Search

The mechanical complexity of flapping wings, their unsteady aerodynamic flow, and challenge of making measurements at the scale of a sub-gram flapping-wing flying insect robot (FIR) make its behavior hard to predict. Knowing the precise mapping from voltage input to torque output, however, can be used to improve their mechanical and flight controller design. To address this challenge, we created a sensitive force-torque sensor based on a flexured gimbal that only requires a standard motion capture system or accelerometer for readout. Our device precisely and accurately measures pitch and roll torques simultaneously, as well as thrust, on a tethered flapping-wing FIR in response to changing voltage input signals. With it, we were able to measure cross-axis coupling of both torque and thrust input commands on a 180 mg FIR, the UW Robofly. We validated these measurements using free-flight experiments. Our results showed that roll and pitch have maximum cross-axis coupling errors of 8.58% and 17.24%, respectively, relative to the range of torque that is possible. Similarly, varying the pitch and roll commands resulted in up to a 5.78% deviation from the commanded thrust, across the entire commanded torque range. Our system, the first to measure two torque axes simultaneously, shows that torque commands have a negligible cross-axis coupling on both torque and thrust., Comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

In the field of materials science and manufacturing, a vast amount of heterogeneous data exists, encompassing measurement and simulation data, machine data, publications, and more. This data serves as the bedrock of valuable knowledge that can be leveraged for various engineering applications. However, efficiently storing and handling such diverse data remain significantly challenging, often due to the lack of standardization and integration across different organizational units. Addressing these issues is crucial for fully utilizing the potential of data-driven approaches in these fields. In this paper, we present a novel technology stack named Dataspace Management System (DSMS) for powering dataspace solutions. The core of DSMS lies on its distinctive knowledge management approach tuned to meet the specific demands of the materials science and manufacturing domain, all while adhering to the FAIR principles. This includes data integration, linkage, exploration, visualization, processing, and enrichment, in order to support engineers in decision-making and in solving design and optimization problems. We provide an architectural overview and describe the core components of DSMS. Additionally, we demonstrate the applicability of DSMS to typical data processing tasks in materials science through use cases from two research projects, namely StahlDigital and KupferDigital, both part of the German MaterialDigital initiative.

Light activated local stimulation and sensing of biological cells offers enormous potential for minimally invasive bioelectronic interfaces. Organic semiconductors are a promising material class to achieve this kind of transduction due to their optoelectronic properties and biocompatibility. Here we investigate which material properties are necessary to keep the optical excitation localized. This is critical to single cell transduction with high spatial resolution. As a model system we use organic photocapacitors for cell stimulation made of the small molecule semiconductors H2Pc and PTCDI. We investigate the spatial broadening of the localized optical excitation with photovoltage microscopy measurements. Our experimental data combined with modelling show that resolution losses due to the broadening of the excitation are directly related to the effective diffusion length of charge carriers generated at the heterojunction. With additional transient photovoltage measurements we find that the H2Pc/PTCDI heterojunction offers a small diffusion length of lambda = 1.5 +/- 0.1 um due to the small mobility of charge carriers along the heterojunction. Instead covering the heterojunction with a layer of PEDOT:PSS improves the photocapacitor performance but increases the carrier diffusion length to lambda = 7.0 +/- 0.3 um due to longer lifetime and higher carrier mobility. Furthermore, we introduce electrochemical photocurrent microscopy experiments to demonstrate micrometric resolution with the pn-junction under realistic aqueous operation conditions. This work offers valuable insights into the physical mechanisms governing the excitation and transduction profile and provide design principles for future organic semiconductor junctions, aiming to achieve high efficiency and high spatial resolution.

Asset Administration Shells are trending in Industry 4.0. In February 2024, the Industrial Digital Twin Association announced 84 and released 18 AAS submodel specifications. As an enabler on programming level, dedicated APIs are needed, for which, at this level of scale, automated creation is desirable. In this paper, we present a model-driven approach, which transforms extracted information from IDTA specifications into an intermediary meta-model and, from there, generates API code and tests. We show we can process all current IDTA specifications successfully leading in total to more than 50000 lines of code. However, syntactical variations and issues in the specifications impose obstacles that require human intervention or AI support. We also discuss experiences that we made and lessons learned., Comment: 8 pages, 1 figure

Observing spatial entanglement in the Bose-Marletto-Vedral (BMV) experiment would demonstrate the existence of non-classical properties of the gravitational field. We show that the special relativistic invariance of the linear regime of general relativity implies that all the components of the gravitational potential must be non-classical. This is simply necessary in order to describe the BMV entanglement consistently across different inertial frames of reference. On the other hand, we show that the entanglement in accelerated frames could differ from that in stationary frames., Comment: 5 pages, 3 figures

2D topological insulators promise novel approaches towards electronic, spintronic, and quantum device applications. This is owing to unique features of their electronic band structure, in which bulk-boundary correspondences enforces the existence of 1D spin-momentum locked metallic edge states - both helical and chiral - surrounding an electrically insulating bulk. Forty years since the first discoveries of topological phases in condensed matter, the abstract concept of band topology has sprung into realization with several materials now available in which sizable bulk energy gaps - up to a few hundred meV - promise to enable topology for applications even at room-temperature. Further, the possibility of combining 2D TIs in heterostructures with functional materials such as multiferroics, ferromagnets, and superconductors, vastly extends the range of applicability beyond their intrinsic properties. While 2D TIs remain a unique testbed for questions of fundamental condensed matter physics, proposals seek to control the topologically protected bulk or boundary states electrically, or even induce topological phase transitions to engender switching functionality. Induction of superconducting pairing in 2D TIs strives to realize non-Abelian quasiparticles, promising avenues towards fault-tolerant topological quantum computing. This roadmap aims to present a status update of the field, reviewing recent advances and remaining challenges in theoretical understanding, materials synthesis, physical characterization and, ultimately, device perspectives.

