Search

Experimental searches for axions or dark photons that couple to the standard model photon require photosensors with low noise, broadband sensitivity, and near zero backgrounds. Here, we introduce a Transition Edge Sensor (TES) based photon sensor with a high sensitivity athermal phonon sensor coupled to the substrate on which the TES is deposited. We show that single photons absorbed locally in the TES have $\sim$35% energy deposition in the electronic system of the TES, with $\sim$26% of the photon energy absorbed by the athermal photon sensor (due to athermal phonons leaking out of the TES during the downconversion process). Backgrounds are observed to be largely coupled to either the TES or phonon system. At high energies, these backgrounds can be efficiently discriminated from TES photon absorption events, while at low energies, their leakage into the passage band is well modeled. With significant sensitivity improvements to both the photon and phonon channel, this coincidence technique could be used to suppress backgrounds in bosonic dark matter searches down to energies near the superconducting bandgap of the sensor., Comment: 5 pages, 5 figures

The structural evolution of oxides in dispersion-strengthened superalloys during laser-powder bed fusion is considered in detail. Alloy chemistry and process parameter effects on oxide structure are assessed through a parameter study on the model alloy Ni-20Cr, doped with varying concentrations of Y2O3 and Al. A scaling analysis of mass and momentum transport within the melt pool, presented here, establishes that diffusional structural evolution mechanisms dominate for nanoscale dispersoids, while fluid forces and advection become significant for larger micron-scale slag inclusions. These findings are developed into a theory of dispersoid structural evolution, integrating quantitative models of diffusional processes -- dispersoid dissolution, nucleation, growth, coarsening -- with a reduced order model of time-temperature trajectories of fluid parcels within the melt pool. Calculations of the dispersoid size in single-pass melting reveal a zone in the center of the melt track in which the oxide feedstock fully dissolves. Within this zone the final Y2O3 size is independent of feedstock size and determined by nucleation and growth kinetics. If the dissolution zones of adjacent melt tracks overlap sufficiently with each other to dissolve large oxides, formed during printing or present in the powder feedstock, then the dispersoid structure throughout the build volume is homogeneous and matches that from a single pass within the dissolution zone. Gaps between adjacent dissolution zones result in oxide accumulation into larger slag inclusions. Predictions of final dispersoid size and slag formation using this dissolution zone model match the present experimental data and explain process-structure linkages speculated in the open literature.

This work reports the procedure for modeling piezoelectric acoustic resonators and filters at millimeter wave (mmWave). Different from conventional methods for lower frequency piezoelectric devices, we include both acoustic and electromagnetic (EM) effects, e.g., self-inductance, in both the circuit-level fitting and finite element analysis, toward higher accuracy at higher frequencies. To validate the method, thin-film lithium niobate (LiNbO3) first-order antisymmetric (A1) mode devices are used as the testbed, achieving great agreement for both the standalone resonators and a fifth-order ladder filter. Upon further development, the reported acoustic and EM co-modeling could guide the future design of compact piezoelectric devices at mmWave and beyond., Comment: 6 pages, 11 figures, submitted to IEEE JMEMS

Quantum embedding with transformers is a novel and promising architecture for quantum machine learning to deliver exceptional capability on near-term devices or simulators. The research incorporated a vision transformer (ViT) to advance quantum significantly embedding ability and results for a single qubit classifier with around 3 percent in the median F1 score on the BirdCLEF-2021, a challenging high-dimensional dataset. The study showcases and analyzes empirical evidence that our transformer-based architecture is a highly versatile and practical approach to modern quantum machine learning problems.

Microwave Kinetic Inductance Detectors (MKIDs) have been demonstrated as capable phonon sensors when coupled to crystalline substrates, and have been proposed as detectors for next-generation rare-event searches such as for the direct detection of dark matter. These Kinetic Inductance Phonon Mediated (KIPM) detector designs, favoring large superconducting absorber volumes and high readout powers, are oftentimes limited in their sensitivity by low temperature amplifier noise introduced in the signal readout chain. We report here an effort to couple a wideband Kinetic Inductance Travelling Wave Parametric Amplifier (KI-TWPA), operated near the Standard Quantum Limit of minimal added amplifier noise, to sensors spanning a 70 MHz bandwidth at 3.5 GHz. This results in a ~5x improvement in the inferred detector energy resolution in the best sensor and highlights the potential of constructing O(100) meV resolving phonon-mediated particle detectors. We detail limitations introduced by lossy passive components, degraded RF responsivity, and microphysical noise sources like two-level systems (TLS), in achieving ultimate quantum-limited system noise levels., Comment: 7 pages, 4 figures

