5632 research outputs found
Sort by
A vision transformer for decoding surgeon activity from surgical videos
No full textThe intraoperative activity of a surgeon has substantial impact on postoperative outcomes. However, for most surgical procedures, the details of intraoperative surgical actions, which can vary widely, are not well understood. Here we report a machine learning system leveraging a vision transformer and supervised contrastive learning for the decoding of elements of intraoperative surgical activity from videos commonly collected during robotic surgeries. The system accurately identified surgical steps, actions performed by the surgeon, the quality of these actions and the relative contribution of individual video frames to the decoding of the actions. Through extensive testing on data from three different hospitals located in two different continents, we show that the system generalizes across videos, surgeons, hospitals and surgical procedures, and that it can provide information on surgical gestures and skills from unannotated videos. Decoding intraoperative activity via accurate machine learning systems could be used to provide surgeons with feedback on their operating skills, and may allow for the identification of optimal surgical behaviour and for the study of relationships between intraoperative factors and postoperative outcomes
Quantum Nondemolition Measurements with Optical Parametric Amplifiers for Ultrafast Universal Quantum Information Processing
No full textRealization of a room-temperature ultrafast photon-number-resolving quantum nondemolition (QND) measurement would have significant implications for photonic quantum information processing, enabling, for example, deterministic quantum computation in discrete-variable architectures, but the requirement for strong coupling has hampered the development of scalable implementations. In this work, we propose and analyze a nonlinear-optical route to photon-number-resolving QND using quadratic (i.e., χ⁽²⁾) nonlinear interactions. We show that the coherent pump field driving a frequency-detuned optical parametric amplifier (OPA) experiences displacements conditioned on the number of signal Bogoliubov excitations. A measurement of the pump displacement thus provides a QND measurement of the signal Bogoliubov excitations, projecting the signal mode to a squeezed photon-number state. We then show how our nonlinear OPA dynamics can be utilized to deterministically generate Gottesman-Kitaev-Preskill states with only additional Gaussian resources, offering an all-optical route for fault-tolerant quantum information processing in continuous-variable systems. Finally, we place these QND schemes into a more traditional context by highlighting analogies between the frequency-detuned optical parametric oscillator and multilevel atom-cavity quantum electrodynamics systems by showing how continuous monitoring of the outcoupled pump quadrature induces conditional localization of the intracavity signal mode onto squeezed photon-number states. Our analysis suggests that our proposal may be viable in near-term χ⁽²⁾ nonlinear nanophotonics, highlighting the rich potential of the OPA as a universal tool for ultrafast non-Gaussian quantum state engineering and quantum computation
In Search of the Edge: A Bayesian Exploration of the Detectability of Red Edges in Exoplanet Reflection Spectra
No full textReflection spectroscopy holds great promise for characterizing the atmospheres and surfaces of potentially habitable terrestrial exoplanets. The surface of the modern Earth exhibits a sharp albedo change near 750 nm caused by vegetation—the red edge—which would leave a strong spectral signature if present on an exoplanet. However, the retrieval of wavelength-dependent surface properties from reflection spectra has seen relatively little study. Here, we propose a new surface albedo parameterization capable of retrieving the wavelength location of a priori unknown “edge-like” features. We demonstrate that a wavelength-dependent surface albedo model achieves higher accuracy in retrieving atmospheric composition than a uniform albedo model. Wavelength-dependent surfaces are also generally preferred over a uniform albedo model when retrieving simulated reflection spectra for a modern Earth analog, even for moderate signal-to-noise ratios (S/N = 10) and Earth-like clouds. Further, the location of the modern Earth’s red edge can be robustly and precisely constrained (within 70 nm for S/N = 10). Our results suggest that future space-based direct-imaging missions have the potential to infer surface compositions for rocky exoplanets, including spectral edges similar to those caused by life on the modern Earth
Flares, jets and quasi-periodic outbursts from neutron star merger remnants
No full textUsing numerical relativity simulations with a subgrid dynamo prescription to generate strong initial magnetic fields, we investigate the possibility of launching a jet-like outflow from the hypermassive neutron star (HMNS) during the early stages of the merger, prior to the remnants collapse to a black hole. We demonstrate that buoyant instabilities in the strongly magnetized HMNS can lead to a periodic emission of powerful electromagnetic flares shortly after the merger. These are followed by a collimated mildly relativistic outflow. Both types of outflows feature quasi-periodic kilohertz substructure. These early-time outflows may power precursors to short-duration gamma-ray bursts (SGRB) or in some cases the entire SGRB. While the overall temporal power spectrum we find broadly agrees with the one recently reported for quasi-periodic oscillations in the SGRB GRB910711, our simulations suggest that the periodic electromagnetic substructure is dominated by magnetohydrodynamic shearing processes rather than correlating with the corresponding post-merger gravitational wave signal
