Preparing matrix product states (MPSs) on quantum computers is an essential routine in the simulation of many-body physics. However, widely used schemes based on staircase circuits are often too deep to execute on current hardware. Here, we demonstrate that MPSs with short-range correlations can be prepared with shallow circuits by leveraging heuristics from approximate quantum compiling (AQC). We achieve this with ADAPT-AQC, an adaptive-ansatz preparation algorithm, and introduce a generalized initialization procedure for the existing AQC-Tensor algorithm. We first compare these methods for the task of preparing a molecular electronic structure ground state. We then use them to prepare an antiferromagnetic (AFM) ground state of the 50-site Heisenberg XXZ spin chain near the AFM-XY phase boundary. Through the execution of circuits with up to 59 cz depth and 1251 cz gates, we perform a global quench and observe the relaxation of magnetic ordering in a parameter regime previously inaccessible due to deep ground-state preparation circuits. Our results demonstrate how the integration of quantum and classical resources can push the boundary of what can be studied on quantum computers.
The Superconducting Quantum Materials and Systems Center, a U. S. Department of Energy National Quantum Information Science Research Center, has conducted a comprehensive and coordinated study using superconducting transmon qubit chips with known performance metrics to identify the underlying materials-level sources of device-to-device performance variation. Following qubit coherence measurements, these qubits of varying base superconducting metals and substrates have been examined with various non-destructive and invasive material characterization techniques at Northwestern University, Ames National Laboratory, and Fermilab as part of a blind study. We find trends in variations of the depth of the etched substrate trench, the thickness of the surface oxide, and the geometry of the sidewall, which when combined, lead to correlations with the T1 lifetime across different qubits on the same chip. In addition, we provide a list of features that varied from device to device, for which the impact on performance requires further studies. Finally, we identify two low-temperature characterization techniques that may potentially serve as proxy tools for qubit measurements. These insights provide materials-oriented solutions to not only reduce performance variations across neighboring devices but also to engineer and fabricate devices with optimal geometries to achieve performance metrics beyond the state-of-the-art values.
Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.
We present MIMIC-III-Ext-PPG, a large-scale, quality-assessed photoplethysmography (PPG) dataset derived from the matched waveform subset of MIMIC-III. Our dataset provides 30-second PPG segments with annotations tailored for various cardiovascular and respiratory analyses. In particular, with 6.3 million segments from 6,189 subjects, it represents the largest publicly available resource for heart rhythm classification, with heart rhythm annotations derived from bedside charted observations. For subsets where arterial blood pressure (ABP), respiratory (RESP), and/or electrocardiography (ECG) signals are available, we also provide systolic/diastolic blood pressure, respiratory rate, and heart rate annotations, extracted using best practice from the underlying signals. We provide signal quality assessments for all signals. This ensures a high-quality, publicly available dataset of unprecedented size that can be used as a benchmarking resource for machine learning approaches for a broad range of prediction tasks, which remains easily extendable by leveraging additional clinical metadata from the MIMIC-III clinical database.
Optical atomic clocks based on laser-cooled trapped ions and atoms have advanced rapidly over the past decade. With fractional frequency uncertainties now surpassing 10 −18 , they are some of the most precise measurement tools ever built. Yet researchers are still pursuing new avenues of research to explore the fundamental limits to their stability, accuracy, and reproducibility. In this mini-review, we provide a survey of the current state of the art by describing the fundamental principles and techniques that underpin this progress, the architectures used to realize optical clocks, and the supporting laser technologies that are essential to their operation. We also examine the progress that has been made toward a redefinition of the second in the International System of Units and the inclusion of optical clocks into the global time and frequency metrology infrastructure. Finally, we discuss emerging applications of optical clocks and look at the prospects for making their precision more readily accessible to end users.