Microcalorimeters that leverage microwave kinetic inductance detectors to read out phonon signals in the particle-absorbing target, referred to as kinetic inductance phonon-mediated (KIPM) detectors, offer an attractive detector architecture to probe dark matter (DM) down to the fermionic thermal relic mass limit. A prototype KIPM detector featuring a single aluminum resonator patterned onto a 1-gram silicon substrate was operated in the NEXUS low-background facility at Fermilab for characterization and evaluation of this detector architecture's efficacy for a dark matter search. An energy calibration was performed by exposing the bare substrate to a pulsed source of 470 nm photons, resulting in a baseline resolution on the energy absorbed by the phonon sensor of $2.1\pm0.2$ eV, a factor of two better than the current state-of-the-art, enabled by millisecond-scale quasiparticle lifetimes. However, due to the sub-percent phonon collection efficiency, the resolution on energy deposited in the substrate is limited to $\sigma_E=318 \pm 28$ eV. We further model the signal pulse shape as a function of device temperature to extract quasiparticle lifetimes, as well as the observed noise spectra, both of which impact the baseline resolution of the sensor.

The research explores the potential of quantum deep learning models to address challenging machine learning problems that classical deep learning models find difficult to tackle. We introduce a novel model architecture that combines classical convolutional layers with a quantum neural network, aiming to surpass state-of-the-art accuracy while maintaining a compact model size. The experiment is to classify high-dimensional audio data from the Bird-CLEF 2021 dataset. Our evaluation focuses on key metrics, including training duration, model accuracy, and total model size. This research demonstrates the promising potential of quantum machine learning in enhancing machine learning tasks and solving practical machine learning challenges available today.

Quantum state preparation is vital in quantum computing and information processing. The ability to accurately and reliably prepare specific quantum states is essential for various applications. One of the promising applications of quantum computers is quantum simulation. This requires preparing a quantum state representing the system we are trying to simulate. This research introduces a novel simulation algorithm, the multi-Split-Steps Quantum Walk (multi-SSQW), designed to learn and load complicated probability distributions using parameterized quantum circuits (PQC) with a variational solver on classical simulators. The multi-SSQW algorithm is a modified version of the split-steps quantum walk, enhanced to incorporate a multi-agent decision-making process, rendering it suitable for modeling financial markets. The study provides theoretical descriptions and empirical investigations of the multi-SSQW algorithm to demonstrate its promising capabilities in probability distribution simulation and financial market modeling. Harnessing the advantages of quantum computation, the multi-SSQW models complex financial distributions and scenarios with high accuracy, providing valuable insights and mechanisms for financial analysis and decision-making. The multi-SSQW's key benefits include its modeling flexibility, stable convergence, and instantaneous computation. These advantages underscore its rapid modeling and prediction potential in dynamic financial markets., Comment: 13 pages, 13 figures. arXiv admin note: text overlap with arXiv:2302.12500

This work reports an acoustic filter at 23.5 GHz with a low insertion loss (IL) of 2.38 dB and a 3-dB fractional bandwidth (FBW) of 18.2%, significantly surpassing the state-of-the-art. The device leverages electrically coupled acoustic resonators in 100 nm 128{\deg} Y-cut lithium niobate (LiNbO3) piezoelectric thin film, operating in the first-order antisymmetric (A1) mode. A new film stack, namely transferred thin-film LiNbO3 on silicon (Si) substrate with an intermediate amorphous silicon (a-Si) layer, facilitates the record-breaking performance at millimeter-wave (mmWave). The filter features a compact footprint of 0.56 mm2. In this letter, acoustic and EM consideration, along with material characterization with X-ray diffraction and verified with cross-sectional electron microscopy are reported. Upon further development, the reported filter platform can enable various front-end signal-processing functions at mmWave., Comment: 4 pages, 8 figures, submitted to IEEE JMEMS

We present a new test-time optimization method for estimating dense and long-range motion from a video sequence. Prior optical flow or particle video tracking algorithms typically operate within limited temporal windows, struggling to track through occlusions and maintain global consistency of estimated motion trajectories. We propose a complete and globally consistent motion representation, dubbed OmniMotion, that allows for accurate, full-length motion estimation of every pixel in a video. OmniMotion represents a video using a quasi-3D canonical volume and performs pixel-wise tracking via bijections between local and canonical space. This representation allows us to ensure global consistency, track through occlusions, and model any combination of camera and object motion. Extensive evaluations on the TAP-Vid benchmark and real-world footage show that our approach outperforms prior state-of-the-art methods by a large margin both quantitatively and qualitatively. See our project page for more results: http://omnimotion.github.io/, Comment: ICCV 2023

