Search

This technical report contains the full set of definitions and projection rules of the paper ``Verified Parameterized Choreographies'' by Rubbens et al. It also supplements the artefact.

Physical information processors can learn from examples if they are modified according to an abstract parameter update equation, termed a learning rule. We introduce a physical model for self-learning that encodes the learning rule in the Hamiltonian of the system. The model consists of a network of multi-modal resonators. One of the modes is driven parametrically into a bi-stable regime, forming a coherent Ising machine (CIM) -- that provides the long-term memory that stores learned responses (weights). The CIM is augmented with an additional spinor field that acts as short-term (activation) memory. We numerically demonstrate that, in the presence of suitable nonlinear interactions between the long-term memory Ising machine and the short-term memory auxiliary field, the system autonomously learns from examples., Comment: 8 pages, 4 figures

Cardiovascular disease (CVD) remains a leading cause of death, and primary prevention through personalized interventions is crucial. This paper introduces MyDigiTwin, a framework that integrates health digital twins with personal health environments to empower patients in exploring personalized health scenarios while ensuring data privacy. MyDigiTwin uses federated learning to train predictive models across distributed datasets without transferring raw data, and a novel data harmonization framework addresses semantic and format inconsistencies in health data. A proof-of-concept demonstrates the feasibility of harmonizing and using cohort data to train privacy-preserving CVD prediction models. This framework offers a scalable solution for proactive, personalized cardiovascular care and sets the stage for future applications in real-world healthcare settings.

Metals have high electronic conductivities, but very low Seebeck coefficients, which traditionally make them unsuitable for thermoelectric materials. Recent studies, however, showed that metals can deliver ultra-high thermoelectric power factors (PFs) under certain conditions. In this work, we theoretically examine the electronic structure and electronic transport specifications which allow for such high PFs. Using Boltzmann transport (BTE) simulations and a multi-band electronic structure model, we show that metals with: i) high degree of transport asymmetry between their bands, ii) strong inter-band scattering, and iii) a large degree of band overlap, can provide ultra-high power factors. We show that each of these characteristics adds to the steepness of the transport distribution function of the BTE, which allows for an increase of the Seebeck coefficient to sizable values, simultaneously with an increase in the electrical conductivity. This work generalizes the concept that transport asymmetry (i.e., mixture of energy regions of high and low contributions to the electrical conductivity), through a combination of different band masses, scattering strengths, or energy filtering scenarios, etc., can indeed result in very high thermoelectric power factors, even in the absence of a material bandgap. Under certain conditions, transport asymmetry can over-compensate any performance degradation to the PF due to bipolar conduction and the naturally low Seebeck coefficients that otherwise exist in this class of materials., Comment: paper and supporting information, 50 pages

Background: The MAJORANA DEMONSTRATOR , a modular array of isotopically enriched high-purity germanium (HPGe) detectors, was constructed to demonstrate backgrounds low enough to justify building a tonne-scale experiment to search for the neutrinoless double-beta decay ($\beta\beta(0\nu)$) of $^{76}\mathrm{Ge}$. Purpose: This paper presents a description of the instrument, its commissioning, and operations. It covers the electroforming, underground infrastructure, enrichment, detector fabrication, low-background and construction techniques, electronics, data acquisition, databases, and data processing of the MAJORANA DEMONSTRATOR. Method: The MAJORANA DEMONSTRATOR operated inside an ultra-low radioactivity passive shield at the 4850-foot~level of the Sanford Underground Research Facility (SURF) from 2015-2021. Results and Conclusions: The MAJORANA DEMONSTRATOR achieved the best energy resolution and second-best background level of any $\beta\beta(0\nu)$ search. This enabled it to achieve an ultimate half-life limit on $\beta\beta(0\nu)$ in $^{76}\mathrm{Ge}$ of $8.3\times 10^{25}$~yr (90\% C.L.) and perform a rich set of searches for other physics beyond the Standard Model., Comment: 72 pages