The novel SoLAr concept aims to extend sensitivities of liquid-argon neutrino detectors down to the MeV scale for next-generation detectors. SoLAr plans to accomplish this with a liquid-argon time projection chamber that employs an anode plane with dual charge and light readout, which enables precision matching of light and charge signals for data acquisition and reconstruction purposes. We present the results of a first demonstration of the SoLAr detector concept with a small-scale prototype detector integrating a pixel-based charge readout and silicon photomultipliers on a shared printed circuit board. We discuss the design of the prototype, and its operation and performance, highlighting the capability of such a detector design.

Today, we see a drastic increase in LLM-based user interfaces to support users in various tasks. Also, in programming, we witness a productivity boost with features like LLM-supported code completion and conversational agents to generate code. In this work, we look at the future of computational notebooks by enriching them with LLM support. We propose a spectrum of support, from simple inline code completion to executable code that was the output of a conversation. We showcase five concrete examples for potential user interface designs and discuss their benefits and drawbacks. With this, we hope to inspire the future development of LLM-supported computational notebooks., Comment: 1st ACM CHI Workshop on Human-Notebook Interactions

A significant challenge in measurements of neutrino oscillations is reconstructing the incoming neutrino energies. While modern fully-active tracking calorimeters such as liquid argon time projection chambers in principle allow the measurement of all final state particles above some detection threshold, undetected neutrons remain a considerable source of missing energy with little to no data constraining their production rates and kinematics. We present the first demonstration of tagging neutrino-induced neutrons in liquid argon time projection chambers using secondary protons emitted from neutron-argon interactions in the MicroBooNE detector. We describe the method developed to identify neutrino-induced neutrons and demonstrate its performance using neutrons produced in muon-neutrino charged current interactions. The method is validated using a small subset of MicroBooNE's total dataset. The selection yields a sample with $60\%$ of selected tracks corresponding to neutron-induced secondary protons.

We present a deep learning-based method for estimating the neutrino energy of charged-current neutrino-argon interactions. We employ a recurrent neural network (RNN) architecture for neutrino energy estimation in the MicroBooNE experiment, utilizing liquid argon time projection chamber (LArTPC) detector technology. Traditional energy estimation approaches in LArTPCs, which largely rely on reconstructing and summing visible energies, often experience sizable biases and resolution smearing because of the complex nature of neutrino interactions and the detector response. The estimation of neutrino energy can be improved after considering the kinematics information of reconstructed final-state particles. Utilizing kinematic information of reconstructed particles, the deep learning-based approach shows improved resolution and reduced bias for the muon neutrino Monte Carlo simulation sample compared to the traditional approach. In order to address the common concern about the effectiveness of this method on experimental data, the RNN-based energy estimator is further examined and validated with dedicated data-simulation consistency tests using MicroBooNE data. We also assess its potential impact on a neutrino oscillation study after accounting for all statistical and systematic uncertainties and show that it enhances physics sensitivity. This method has good potential to improve the performance of other physics analyses.

Large Language Models (LLMs) have become widely adopted recently. Research explores their use both as autonomous agents and as tools for software engineering. LLM-integrated applications, on the other hand, are software systems that leverage an LLM to perform tasks that would otherwise be impossible or require significant coding effort. While LLM-integrated application engineering is emerging as new discipline, its terminology, concepts and methods need to be established. This study provides a taxonomy for LLM-integrated applications, offering a framework for analyzing and describing these systems. It also demonstrates various ways to utilize LLMs in applications, as well as options for implementing such integrations. Following established methods, we analyze a sample of recent LLM-integrated applications to identify relevant dimensions. We evaluate the taxonomy by applying it to additional cases. This review shows that applications integrate LLMs in numerous ways for various purposes. Frequently, they comprise multiple LLM integrations, which we term ``LLM components''. To gain a clear understanding of an application's architecture, we examine each LLM component separately. We identify thirteen dimensions along which to characterize an LLM component, including the LLM skills leveraged, the format of the output, and more. LLM-integrated applications are described as combinations of their LLM components. We suggest a concise representation using feature vectors for visualization. The taxonomy is effective for describing LLM-integrated applications. It can contribute to theory building in the nascent field of LLM-integrated application engineering and aid in developing such systems. Researchers and practitioners explore numerous creative ways to leverage LLMs in applications. Though challenges persist, integrating LLMs may revolutionize the way software systems are built.

Tails used as inertial appendages induce body rotations of animals and robots, a phenomenon that is governed largely by the ratio of the body and tail moments of inertia. However, vertebrate tails have more degrees of freedom (e.g., number of joints, rotational axes) than most current theoretical models and robotic tails. To understand how morphology affects inertial appendage function, we developed an optimization-based approach that finds the maximally effective tail trajectory and measures error from a target trajectory. For tails of equal total length and mass, increasing the number of equal-length joints increased the complexity of maximally effective tail motions. When we optimized the relative lengths of tail bones while keeping the total tail length, mass, and number of joints the same, this optimization-based approach found that the lengths match the pattern found in the tail bones of mammals specialized for inertial maneuvering. In both experiments, adding joints enhanced the performance of the inertial appendage, but with diminishing returns, largely due to the total control effort constraint. This optimization-based simulation can compare the maximum performance of diverse inertial appendages that dynamically vary in moment of inertia in 3D space, predict inertial capabilities from skeletal data, and inform the design of robotic inertial appendages.