Three-dimensional Dynamics of Strongly Twisted Magnetar Magnetospheres: Kinking Flux Tubes and Global Eruptions
No full textThe origins of the various outbursts of hard X-rays from magnetars (highly magnetized neutron stars) are still unknown. We identify instabilities in relativistic magnetospheres that can explain a range of X-ray flare luminosities. Crustal surface motions can twist the magnetar magnetosphere by shifting the frozen-in footpoints of magnetic field lines in current-carrying flux bundles. Axisymmetric (2D) magnetospheres exhibit strong eruptive dynamics, i.e., catastrophic lateral instabilities triggered by a critical footpoint displacement of ψ_(crit) ≳ π. In contrast, our new three-dimensional (3D) twist models with finite surface extension capture important non-axisymmetric dynamics of twisted force-free flux bundles in dipolar magnetospheres. Besides the well-established global eruption resulting (as in 2D) from lateral instabilities, such 3D structures can develop helical, kink-like dynamics, and dissipate energy locally (confined eruptions). Up to 25% of the induced twist energy is dissipated and available to power X-ray flares in powerful global eruptions, with most of our models showing an energy release in the range of the most common X-ray outbursts, ≲10⁴³ erg. Such events occur when significant energy builds up while deeply buried in the dipole magnetosphere. Less energetic outbursts likely precede powerful flares, due to intermittent instabilities and confined eruptions of a continuously twisting flux tube. Upon reaching a critical state, global eruptions produce the necessary Poynting-flux-dominated outflows required by models prescribing the fast radio burst production in the magnetar wind—for example, via relativistic magnetic reconnection or shocks
On-chip flow rate sensing via membrane deformation and bistability probed by microwave resonators
No full textPrecise monitoring of fluid flow rates constitutes an integral problem in various lab-on-a-chip applications. While off-chip flow sensors are commonly used, new sensing mechanisms are being investigated to address the needs of increasingly complex lab-on-a-chip platforms which require local and non-intrusive flow rate sensing. In this regard, the deformability of microfluidic components has recently attracted attention as an on-chip sensing mechanism. To develop an on-chip flow rate sensor, here we utilized the mechanical deformations of a 220 nm thick Silicon Nitride membrane integrated with the microfluidic channel. Applied pressure and fluid flow induce different modes of deformations on the membrane, which are electronically probed by an integrated microwave resonator. The flow changes the capacitance, and in turn resonance frequency, of the microwave resonator. By tracking the resonance frequency, liquid flow was probed with the device. In addition to responding to applied pressure by deflection, the membrane also exhibits periodic pulsation motion under fluid flow at a constant rate. The two separate mechanisms, deflection and pulsation, constitute sensing mechanisms for pressure and flow rate. Using the same device architecture, we also detected pressure-induced deformations by a gas to draw further insight into the sensing mechanism of the membrane. Flow rate measurements based on the deformation and instability of thin membranes demonstrate the transduction potential of microwave resonators for fluid–structure interactions at micro- and nanoscales
²H-²H clumping in molecular hydrogen method and preliminary results
No full textWe present a method for measuring the clumped isotope composition of molecular hydrogen (H₂) using a high-resolution mass spectrometer, the Thermo 253 Ultra, improved to address subtle artifacts arising from instrument baselines and non-linear responses. We also present methods for purification and concentration of H₂ from natural and experimental samples, tailored to this measurement. We document the accuracy of the method through comparison to established methods for the determination of δD values, and through measurements of H₂ gases of widely varying D content that were driven to isotopic equilibrium with respect to their distributions of isotopologues by heating in the presence of a catalyst. Experimental reproducibility of δD and ΔDD values over months averages ±0.5 and ± 6.9 ‰, respectively (1σ) — both small fractions of common natural variations. We explore methods of gas purification and handling, and show that preferred methods result in low (0–4 ‰) changes in δD and undetectable changes in ΔDD. Our methods and data processing procedures were further tested by comparing measurements of mixtures of H₂ gases that varied widely in δD and ∆DD with a model describing proportions of isotopologues in such mixtures. Application of these methods to H₂ that is residual to laboratory consumption by cultured methanogens shows that metabolic ‘back reaction’ (metabolic production of H₂ from water-derived protons during net H2 consumption) is responsible for driving the ΔDD value of residual H₂ toward equilibrium at environmental temperatures. Finally, we report the first measurements of the clumped isotope composition of molecular hydrogen in natural geological samples collected from high and low temperature submarine hydrothermal vents (Lost city, Rainbow, Ashadze) and an intracontinental natural reservoir in Mali; initial findings suggest that ∆DD of H₂ generally records temperatures of fluid venting or long-term storage, even in cases where the δD of H₂ has not equilibrated with water at those temperatures. This study establishes the first clumped isotope systematics of molecular hydrogen based on both experimental and natural samples, including key processes in the biogeochemical cycle of H₂