We present a multifractal artificial terrain generation method that uses the 3D Weierstrass-Mandelbrot function to control roughness. By varying the fractal dimension used in terrain generation across three different values, we generate 60 unique off-road terrains. We use gradient maps to categorize the roughness of each terrain, consisting of low-, semi-, and high-roughness areas. To test how the fractal dimension affects the difficulty of vehicle traversals, we measure the success rates, vertical accelerations, pitch and roll rates, and traversal times of an autonomous ground vehicle traversing 20 randomized straight-line paths in each terrain. As we increase the fractal dimension from 2.3 to 2.45 and from 2.45 to 2.6, we find that the median area of low-roughness terrain decreases 13.8% and 7.16%, the median area of semi-rough terrain increases 11.7% and 5.63%, and the median area of high-roughness terrain increases 1.54% and 3.33%, all respectively. We find that the median success rate of the vehicle decreases 22.5% and 25% as the fractal dimension increases from 2.3 to 2.45 and from 2.45 to 2.6, respectively. Successful traversal results show that the median root-mean-squared vertical accelerations, median root-mean-squared pitch and roll rates, and median traversal times all increase with the fractal dimension., Comment: This work has been accepted for publication in ASME Journal of Autonomous Vehicles and Systems

The Majorana Demonstrator was an ultra-low-background experiment designed for neutrinoless double-beta decay ($0\nu\beta\beta$) investigation in $^{76}$Ge. Located at the Sanford Underground Research Facility in Lead, South Dakota, the Demonstrator utilized modular high-purity Ge detector arrays within shielded vacuum cryostats, operating deep underground. The arrays, with a capacity of up to 40.4 kg (27.2 kg enriched to $\sim 88\%$ in $^{76}$Ge), have accumulated the full data set, totaling 64.5 kg yr of enriched active exposure and 27.4 kg yr of exposure for natural detectors. Our updated search improves previously explored three-nucleon decay modes in Ge isotopes, setting new half-life limits of $1.27\times10^{26}$ years (90\% confidence level) for $^{76}$Ge($ppp$) $\rightarrow$ $^{73}$Cu e$^+\pi^+\pi^+$ and $^{76}$Ge($ppn$) $\rightarrow$ $^{73}$Zn e$^+\pi^+$. The half-life limit for the invisible tri-proton decay mode of $^{76}$Ge is found to be $1.4\times10^{25}$ yr. Furthermore, we have updated limits for corresponding multi-nucleon decays.

Open-domain semantic parsing remains a challenging task, as models often rely on heuristics and struggle to handle unseen concepts. In this paper, we investigate the potential of large language models (LLMs) for this task and introduce Retrieval-Augmented Semantic Parsing (RASP), a simple yet effective approach that integrates external lexical knowledge into the parsing process. Our experiments not only show that LLMs outperform previous encoder-decoder baselines for semantic parsing, but that RASP further enhances their ability to predict unseen concepts, nearly doubling the performance of previous models on out-of-distribution concepts. These findings highlight the promise of leveraging large language models and retrieval mechanisms for robust and open-domain semantic parsing., Comment: Submitted to ARR

Extended reality technology has become a useful tool in many applications, but still suffers from visual deviations that can hamper the utility of the technology. This paper discusses the types of persisting visual deviations experienced when observing the natural world through video see-through head-mounted displays. A generalizable method to measure the effect of these deviations on real-world interaction is designed and used in a human-in-the-loop experiment. The experiment compared video see-through sight through an head-mounted display with normal eyesight in a static set-up, focusing on (camera) lens distortions and display deviations. Participants interacted with a real touchscreen, locating the position of flashed markers shortly after disappearance comparing both conditions to check for deviations in position and time. Results show significant larger mean distance errors between the interaction locations and the original marker positions for video see-through compared to normal eyesight. Moreover, errors increase towards the screen periphery. No significant distance error improvement over time was found, however, response times did significantly decrease for both types of sight., Comment: 10 pages. Preprint. A shortened 4-page version of this paper was accepted to the IEEE ISMAR2024 poster track

