and demonstrate experimentally a chiral transition metal dichalcogenide (TMDC) metasurface with broken out-of-plane symmetry, allowing for a selective formation of self-hybridized exciton-polaritons with specific chirality. Our metasurface cavity maintains maximum chirality for oblique incidence up to 20{\deg}, significantly outperforming all previously known designs, thereby turning the angle of incidence from a constraint to a new degree of freedom for sub-nanometer precise control of resonance wavelengths. Moreover, we study the chiral strong-coupling regime in nonlinear experiments and show the polariton-driven nature of chiral third-harmonic generation. Our results demonstrate a clear pathway towards novel quantum material engineering with implications in a wide range of photonics research, such as superconductivity and valleytronics., Comment: 34 pages, 4 figures, and 12 supporting figures