Assessing tumor response to systemic therapies is one of the main applications of PET/CT. Routinely, only a small subset of index lesions out of multiple lesions is analyzed. However, this operator dependent selection may bias the results due to possible significant inter-metastatic heterogeneity of response to therapy. Automated, AI based approaches for lesion tracking hold promise in enabling the analysis of many more lesions and thus providing a better assessment of tumor response. This work introduces a Siamese CNN approach for lesion tracking between PET/CT scans. Our approach is applied on the laborious task of tracking a high number of bone lesions in full-body baseline and follow-up [68Ga]Ga- or [18F]F-PSMA PET/CT scans after two cycles of [177Lu]Lu-PSMA therapy of metastatic castration resistant prostate cancer patients. Data preparation includes lesion segmentation and affine registration. Our algorithm extracts suitable lesion patches and forwards them into a Siamese CNN trained to classify the lesion patch pairs as corresponding or non-corresponding lesions. Experiments have been performed with different input patch types and a Siamese network in 2D and 3D. The CNN model successfully learned to classify lesion assignments, reaching a lesion tracking accuracy of 83 % in its best configuration with an AUC = 0.91. For remaining lesions the pipeline accomplished a re-identification rate of 89 %. We proved that a CNN may facilitate the tracking of multiple lesions in PSMA PET/CT scans. Future clinical studies are necessary if this improves the prediction of the outcome of therapies., Comment: 25 pages, 9 figures, 3 tables

Recent advancements in generative models have highlighted the crucial role of image tokenization in the efficient synthesis of high-resolution images. Tokenization, which transforms images into latent representations, reduces computational demands compared to directly processing pixels and enhances the effectiveness and efficiency of the generation process. Prior methods, such as VQGAN, typically utilize 2D latent grids with fixed downsampling factors. However, these 2D tokenizations face challenges in managing the inherent redundancies present in images, where adjacent regions frequently display similarities. To overcome this issue, we introduce Transformer-based 1-Dimensional Tokenizer (TiTok), an innovative approach that tokenizes images into 1D latent sequences. TiTok provides a more compact latent representation, yielding substantially more efficient and effective representations than conventional techniques. For example, a 256 x 256 x 3 image can be reduced to just 32 discrete tokens, a significant reduction from the 256 or 1024 tokens obtained by prior methods. Despite its compact nature, TiTok achieves competitive performance to state-of-the-art approaches. Specifically, using the same generator framework, TiTok attains 1.97 gFID, outperforming MaskGIT baseline significantly by 4.21 at ImageNet 256 x 256 benchmark. The advantages of TiTok become even more significant when it comes to higher resolution. At ImageNet 512 x 512 benchmark, TiTok not only outperforms state-of-the-art diffusion model DiT-XL/2 (gFID 2.74 vs. 3.04), but also reduces the image tokens by 64x, leading to 410x faster generation process. Our best-performing variant can significantly surpasses DiT-XL/2 (gFID 2.13 vs. 3.04) while still generating high-quality samples 74x faster., Comment: A compact 1D Image Tokenization method, leading to SOTA generation performance while being substantially faster. Project page at https://yucornetto.github.io/projects/titok.html

SBND is the near detector of the Short-Baseline Neutrino program at Fermilab. Its location near to the Booster Neutrino Beam source and relatively large mass will allow the study of neutrino interactions on argon with unprecedented statistics. This paper describes the expected performance of the SBND photon detection system, using a simulated sample of beam neutrinos and cosmogenic particles. Its design is a dual readout concept combining a system of 120 photomultiplier tubes, used for triggering, with a system of 192 X-ARAPUCA devices, located behind the anode wire planes. Furthermore, covering the cathode plane with highly-reflective panels coated with a wavelength-shifting compound recovers part of the light emitted towards the cathode, where no optical detectors exist. We show how this new design provides a high light yield and a more uniform detection efficiency, an excellent timing resolution and an independent 3D-position reconstruction using only the scintillation light. Finally, the whole reconstruction chain is applied to recover the temporal structure of the beam spill, which is resolved with a resolution on the order of nanoseconds., Comment: 21 pages, 17 figures

The Microwave SQUID Multiplexer ({\mu}MUX) is the device of choice for the readout of a large number of Low-Temperature Detectors in a wide variety of experiments within the fields of astronomy and particle physics. While it offers large multiplexing factors, the system noise performance is highly dependent on the cold and warm-readout electronic systems used to read it out, as well as the demodulation domain and parameters chosen. In order to understand the impact of the readout systems in the overall detection system noise performance, first we extended the available {\mu}MUX simulation frameworks including additive and multiplicative noise sources in the probing tones (i.e. phase and amplitude noise), along with the capability of demodulating the scientific data, either in resonator's phase or scattering amplitude. Then, considering the additive noise as a dominant noise source, the optimum readout parameters to achieve minimum system noise were found for both open-loop and flux-ramp demodulation schemes in the aforementioned domains. Later, we evaluated the system noise sensitivity to multiplicative noise sources under the optimum readout parameters. Finally, as a case study, we evaluated the optimal demodulation domain and expected system noise level for a typical Software-Defined Radio (SDR) readout system. This work leads to an improved system performance prediction and noise engineering based on the available readout electronics and selected demodulation domain., Comment: This article has been submitted to the Journal of Applied Physics. This is the revised version (V1)

The usual way to interpret language models (LMs) is to test their performance on different benchmarks and subsequently infer their internal processes. In this paper, we present an alternative approach, concentrating on the quality of LM processing, with a focus on their language abilities. To this end, we construct 'linguistic task spaces' -- representations of an LM's language conceptualisation -- that shed light on the connections LMs draw between language phenomena. Task spaces are based on the interactions of the learning signals from different linguistic phenomena, which we assess via a method we call 'similarity probing'. To disentangle the learning signals of linguistic phenomena, we further introduce a method called 'fine-tuning via gradient differentials' (FTGD). We apply our methods to language models of three different scales and find that larger models generalise better to overarching general concepts for linguistic tasks, making better use of their shared structure. Further, the distributedness of linguistic processing increases with pre-training through increased parameter sharing between related linguistic tasks. The overall generalisation patterns are mostly stable throughout training and not marked by incisive stages, potentially explaining the lack of successful curriculum strategies for LMs., Comment: To be published at ACL 2024 (main)