Quantum computing holds great potential for advancing the limitations of machine learning algorithms to handle higher dimensions of data and reduce overall training parameters in deep learning (DL) models. This study uses a trainable variational quantum circuit (VQC) on a gate-based quantum computing model to investigate the potential for quantum benefit in a model-free reinforcement learning problem. Through a comprehensive investigation and evaluation of the current model and capabilities of quantum computers, we designed and trained a novel hybrid quantum neural network based on the latest Qiskit and PyTorch framework. We compared its performance with a full-classical CNN with and without an incorporated VQC. Our research provides insights into the potential of deep quantum learning to solve a maze problem and, potentially, other reinforcement learning problems. We conclude that reinforcement learning problems can be practical with reasonable training epochs. Moreover, a comparative study of full-classical and hybrid quantum neural networks is discussed to understand these two approaches' performance, advantages, and disadvantages to deep-Q learning problems, especially on larger-scale maze problems larger than 4x4.

Photorealistic object appearance modeling from 2D images is a constant topic in vision and graphics. While neural implicit methods (such as Neural Radiance Fields) have shown high-fidelity view synthesis results, they cannot relight the captured objects. More recent neural inverse rendering approaches have enabled object relighting, but they represent surface properties as simple BRDFs, and therefore cannot handle translucent objects. We propose Object-Centric Neural Scattering Functions (OSFs) for learning to reconstruct object appearance from only images. OSFs not only support free-viewpoint object relighting, but also can model both opaque and translucent objects. While accurately modeling subsurface light transport for translucent objects can be highly complex and even intractable for neural methods, OSFs learn to approximate the radiance transfer from a distant light to an outgoing direction at any spatial location. This approximation avoids explicitly modeling complex subsurface scattering, making learning a neural implicit model tractable. Experiments on real and synthetic data show that OSFs accurately reconstruct appearances for both opaque and translucent objects, allowing faithful free-viewpoint relighting as well as scene composition., Comment: Journal extension of arXiv:2012.08503 (TMLR 2023). The first two authors contributed equally to this work. Project page: https://kovenyu.com/osf/

Understanding and control of Laser-driven Free Electron Lasers remain to be difficult problems that require highly intensive experimental and theoretical research. The gap between simulated and experimentally collected data might complicate studies and interpretation of obtained results. In this work we developed a deep learning based surrogate that could help to fill in this gap. We introduce a surrogate model based on normalising flows for conditional phase-space representation of electron clouds in a FEL beamline. Achieved results let us discuss further benefits and limitations in exploitability of the models to gain deeper understanding of fundamental processes within a beamline., Comment: 7 pages, 5 figures

In recent years, there has been an emerging trend of combining two innovations in computer science and physics to achieve better computation capability. Exploring the potential of quantum computation to achieve highly efficient performance in various tasks is a vital development in engineering and a valuable question in sciences, as it has a significant potential to provide exponential speedups for technologically complex problems that are specifically advantageous to quantum computers. However, one key issue in unleashing this potential is constructing an efficient approach to load classical data into quantum states that can be executed by quantum computers or quantum simulators on classical hardware. Therefore, the split-step quantum walks (SSQW) algorithm was proposed to address this limitation. We facilitate SSQW to design parameterized quantum circuits (PQC) that can generate probability distributions and optimize the parameters to achieve the desired distribution using a variational solver. A practical example of implementing SSQW using Qiskit has been released as open-source software. Showing its potential as a promising method for generating desired probability amplitude distributions highlights the potential application of SSQW in option pricing through quantum simulation., Comment: 11 pages, 7 figures

We present a GPU solution for exact maximal clique enumeration (MCE) that performs a search tree traversal following the Bron-Kerbosch algorithm. Prior works on parallelizing MCE on GPUs perform a breadth-first traversal of the tree, which has limited scalability because of the explosion in the number of tree nodes at deep levels. We propose to parallelize MCE on GPUs by performing depth-first traversal of independent subtrees in parallel. Since MCE suffers from high load imbalance and memory capacity requirements, we propose a worker list for dynamic load balancing, as well as partial induced subgraphs and a compact representation of excluded vertex sets to regulate memory consumption. Our evaluation shows that our GPU implementation on a single GPU outperforms the state-of-the-art parallel CPU implementation by a geometric mean of 4.9x (up to 16.7x), and scales efficiently to multiple GPUs. Our code has been open-sourced to enable further research on accelerating MCE.