Neutrinoless double-beta decay ($0\nu\beta\beta$) is a rare nuclear process that, if observed, will provide insight into the nature of neutrinos and help explain the matter-antimatter asymmetry in the universe. The Large Enriched Germanium Experiment for Neutrinoless Double-Beta Decay (LEGEND) will operate in two phases to search for $0\nu\beta\beta$. The first (second) stage will employ 200 (1000) kg of High-Purity Germanium (HPGe) enriched in $^{76}$Ge to achieve a half-life sensitivity of 10$^{27}$ (10$^{28}$) years. In this study, we present a semi-supervised data-driven approach to remove non-physical events captured by HPGe detectors powered by a novel artificial intelligence model. We utilize Affinity Propagation to cluster waveform signals based on their shape and a Support Vector Machine to classify them into different categories. We train, optimize, test our model on data taken from a natural abundance HPGe detector installed in the Full Chain Test experimental stand at the University of North Carolina at Chapel Hill. We demonstrate that our model yields a maximum physics event sacrifice of $0.024 ^{+0.004}_{-0.003} \%$ when performing data cleaning cuts. Our model is being used to accelerate data cleaning development for LEGEND-200., Comment: 11 pages, 13 figures

$^{76}$Ge can $\beta\beta$ decay into three possible excited states of $^{76}$Se, with the emission of two or, if the neutrino is Majorana, zero neutrinos. None of these six transitions have yet been observed. The MAJORANA DEMONSTRATOR was designed to study $\beta\beta$ decay of $^{76}$Ge using a low background array of high purity germanium detectors. With 98.2 kg-y of isotopic exposure, the DEMONSTRATOR sets the strongest half-life limits to date for all six transition modes. For $2\nu\beta\beta$ to the $0^+_1$ state of $^{76}$Se, this search has begun to probe for the first time half-life values predicted using modern many-body nuclear theory techniques, setting a limit of $T_{1/2}>1.5\times10^{24}$ y (90% CL)., Comment: 8 pages, 3 figures

The background index is an important quantity which is used in projecting and calculating the half-life sensitivity of neutrinoless double-beta decay ($0\nu\beta\beta$) experiments. A novel analysis framework is presented to calculate the background index using the specific activities, masses and simulated efficiencies of an experiment's components as distributions. This Bayesian framework includes a unified approach to combine specific activities from assay. Monte Carlo uncertainty propagation is used to build a background index distribution from the specific activity, mass and efficiency distributions. This analysis method is applied to the MAJORANA DEMONSTRATOR, which deployed arrays of high-purity Ge detectors enriched in $^{76}$Ge to search for $0\nu\beta\beta$. The framework projects a mean background index of $\left[8.95 \pm 0.36\right] \times 10^{-4}$cts/(keV kg yr) from $^{232}$Th and $^{238}$U in the DEMONSTRATOR's components., Comment: 9 pages, 3 figures

Circuits are supposed to accurately describe how a neural network performs a specific task, but do they really? We evaluate three circuits found in the literature (IOI, greater-than, and docstring) in an adversarial manner, considering inputs where the circuit's behavior maximally diverges from the full model. Concretely, we measure the KL divergence between the full model's output and the circuit's output, calculated through resample ablation, and we analyze the worst-performing inputs. Our results show that the circuits for the IOI and docstring tasks fail to behave similarly to the full model even on completely benign inputs from the original task, indicating that more robust circuits are needed for safety-critical applications., Comment: 19 pages, 10 figures

We performed closed-loop lab testing of large-format deformable mirrors (DMs) with hybrid variable reluctance actuators. TNO has been developing the hybrid variable reluctance actuators in support for a new generation of adaptive secondary mirrors (ASMs), which aim to be more robust and reliable. Compared to the voice coil actuators, this new actuator technology has a higher current to force efficiency, and thus can support DMs with thicker facesheets. Before putting this new technology on-sky, it is necessary to understand how to control it and how it behaves in closed-loop. We performed closed-loop tests with the Shack-Hartmann wavefront sensor with three large-format deformable mirrors that use the TNO actuators: DM3, FLASH, and IRTF-ASM-1 ASM. The wavefront sensor and the real-time control systems were developed for the NASA Infrared Telescope Facility (IRTF) and the UH 2.2-meter telescope ASMs. We tested IRTF-ASM-1 on-sky and proved that it meets all of our performance requirements. This work presents our lab setup for the experiments, the techniques we have employed to drive these new ASMs, the results of our closed-loop lab tests for FLASH and IRTF-ASM-1, and the on-sky closed-loop results of IRTF-ASM-1 ASM., Comment: 15 pages, 12 figures, SPIE Astronomical Telescopes + Instrumentation 2024