Modeling dynamical systems, e.g. in climate and engineering sciences, often necessitates solving partial differential equations. Neural operators are deep neural networks designed to learn nontrivial solution operators of such differential equations from data. As for all statistical models, the predictions of these models are imperfect and exhibit errors. Such errors are particularly difficult to spot in the complex nonlinear behaviour of dynamical systems. We introduce a new framework for approximate Bayesian uncertainty quantification in neural operators using function-valued Gaussian processes. Our approach can be interpreted as a probabilistic analogue of the concept of currying from functional programming and provides a practical yet theoretically sound way to apply the linearized Laplace approximation to neural operators. In a case study on Fourier neural operators, we show that, even for a discretized input, our method yields a Gaussian closure--a structured Gaussian process posterior capturing the uncertainty in the output function of the neural operator, which can be evaluated at an arbitrary set of points. The method adds minimal prediction overhead, can be applied post-hoc without retraining the neural operator, and scales to large models and datasets. We showcase the efficacy of our approach through applications to different types of partial differential equations.

We consider the discretized version of the sigma-model with supersphere target space $OSp(N+2m|2m)/OSp(N+2m-1|2m)$ and present a preliminary numerical study of bosonic and fermionic two-point functions for the cases $OSp(3|2)$ and $OSp(5|2)$. We observe consistency with the expectations of this supersymmetric setup and discuss the sign problem., Comment: 12 pages, 14 figures, Contribution to the Proceedings European network for Particle physics, Lattice field theory and Extreme computing

Most Zero-shot Multi-speaker TTS (ZS-TTS) systems support only a single language. Although models like YourTTS, VALL-E X, Mega-TTS 2, and Voicebox explored Multilingual ZS-TTS they are limited to just a few high/medium resource languages, limiting the applications of these models in most of the low/medium resource languages. In this paper, we aim to alleviate this issue by proposing and making publicly available the XTTS system. Our method builds upon the Tortoise model and adds several novel modifications to enable multilingual training, improve voice cloning, and enable faster training and inference. XTTS was trained in 16 languages and achieved state-of-the-art (SOTA) results in most of them., Comment: Accepted at INTERSPEECH 2024

The expected regret of any reinforcement learning algorithm is lower bounded by $\Omega\left(\sqrt{DXAT}\right)$ for undiscounted returns, where $D$ is the diameter of the Markov decision process, $X$ the size of the state space, $A$ the size of the action space and $T$ the number of time steps. However, this lower bound is general. A smaller regret can be obtained by taking into account some specific knowledge of the problem structure. In this article, we consider an admission control problem to an $M/M/c/S$ queue with $m$ job classes and class-dependent rewards and holding costs. Queuing systems often have a diameter that is exponential in the buffer size $S$, making the previous lower bound prohibitive for any practical use. We propose an algorithm inspired by UCRL2, and use the structure of the problem to upper bound the expected total regret by $O(S\log T + \sqrt{mT \log T})$ in the finite server case. In the infinite server case, we prove that the dependence of the regret on $S$ disappears.

In quantum information, nonlocal games are particularly useful for differentiating classical, quantum, and non-signalling correlations. An example of differentiation is given by the principle of no-collapse of communication complexity, which is often interpreted as necessary for a feasible physical theory. It is satisfied by quantum correlations but violated by some non-signalling ones. In this work, we investigate this principle in the context of three nonlocal games related to graph theory, starting from the well-known graph isomorphism and graph coloring games, and introducing a new game, the vertex distance game, with a parameter $D\in\mathbb N$, that generalizes the former two to some extent. For these three games, we prove that perfect non-signalling strategies collapse communication complexity under favorable conditions. We also define a refinement of fractional isomorphism of graphs, namely D-fractional isomorphisms, and we show that this characterizes perfect non-signalling strategies for the vertex distance game. Surprisingly, we observe that non-signalling strategies provide a finer distinction for the new game compared to classical and quantum strategies since the parameter D is visible only in the non-signalling setting., Comment: 42 pages, 9 figures

Modern diesel engines temporarily use a very late post-injection in the combustion cycle to either generate heat for a diesel particulate filter regeneration or purge a lean NOx trap. In some configurations, unburned fuel is left at the cylinder walls and is transported via the piston rings toward the lower crankcase region, where fuel may dilute the oil. Reduced oil lubrication shortens the oil service intervals and increases friction. Beside diesel fuel, this problem may also occur for other types of liquid fuels such as alcohols and e-fuels. The exact transport mechanism of the unburned fuel via the piston ring pack grooves and cylinder wall is hard to measure experimentally, motivating numerical flow simulation in early design stages for an in-depth understanding of the involved processes. A new CFD simulation methodology has been developed to investigate the transient, compressible, multiphase flow around the piston ring pack, through the gap between piston and liner, and its impact on fuel or oil transport. The modern level-set approach is used for the multiphase physics, which directly captures the sharp interface between blow-by gas and fuel or oil. Transient blow-by and two-phase flow simulations have been extensively applied to a Ford 2.0L I4 diesel test engine. The results confirm the validity of the flow compressibility assumption and highlight the sensitivity of the fuel leakage regarding piston sealing ring movement and highly resolved meshes for the multiphase flow. Based on the simulation results, design recommendations for piston and piston ring geometry are provided to reduce the fuel transport toward the crankcase.