A new high-pressure high-temperature deformation apparatus to study the brittle to ductile transition in rocks
No full textUnderstanding the micro-mechanisms underlying the localized–ductile transition (LDT) as well as the brittle–plastic transition (BPT) has become crucial for our wider understanding of crustal processes and seismicity. Given how difficult in situ observations of these transitions are to perform, laboratory experiments might be our only way to investigate the processes active under these conditions (high T and high P). Here, we present Triaxial AppaRatus for GEoThermal energy, a new gas-based triaxial apparatus located at EPFL in Switzerland that was specifically designed to operate under conditions where both the LDT and BPT can occur in geomaterials. We show that the machine is capable of deforming rock samples at confining pressures of up to 400 MPa, temperatures of up to 800 °C, and pore pressures (liquid or gas) of up to 300 MPa while keeping the temperature gradient along samples of 40 mm in length and 20 mm in diameter minimal (less than 30 at 700 °C). Most importantly, the maximum load is 1000 kN (stresses as high as 2.2 GPa on 24 mm samples and 3 GPa on 20 mm samples), allowing for the deformation of very competent rock samples. Moreover, during deformation, the pair of syringe pore pressure pumps allow for continuous permeability or dilatancy recording. We benchmarked our machine against existing data in the literature and show that it accurately and precisely records stress, strain, permeability, pressure, and temperature
H/V Analysis in Juchitán de Zaragoza, Oaxaca, Following the 2017 M 8.2 Tehuantepec, México, Earthquake
No full textIn September 2017, over 450 lives were lost in Mexico as a result of two unusual, large-magnitude, normal earthquakes. On 7 September, an M 8.2 earthquake occurred offshore of the State of Oaxaca in the Gulf of Tehuantepec, one of the largest extensional earthquakes to have occurred in a subduction zone. Twelve days later on 19 September an M 7.1 damaging earthquake struck near Puebla and Morelos, over 600 km away. Both earthquakes occurred in the downgoing Cocos plate, which is subducting beneath the North American plate. The first large event was followed on 23 September by a shallow M 6.1 extensional earthquake near Juchitán de Zaragoza, Oaxaca. Researchers from Mexico and the United States collaborated to deploy a temporary seismic network to study the aftershocks of the M 8.2 Tehuantepec, Mexico, earthquake, which included a three-week deployment of 51 Magseis Fairfield Z-Land 5-Hz three-component nodal seismometers (“nodes”) near Juchitán and a 6-month deployment of 10 Nanometrics Trillium 120PA broadband seismometers with Reftek RT130 dataloggers for 6 months. In this article, we analyze the capabilities of the nodes to calculate the horizontal/vertical (H/V) spectral ratio and relative amplification using both microtremors and earthquakes and validate the results calculated with the nodes using data from broadband stations from this and previous deployments in the area. We create maps showing a correlation of the distribution of the fundamental frequency and relative amplification of the soil and compare them with the geology and the damage caused by the September 2017 earthquakes. There is a lack of public awareness and discrepancies in the construction procedures in the region, and we find that the majority of damaged houses in the area of study followed the location of river beds and tended to be in places with low resonance frequencies despite being in a low amplification zone
Low temperature synthesis of new highly graphitized N-doped carbon for Pt fuel cell supports, satisfying DOE 2025 durability standards for both catalyst and support
No full textFor polymer electrolyte membrane fuel cells (PEMFCs), the state-of-the-art electrocatalysts are based on carbon-supported Pt group metals. However, current carbon supports suffer from carbon corrosion during repeated start-stop operations, causing performance degradation. We report a new strategy to produce highly graphitized carbon with controllable N-doping that uses low-temperature synthesis (650 ℃) from g-C₃N₄ carbon-nitrogen precursor with pyrolysis using Mg. The high graphiticity is confirmed by high-intensity 2D Raman peak with low I_D/I_G (0.57), pronounced graphitic XRD planes, and excellent conductivity. Without further post-treatment, this highly graphitized N-doped carbon (HGNC) material combines high pyrrolic-N content with high porosity. Supporting Pt on HGNC exhibits excellent oxygen reduction activity for PEMFC with greatly improved durability as proved by real-time loss measurements of Pt and carbon, the first to surpass the DOE 2025 durability targets for both catalyst and support. The Pt/HGNC-65 shows 32% and 24% drop in mass activity after accelerated durability tests of both electrocatalyst and support, respectively, which are less than DOE target of 40% loss. The atomistic basis for this durability is explained via quantum mechanics-based molecular dynamics simulations. Interestingly, it is found that pyrrolic-N strongly interacts with Pt, making the Pt catalyst more stable during fuel cell reaction