We are developing an adaptive secondary mirror (ASM) that uses a new actuator technology created by the Netherlands Organization for Applied Scientific Research (TNO). The TNO hybrid variable reluctance actuators have more than an order of magnitude better efficiency over the traditional voice coil actuators that have been used on existing ASMs and show potential for improving the long-term robustness and reliability of ASMs. To demonstrate the performance, operations, and serviceability of TNO's actuators in an observatory, we have developed a 36-actuator prototype ASM for the NASA Infrared Telescope Facility (IRTF) called IRTF-ASM-1. IRTF-ASM-1 provides the first on-sky demonstration of this approach and will help us evaluate the long-term performance and use of this technology in an astronomical facility environment. We present calibration and performance results with the ASM in a Meniscus Hindle Sphere lens setup as well as preliminary on-sky results on IRTF. IRTF-ASM-1 achieved stable closed-loop performance on-sky with H-band Strehl ratios of 35-40% in long-exposure images under a variety of seeing conditions., Comment: 15 pages, 12 figures, SPIE Astronomical Telescopes + Instrumentation: Adaptive Optics Systems IX

Discourse Representation Theory (DRT) distinguishes itself from other semantic representation frameworks by its ability to model complex semantic and discourse phenomena through structural nesting and variable binding. While seq2seq models hold the state of the art on DRT parsing, their accuracy degrades with the complexity of the sentence, and they sometimes struggle to produce well-formed DRT representations. We introduce the AMS parser, a compositional, neurosymbolic semantic parser for DRT. It rests on a novel mechanism for predicting quantifier scope. We show that the AMS parser reliably produces well-formed outputs and performs well on DRT parsing, especially on complex sentences.

For many years, designs for "Neuromorphic" or brain-like processors have been motivated by achieving extreme energy efficiency, compared with von-Neumann and tensor processor devices. As part of their design language, Neuromorphic processors take advantage of weight, parameter, state and activity sparsity. In the extreme case, neural networks based on these principles mimic the sparse activity oof biological nervous systems, in ``Spiking Neural Networks'' (SNNs). Few benchmarks are available for Neuromorphic processors, that have been implemented for a range of Neuromorphic and non-Neuromorphic platforms, which can therefore demonstrate the energy benefits of Neuromorphic processor designs. Here we describes the implementation of a spoken audio keyword-spotting (KWS) benchmark "Aloha" on the Xylo Audio 2 (SYNS61210) Neuromorphic processor device. We obtained high deployed quantized task accuracy, (95%), exceeding the benchmark task accuracy. We measured real continuous power of the deployed application on Xylo. We obtained best-in-class dynamic inference power ($291\mu$W) and best-in-class inference efficiency ($6.6\mu$J / Inf). Xylo sets a new minimum power for the Aloha KWS benchmark, and highlights the extreme energy efficiency achievable with Neuromorphic processor designs. Our results show that Neuromorphic designs are well-suited for real-time near- and in-sensor processing on edge devices.

Content moderators play a key role in keeping the conversation on social media healthy. While the high volume of content they need to judge represents a bottleneck to the moderation pipeline, no studies have explored how models could support them to make faster decisions. There is, by now, a vast body of research into detecting hate speech, sometimes explicitly motivated by a desire to help improve content moderation, but published research using real content moderators is scarce. In this work we investigate the effect of explanations on the speed of real-world moderators. Our experiments show that while generic explanations do not affect their speed and are often ignored, structured explanations lower moderators' decision making time by 7.4%., Comment: 11 pages, 14 figures, to be published at ACL 2024

We give a holomorphic quartic polynomial in the overlap variables whose zeros on the torus are precisely the Weyl-Heisenberg SICs (symmetric informationally complete positive operator valued measures). By way of comparison, all the other known systems of equations that determine a Weyl-Heisenberg SIC involve variables and their complex conjugates. We also give a related interesting result about the powers of the projective Fourier transform of the group G = Z d x Z d ., Comment: 19 Pages