The manifold hypothesis presumes that high-dimensional data lies on or near a low-dimensional manifold. While the utility of encoding geometric structure has been demonstrated empirically, rigorous analysis of its impact on the learnability of neural networks is largely missing. Several recent results have established hardness results for learning feedforward and equivariant neural networks under i.i.d. Gaussian or uniform Boolean data distributions. In this paper, we investigate the hardness of learning under the manifold hypothesis. We ask which minimal assumptions on the curvature and regularity of the manifold, if any, render the learning problem efficiently learnable. We prove that learning is hard under input manifolds of bounded curvature by extending proofs of hardness in the SQ and cryptographic settings for Boolean data inputs to the geometric setting. On the other hand, we show that additional assumptions on the volume of the data manifold alleviate these fundamental limitations and guarantee learnability via a simple interpolation argument. Notable instances of this regime are manifolds which can be reliably reconstructed via manifold learning. Looking forward, we comment on and empirically explore intermediate regimes of manifolds, which have heterogeneous features commonly found in real world data.

Background: Automatic treatment planning promises many benefits for both research and clinical environments. For clinics, autoplanning promises to reduce planning time and achieve more comparable treatment plans and thereby reduce inter-planner variability. Further, it can assist clinicians in quality assurance by providing a minimum plan quality standard. Finally, autoplanning is an essential part of patient selection, which is crucial for the advancement of proton therapy itself. Methods: A retrospective planning study using a cohort of 17 head-and-neck cancer patients treated at our institute. The clinically accepted plans created by dosimetrists (d-plans) were compared to automatically generated JulianA plans (j-plans). Both methods used the same beam arrangement. The plans were analysed by two expert reviewers without knowing how each plan was created. They assessed the plan quality and stated a preference. Results: All of the j-plans were deemed rather or clearly acceptable, resulting in a higher acceptability than the d-plans. The j-plan was considered superior in 14 (82.4%) cases, of equal quality for 1 (5.9%) and inferior to the d-plan for only 2 (11.8%) of the cases. The reviewers concluded that JulianA achieves more conformal dose distributions for the 15 (88.2%) cases where the j-plans were at least as good as the d-plans. Conclusions: The results show that the JulianA is ready to be used as a clinical quality assurance tool and research platform at our institute. While these results are encouraging, further research is needed to reduce the number of spots further and introduce robustness considerations into the optimisation algorithm in order to employ it on a daily basis for patient treatment.

The BICEP/$\textit{Keck}$ (BK) series of cosmic microwave background (CMB) polarization experiments has, over the past decade and a half, produced a series of field-leading constraints on cosmic inflation via measurements of the "B-mode" polarization of the CMB. Primordial B modes are directly tied to the amplitude of primordial gravitational waves (PGW), their strength parameterized by the tensor-to-scalar ratio, $r$, and thus the energy scale of inflation. Having set the most sensitive constraints to-date on $r$, $\sigma(r)=0.009$ ($r_{0.05}<0.036, 95\%$ C.L.) using data through the 2018 observing season ("BK18"), the BICEP/$\textit{Keck}$ program has continued to improve its dataset in the years since. We give a brief overview of the BK program and the "BK18" result before discussing the program's ongoing efforts, including the deployment and performance of the $\textit{Keck Array}$'s successor instrument, BICEP Array, improvements to data processing and internal consistency testing, new techniques such as delensing, and how those will ultimately serve to allow BK reach $\sigma(r) \lesssim 0.003$ using data through the 2027 observing season., Comment: 9 pages, 5 figures. Contribution to the 2024 Cosmology session of the 58th Rencontres de Moriond

Near-surface nitrogen-vacancy centres are critical to many diamond-based quantum technologies such as information processors and nanosensors. Surface defects play an important role in the design and performance of these devices. The targeted creation of defects is central to proposed bottom-up approaches to nanofabrication of quantum diamond processors, and uncontrolled surface defects may generate noise and charge trapping which degrade shallow NV device performance. Surface preparation protocols may be able to control the production of desired defects and eliminate unwanted defects, but only if their atomic structure can first be conclusively identified. This work uses a combination of scanning tunnelling microscopy (STM) imaging and first-principles simulations to identify several surface defects on H:C(100)-2x1 surfaces prepared using chemical vapour deposition (CVD). The atomic structure of these defects is elucidated, from which the microscopic origins of magnetic noise and charge trapping is determined based on modelling of their paramagnetic properties and acceptor states. Rudimentary control of these deleterious properties is demonstrated through STM tip-induced manipulation of the defect structure. Furthermore, the results validate accepted models for CVD diamond growth by identifying key adsorbates responsible for nucleation of new layers., Comment: 15 pages, 14 figures, combined manuscript and supporting information

Most popular benchmarks for comparing LLMs rely on a limited set of prompt templates, which may not fully capture the LLMs' abilities and can affect the reproducibility of results on leaderboards. Many recent works empirically verify prompt sensitivity and advocate for changes in LLM evaluation. In this paper, we consider the problem of estimating the performance distribution across many prompt variants instead of finding a single prompt to evaluate with. We introduce PromptEval, a method for estimating performance across a large set of prompts borrowing strength across prompts and examples to produce accurate estimates under practical evaluation budgets. The resulting distribution can be used to obtain performance quantiles to construct various robust performance metrics (e.g., top 95% quantile or median). We prove that PromptEval consistently estimates the performance distribution and demonstrate its efficacy empirically on three prominent LLM benchmarks: MMLU, BIG-bench Hard, and LMentry. For example, PromptEval can accurately estimate performance quantiles across 100 prompt templates on MMLU with a budget equivalent to two single-prompt evaluations. Our code and data can be found at https://github.com/felipemaiapolo/prompt-eval.