Since the start of full science operations from 2004, the Submillimeter Array has been implementing plans to expand IF bandwidths and upgrade receivers and cryostats. Metal mesh low-pass filters were designed to block infrared (IR) radiation to reduce the thermal load on the cryostats. Filters were fabricated on a quartz wafer through photolithography and coated with anti-reflection (AR) material. The filters were tested from 200 to 400 GHz to verify their passband performances. The measurement results were found to be in good agreement with EM simulation results. They were tested in the far-infrared (FIR) frequency range to verify out-of-band rejection. The IR reflectivity was found to be approximately 70%, which corresponded to the percentage of the area blocked by metal., Comment: 9pages, 14 figures

The performance of superconducting qubits is degraded by a poorly characterized set of energy sources breaking the Cooper pairs responsible for superconductivity, creating a condition often called ``quasiparticle poisoning". Both superconducting qubits and low threshold dark matter calorimeters have observed excess bursts of quasiparticles or phonons that decrease in rate with time. Here, we show that a silicon crystal glued to its holder exhibits a rate of low-energy phonon events that is more than two orders of magnitude larger than in a functionally identical crystal suspended from its holder in a low-stress state. The excess phonon event rate in the glued crystal decreases with time since cooldown, consistent with a source of phonon bursts which contributes to quasiparticle poisoning in quantum circuits and the low-energy events observed in cryogenic calorimeters. We argue that relaxation of thermally induced stress between the glue and crystal is the source of these events., Comment: 19 pages (main + supplementary), 6 figures. W. A. Page and R. K. Romani contributed equally to this work. Correspondence should be addressed to R. K. Romani

Objects play a crucial role in our everyday activities. Though multisensory object-centric learning has shown great potential lately, the modeling of objects in prior work is rather unrealistic. ObjectFolder 1.0 is a recent dataset that introduces 100 virtualized objects with visual, acoustic, and tactile sensory data. However, the dataset is small in scale and the multisensory data is of limited quality, hampering generalization to real-world scenarios. We present ObjectFolder 2.0, a large-scale, multisensory dataset of common household objects in the form of implicit neural representations that significantly enhances ObjectFolder 1.0 in three aspects. First, our dataset is 10 times larger in the amount of objects and orders of magnitude faster in rendering time. Second, we significantly improve the multisensory rendering quality for all three modalities. Third, we show that models learned from virtual objects in our dataset successfully transfer to their real-world counterparts in three challenging tasks: object scale estimation, contact localization, and shape reconstruction. ObjectFolder 2.0 offers a new path and testbed for multisensory learning in computer vision and robotics. The dataset is available at https://github.com/rhgao/ObjectFolder., Comment: In CVPR 2022. Gao, Si, and Chang contributed equally to this work. Project page: https://ai.stanford.edu/~rhgao/objectfolder2.0/

Polyphone disambiguation is the most crucial task in Mandarin grapheme-to-phoneme (g2p) conversion. Previous studies have approached this problem using pre-trained language models, restricted output, and extra information from Part-Of-Speech (POS) tagging. Inspired by these strategies, we propose a novel approach, called g2pW, which adapts learnable softmax-weights to condition the outputs of BERT with the polyphonic character of interest and its POS tagging. Rather than using the hard mask as in previous works, our experiments show that learning a soft-weighting function for the candidate phonemes benefits performance. In addition, our proposed g2pW does not require extra pre-trained POS tagging models while using POS tags as auxiliary features since we train the POS tagging model simultaneously with the unified encoder. Experimental results show that our g2pW outperforms existing methods on the public CPP dataset. All codes, model weights, and a user-friendly package are publicly available., Comment: Accepted in INTERSPEECH 2022

In this paper, we present a re-analysis of SuperCDMS data using a profile likelihood approach to search for sub-GeV dark matter particles (DM) through two inelastic scattering channels: bremsstrahlung radiation and the Migdal effect. By considering possible inelastic scattering channels, experimental sensitivity can be extended to DM masses that would otherwise be undetectable through the DM-nucleon elastic scattering channel, given the energy threshold of current experiments. We exclude DM masses down to $220~\textrm{MeV}/c^2$ at $2.7 \times 10^{-30}~\textrm{cm}^2$ via the bremsstrahlung channel. The Migdal channel search excludes DM masses down to $30~\textrm{MeV}/c^2$ at $5.0 \times 10^{-30}~\textrm{cm}^2$., Comment: This paper is being withdrawn due to an error in data selection during the analysis. Although incorrect, the limits are roughly representative of the sensitivity. The new corrected version of the result will be uploaded once ready

Training recognition models with synthetic images have achieved remarkable results in text recognition. However, recognizing text from real-world images still faces challenges due to the domain shift between synthetic and real-world text images. One of the strategies to eliminate the domain difference without manual annotation is unsupervised domain adaptation (UDA). Due to the characteristic of sequential labeling tasks, most popular UDA methods cannot be directly applied to text recognition. To tackle this problem, we proposed a UDA method with minimizing latent entropy on sequence-to-sequence attention-based models with classbalanced self-paced learning. Our experiments show that our proposed framework achieves better recognition results than the existing methods on most UDA text recognition benchmarks. All codes are publicly available.