Current open-domain neural semantics parsers show impressive performance. However, closer inspection of the symbolic meaning representations they produce reveals significant weaknesses: sometimes they tend to merely copy character sequences from the source text to form symbolic concepts, defaulting to the most frequent word sense based in the training distribution. By leveraging the hierarchical structure of a lexical ontology, we introduce a novel compositional symbolic representation for concepts based on their position in the taxonomical hierarchy. This representation provides richer semantic information and enhances interpretability. We introduce a neural "taxonomical" semantic parser to utilize this new representation system of predicates, and compare it with a standard neural semantic parser trained on the traditional meaning representation format, employing a novel challenge set and evaluation metric for evaluation. Our experimental findings demonstrate that the taxonomical model, trained on much richer and complex meaning representations, is slightly subordinate in performance to the traditional model using the standard metrics for evaluation, but outperforms it when dealing with out-of-vocabulary concepts. This finding is encouraging for research in computational semantics that aims to combine data-driven distributional meanings with knowledge-based symbolic representations., Comment: This manuscript has been accepted by Computational Linguistics journal on 2024-09-07

The Parallel Meaning Bank (PMB) serves as a corpus for semantic processing with a focus on semantic parsing and text generation. Currently, we witness an excellent performance of neural parsers and generators on the PMB. This might suggest that such semantic processing tasks have by and large been solved. We argue that this is not the case and that performance scores from the past on the PMB are inflated by non-optimal data splits and test sets that are too easy. In response, we introduce several changes. First, instead of the prior random split, we propose a more systematic splitting approach to improve the reliability of the standard test data. Second, except for the standard test set, we also propose two challenge sets: one with longer texts including discourse structure, and one that addresses compositional generalization. We evaluate five neural models for semantic parsing and meaning-to-text generation. Our results show that model performance declines (in some cases dramatically) on the challenge sets, revealing the limitations of neural models when confronting such challenges.

User embeddings play a crucial role in user engagement forecasting and personalized services. Recent advances in sequence modeling have sparked interest in learning user embeddings from behavioral data. Yet behavior-based user embedding learning faces the unique challenge of dynamic user modeling. As users continuously interact with the apps, user embeddings should be periodically updated to account for users' recent and long-term behavior patterns. Existing methods highly rely on stateless sequence models that lack memory of historical behavior. They have to either discard historical data and use only the most recent data or reprocess the old and new data jointly. Both cases incur substantial computational overhead. To address this limitation, we introduce User Stateful Embedding (USE). USE generates user embeddings and reflects users' evolving behaviors without the need for exhaustive reprocessing by storing previous model states and revisiting them in the future. Furthermore, we introduce a novel training objective named future W-behavior prediction to transcend the limitations of next-token prediction by forecasting a broader horizon of upcoming user behaviors. By combining it with the Same User Prediction, a contrastive learning-based objective that predicts whether different segments of behavior sequences belong to the same user, we further improve the embeddings' distinctiveness and representativeness. We conducted experiments on 8 downstream tasks using Snapchat users' behavioral logs in both static (i.e., fixed user behavior sequences) and dynamic (i.e., periodically updated user behavior sequences) settings. We demonstrate USE's superior performance over established baselines. The results underscore USE's effectiveness and efficiency in integrating historical and recent user behavior sequences into user embeddings in dynamic user modeling.

We prove that, given an average power, the ascent time is minimized if a cyclist maintains a constant ground speed regardless of the slope. Herein, minimizing the time is equivalent to maximizing -- for a given uphill -- the corresponding mean ascent velocity (VAM: velocit\`a ascensionale media), which is a common training metric. We illustrate the proof with numerical examples, and show that, in general, maintaining a constant instantaneous power results in longer ascent times; both strategies result in the same time if the slope is constant. To remain within the athlete's capacity, we examine the effect of complementing the average-power constraint with a maximum-power constraint. Even with this additional constraint, the ascent time is the shortest with a modified constant-speed -- not constant-power -- strategy; as expected, both strategies result in the same time if the maximum and average powers are equal to one another. Given standard available information -- including level of fitness, quantified by the power output, and ascent profile -- our results allow to formulate reliable and convenient strategies of uphill timetrials.