The duality of Jackiw-Teitelboim (JT) gravity and a double scaled matrix integral has led to studies of the canonical spectral form factor (SFF) in the so called $\tau-$scaled limit of large times, $t \to \infty$, and fixed temperature in order to demonstrate agreement with universal random matrix theory (RMT). Though this has been established for the unitary case, extensions to other symmetry classes requires the inclusion of unorientable manifolds in the sum over geometries, necessary to address time reversal invariance, and regularization of the corresponding prime geometrical objects, the Weil-Petersson (WP) volumes. We report here how universal signatures of quantum chaos, witnessed by the fidelity to the Gaussian orthogonal ensemble, emerge for the low-energy limit of unorientable JT gravity, i.e. the Airy model/topological gravity. To this end, we implement the loop equations for the corresponding dual (double-scaled) matrix model and find the generic form of the Airy WP volumes, supported by calculations using unorientable Kontsevich graphs. In an apparent violation of the gravity/chaos duality, the $\tau-$scaled SFF on the gravity side acquires both logarithmic and power law contributions in $t$, not manifestly present on the RMT side. We show the expressions can be made to agree by means of bootstrapping-like relations hidden in the asymptotic expansions of generalized hypergeometric functions. Thus, we are able to establish strong evidence of the quantum chaotic nature of unorientable topological gravity., Comment: version accepted for publication in JHEP, minor changes of text and references to the previous version, 34 pages+appendices

The rapid proliferation of generative AI has raised questions about the competitiveness of lower-parameter, locally tunable, open-weight models relative to high-parameter, API-guarded, closed-weight models in terms of performance, domain adaptation, cost, and generalization. Centering under-resourced yet risk-intolerant settings in government, research, and healthcare, we see for-profit closed-weight models as incompatible with requirements for transparency, privacy, adaptability, and standards of evidence. Yet the performance penalty in using open-weight models, especially in low-data and low-resource settings, is unclear. We assess the feasibility of using smaller, open-weight models to replace GPT-4-Turbo in zero-shot, few-shot, and fine-tuned regimes, assuming access to only a single, low-cost GPU. We assess value-sensitive issues around bias, privacy, and abstention on three additional tasks relevant to those topics. We find that with relatively low effort, very low absolute monetary cost, and relatively little data for fine-tuning, small open-weight models can achieve competitive performance in domain-adapted tasks without sacrificing generality. We then run experiments considering practical issues in bias, privacy, and hallucination risk, finding that open models offer several benefits over closed models. We intend this work as a case study in understanding the opportunity cost of reproducibility and transparency over for-profit state-of-the-art zero shot performance, finding this cost to be marginal under realistic settings., Comment: Accepted at the ACM Conference on Fairness, Accountability, and Transparency (FAccT) 2024

In this work, we consider data association problems involving multi-object tracking (MOT). In particular, we address the challenges arising from object occlusions. We propose a framework called approximate dynamic programming track (ADPTrack), which applies dynamic programming principles to improve an existing method called the base heuristic. Given a set of tracks and the next target frame, the base heuristic extends the tracks by matching them to the objects of this target frame directly. In contrast, ADPTrack first processes a few subsequent frames and applies the base heuristic starting from the next target frame to obtain tentative tracks. It then leverages the tentative tracks to match the objects of the target frame. This tends to reduce the occlusion-based errors and leads to an improvement over the base heuristic. When tested on the MOT17 video dataset, the proposed method demonstrates a 0.7% improvement in the association accuracy (IDF1 metric) over a state-of-the-art method that is used as the base heuristic. It also obtains improvements with respect to all the other standard metrics. Empirically, we found that the improvements are particularly pronounced in scenarios where the video data is obtained by fixed-position cameras.

High-harmonic generation is a light up-conversion process occurring in a strong laser field, leading to coherent bursts of extreme ultrashort broadband radiation [1]. As a new perspective, we propose that ultrafast strong-field electronic or photonic processes such as high-harmonic generation can potentially generate non-classical states of light well before the decoherence of the system occurs [2, 3]. This could address fundamental challenges in quantum technology such as scalability, decoherence or the generation of massively entangled states [4]. Here, we report experimental evidence of the non-classical nature of the harmonic emission in several semiconductors excited by a femtosecond infrared laser. By investigating single- and double beam intensity cross-correlation [5], we measure characteristic, non-classical features in the single photon statistics. We observe two-mode squeezing in the generated harmonic radiation, which depends on the laser intensity that governs the transition from Super-Poissonian to Poissonian photon statistics. The measured violation of the Cauchy-Schwarz inequality realizes a direct test of multipartite entanglement in high-harmonic generation [6]. This result is supported by the theory of multimodal detection and the Hamiltonian from which the effective squeezing modes of the harmonics can be derived [7, 8]. With this work, we show experimentally that high-harmonic generation is a new quantum bosonic platform that intrinsically produces non-classical states of light with unique features such as multipartite broadband entanglement or multimode squeezing. The source operates at room temperature using standard semiconductors and a standard commercial fiber laser, opening new routes for the quantum industry, such as optical quantum computing, communication and imaging., Comment: 39 pages, 11 figures

Language-conditioned robotic skills make it possible to apply the high-level reasoning of Large Language Models (LLMs) to low-level robotic control. A remaining challenge is to acquire a diverse set of fundamental skills. Existing approaches either manually decompose a complex task into atomic robotic actions in a top-down fashion, or bootstrap as many combinations as possible in a bottom-up fashion to cover a wider range of task possibilities. These decompositions or combinations, however, require an initial skill library. For example, a ``grasping'' capability can never emerge from a skill library containing only diverse ``pushing'' skills. Existing skill discovery techniques with reinforcement learning acquire skills by an exhaustive exploration but often yield non-meaningful behaviors. In this study, we introduce a novel framework for skill discovery that is entirely driven by LLMs. The framework begins with an LLM generating task proposals based on the provided scene description and the robot's configurations, aiming to incrementally acquire new skills upon task completion. For each proposed task, a series of reinforcement learning processes are initiated, utilizing reward and success determination functions sampled by the LLM to develop the corresponding policy. The reliability and trustworthiness of learned behaviors are further ensured by an independent vision-language model. We show that starting with zero skill, the skill library emerges and expands to more and more meaningful and reliable skills, enabling the robot to efficiently further propose and complete advanced tasks. Project page: \url{https://agentic-skill-discovery.github.io}., Comment: Webpage see https://agentic-skill-discovery.github.io/

The ability to perform fast and accurate atomistic simulations is crucial for advancing the chemical sciences. By learning from high-quality data, machine-learned interatomic potentials achieve accuracy on par with ab initio and first-principles methods at a fraction of their computational cost. The success of machine-learned interatomic potentials arises from integrating inductive biases such as equivariance to group actions on an atomic system, e.g., equivariance to rotations and reflections. In particular, the field has notably advanced with the emergence of equivariant message-passing architectures. Most of these models represent an atomic system using spherical tensors, tensor products of which require complicated numerical coefficients and can be computationally demanding. This work introduces higher-rank irreducible Cartesian tensors as an alternative to spherical tensors, addressing the above limitations. We integrate irreducible Cartesian tensor products into message-passing neural networks and prove the equivariance of the resulting layers. Through empirical evaluations on various benchmark data sets, we consistently observe on-par or better performance than that of state-of-the-art spherical models.

Fiber-based distribution of triggered, entangled, single-photon pairs is a key requirement for the future development of terrestrial quantum networks. In this context, semiconductor quantum dots (QDs) are promising candidates for deterministic sources of on-demand polarization-entangled photon pairs. So far, the best QD polarization-entangled-pair sources emit in the near-infrared wavelength regime, where the transmission distance in deployed fibers is limited. Here, to be compatible with existing fiber network infrastructures, bi-directional polarization-conserving quantum frequency conversion (QFC) is employed to convert the QD emission from \unit[780]{nm} to telecom wavelengths. We show the preservation of polarization entanglement after QFC (fidelity to Bell state $F_{\phi^+, conv}=0.972\pm0.003$) of the biexciton transition. As a step towards real-world applicability, high entanglement fidelities ($F_{\phi^+, loop}=0.945\pm0.005$) after the propagation of one photon of the entangled pair along a \unit[35.8]{km} field installed standard single mode fiber link are reported. Furthermore, we successfully demonstrate a second polarization-conversing QFC step back to \unit[780]{nm} preserving entanglement ($F_{\phi^+, back}=0.903\pm0.005$). This further prepares the way for interfacing quantum light to various quantum memories.

Advances in latent diffusion models (LDMs) have revolutionized high-resolution image generation, but the design space of the autoencoder that is central to these systems remains underexplored. In this paper, we introduce LiteVAE, a family of autoencoders for LDMs that leverage the 2D discrete wavelet transform to enhance scalability and computational efficiency over standard variational autoencoders (VAEs) with no sacrifice in output quality. We also investigate the training methodologies and the decoder architecture of LiteVAE and propose several enhancements that improve the training dynamics and reconstruction quality. Our base LiteVAE model matches the quality of the established VAEs in current LDMs with a six-fold reduction in encoder parameters, leading to faster training and lower GPU memory requirements, while our larger model outperforms VAEs of comparable complexity across all evaluated metrics (rFID, LPIPS, PSNR, and SSIM).

The Super-Kamiokande and T2K collaborations present a joint measurement of neutrino oscillation parameters from their atmospheric and beam neutrino data. It uses a common interaction model for events overlapping in neutrino energy and correlated detector systematic uncertainties between the two datasets, which are found to be compatible. Using 3244.4 days of atmospheric data and a beam exposure of $19.7(16.3) \times 10^{20}$ protons on target in (anti)neutrino mode, the analysis finds a 1.9$\sigma$ exclusion of CP-conservation (defined as $J_{CP}=0$) and a preference for the normal mass ordering., Comment: 10 pages, 3 figures

The spread of microbial infections is governed by the self-organization of bacteria on surfaces. Limitations of live imaging techniques make collective behaviors in clinically relevant systems challenging to quantify. Here, novel experimental and image analysis techniques for high-fidelity single-cell segmentation of bacterial colonies are developed. Machine learning-based segmentation models are trained solely using synthetic microscopy images that are processed to look realistic using state-of-the-art image-to-image translation methods, requiring no biophysical modeling. Accurate single-cell segmentation is achieved for densely packed single-species colonies and multi-species colonies of common pathogenic bacteria, even under suboptimal imaging conditions and for both brightfield and confocal laser scanning microscopy. The resulting data provide quantitative insights into the self-organization of bacteria on soft surfaces. Thanks to their high adaptability and relatively simple implementation, these methods promise to greatly facilitate quantitative descriptions of bacterial infections in varied environments., Comment: 12 pages, 5 figures

Recent quasar absorption line observations suggest that reionization may end as late as $z \approx 5.3$. As a means to search for large neutral hydrogen islands at $z<6$, we revisit long dark gaps in the Ly$\beta$ forest in VLT/X-Shooter and Keck/ESI quasar spectra. We stack the Ly$\alpha$ forest corresponding to both edges of these Ly$\beta$ dark gaps and identify a damping wing-like extended absorption profile. The average redshift of the stacked forest is $z=5.8$. By comparing these observations with reionization simulations, we infer that such a damping wing-like feature can be naturally explained if these gaps are at least partially created by neutral islands. Conversely, simulated dark gaps lacking neutral hydrogen struggle to replicate the observed damping wing features. Furthermore, this damping wing-like profile implies that the volume-averaged neutral hydrogen fraction must be $\langle x_{\rm HI} \rangle \geq 6.1 \pm 3.9\%$ at $z = 5.8$. Our results offer robust evidence that reionization extends below $z=6$., Comment: 8 pages, 5 figures, 1 table; accepted for publication in MNRAS Letters

We partially characterize nuclearity for the recently introduced class of hypergraph C*-algebras using a tailor-made hypergraph minor relation. The latter is generated by certain operations on hypergraphs which resemble the moves on directed graphs used by Eilers, Restorff, Ruiz and S{\o}rensen to classify unital graph C*-algebras. In particular, we obtain a new proof of the fact that every graph C*-algebra associated to a finite graph is nuclear.

In this work, we recover the underlying 3D structure of non-geometrically consistent scenes. We focus our analysis on hand-drawn images from cartoons and anime. Many cartoons are created by artists without a 3D rendering engine, which means that any new image of a scene is hand-drawn. The hand-drawn images are usually faithful representations of the world, but only in a qualitative sense, since it is difficult for humans to draw multiple perspectives of an object or scene 3D consistently. Nevertheless, people can easily perceive 3D scenes from inconsistent inputs! In this work, we correct for 2D drawing inconsistencies to recover a plausible 3D structure such that the newly warped drawings are consistent with each other. Our pipeline consists of a user-friendly annotation tool, camera pose estimation, and image deformation to recover a dense structure. Our method warps images to obey a perspective camera model, enabling our aligned results to be plugged into novel-view synthesis reconstruction methods to experience cartoons from viewpoints never drawn before. Our project page is https://toon3d.studio ., Comment: Please see our project page: https://toon3d.studio

We give a definition of hypergraph C*-algebras. These generalize the well-known graph C*-algebras as well as ultragraph C*-algebras. In contrast to those objects, hypergraph C*-algebras are not always nuclear. We provide a number of non-nuclear examples, we prove a Gauge-Invariant Uniqueness Theorem for a subclass of hypergraph C*-algebras and we study moves on hypergraphs which generalize the moves in the theory of graph C*-algebras., Comment: 30 pages + appendix; v2: updates in the references

The IoT and Business Process Management (BPM) communities co-exist in many shared application domains, such as manufacturing and healthcare. The IoT community has a strong focus on hardware, connectivity and data; the BPM community focuses mainly on finding, controlling, and enhancing the structured interactions among the IoT devices in processes. While the field of Process Mining deals with the extraction of process models and process analytics from process event logs, the data produced by IoT sensors often is at a lower granularity than these process-level events. The fundamental questions about extracting and abstracting process-related data from streams of IoT sensor values are: (1) Which sensor values can be clustered together as part of process events?, (2) Which sensor values signify the start and end of such events?, (3) Which sensor values are related but not essential? This work proposes a framework to semi-automatically perform a set of structured steps to convert low-level IoT sensor data into higher-level process events that are suitable for process mining. The framework is meant to provide a generic sequence of abstract steps to guide the event extraction, abstraction, and correlation, with variation points for plugging in specific analysis techniques and algorithms for each step. To assess the completeness of the framework, we present a set of challenges, how they can be tackled through the framework, and an example on how to instantiate the framework in a real-world demonstration from the field of smart manufacturing. Based on this framework, future research can be conducted in a structured manner through refining and improving individual steps.

The development of generative artificial intelligence for human motion generation has expanded rapidly, necessitating a unified evaluation framework. This paper presents a detailed review of eight evaluation metrics for human motion generation, highlighting their unique features and shortcomings. We propose standardized practices through a unified evaluation setup to facilitate consistent model comparisons. Additionally, we introduce a novel metric that assesses diversity in temporal distortion by analyzing warping diversity, thereby enhancing the evaluation of temporal data. We also conduct experimental analyses of three generative models using a publicly available dataset, offering insights into the interpretation of each metric in specific case scenarios. Our goal is to offer a clear, user-friendly evaluation framework for newcomers, complemented by publicly accessible code.

We present an in-depth analysis of the LISA Pathfinder differential acceleration performance over the entire course of its science operations, spanning approximately 500 days. We find that: 1) the evolution of the Brownian noise that dominates the acceleration amplitude spectral density (ASD), for frequencies $f\gtrsim 1\,\text{mHz}$, is consistent with the decaying pressure due to the outgassing of a single gaseous species. 2) between $f=36\,\mu\text{Hz}$ and $1\,\text{mHz}$, the acceleration ASD shows a $1/f$ tail in excess of the Brownian noise of almost constant amplitude, with $\simeq 20\%$ fluctuations over a period of a few days, with no particular time pattern over the course of the mission; 3) at the lowest considered frequency of $f=18\,\mu\text{Hz}$, the ASD significantly deviates from the $1/f$ behavior, because of temperature fluctuations that appear to modulate a quasi-static pressure gradient, sustained by the asymmetries of the outgassing pattern. We also present the results of a projection of the observed acceleration noise on the potential sources for which we had either a direct correlation measurement, or a quantitative estimate from dedicated experiments. These sources account for approximately $40\%$ of the noise power in the $1/f$ tail. Finally, we analyze the possible sources of the remaining unexplained fraction, and identify the possible measures that may be taken to keep those under control in LISA.

Most Bundle Adjustment (BA) solvers like the Levenberg-Marquardt algorithm require a good initialization. Instead, initialization-free BA remains a largely uncharted territory. The under-explored Variable Projection algorithm (VarPro) exhibits a wide convergence basin even without initialization. Coupled with object space error formulation, recent works have shown its ability to solve small-scale initialization-free bundle adjustment problem. To make such initialization-free BA approaches scalable, we introduce Power Variable Projection (PoVar), extending a recent inverse expansion method based on power series. Importantly, we link the power series expansion to Riemannian manifold optimization. This projective framework is crucial to solve large-scale bundle adjustment problems without initialization. Using the real-world BAL dataset, we experimentally demonstrate that our solver achieves state-of-the-art results in terms of speed and accuracy. To our knowledge, this work is the first to address the scalability of BA without initialization opening new venues for initialization-free structure-from-motion.