Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Binaries”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

The ‘Giraffe’: discovery of a stripped red giant in an interacting binary with an ∼2 M ⊙ lower giant

We report the discovery of a stripped giant + lower giant Algol-type binary, 2M04123153+6738486 (2M0412), identified during a search for non-interacting compact object – star binaries. 2M0412 is an evolved ( T eff, giant ≃ 4000 K), luminous ( L giant ≃ 150 L ⊙ ) red giant in a circular P = 81.2 d binary. 2M0412 is a known variable star previously classified as a semiregular variable. The cross-correlation functions of follow-up Keck/HIRES and LBT/Potsdam Echelle Polarimetric and Spectroscopic Instrument (PEPSI) spectra show an radial velocity variable second component with implied mass ratio q = M giant / M comp ≃ 0.20 ± 0.01. The All-Sky Automated Survey for SuperNovae (ASAS-SN), Asteroid Terrestrial-impact Last Alert System (ATLAS), Transiting Exoplanet Survey Telescope ( TESS ), and Zwicky Transient Facility (ZTF) light curves show that the giant is a Roche lobe filling ellipsoidal variable with an inclination of 49.4 ± 0.3°, and a giant mass of M giant = 0.38 ± 0.01 M ⊙ for a distance of ≃ 3.7 kpc. The mass of the giant indicates that its envelope has been partially stripped. The giant companion on the lower red giant branch has a mass of M comp = 1.91 ± 0.03 M ⊙ with T eff, comp ≃ 5000 K, L comp ≃ 60 L ⊙ , and R comp ≃ 11 R ⊙ . The lower giant contributes ${\sim }35{{\ \rm per\ cent}}$ of the flux in the V band. We also identify an orbital phase dependent, broad $\rm H\,\alpha$ emission line which could indicate ongoing accretion from the stripped red giant on to the companion.

(stars:) binaries: spectroscopic↗
DOE OSTI2021

Signatures of deconfined quark phases in binary neutron star mergers

We investigate the quark deconfinement phase transition in the context of binary neutron star (BNS) mergers. We employ a new finite-temperature composition-dependent equation of state (EOS) with a first order phase transition between hadrons and deconfined quarks to perform numerical relativity simulations of BNS mergers. Here, the softening of the EOS due to the phase transition causes the merger remnants to be more compact and to collapse to a black hole (BH) at earlier times. The phase transition is imprinted on the postmerger gravitational wave (GW) signal duration, amplitude, and peak frequency. However, this imprint is only detectable for binaries with sufficiently long-lived remnants. Moreover, the phase transition does not result in significant deviations from quasi-universal relations for the postmerger GW peak frequency. We also study the impact of the phase transition on dynamical ejecta, remnant accretion disk masses, r-process nucleosynthetic yields and associated electromagnetic (EM) counterparts. While there are differences in the EM counterparts and nucleosynthesis yields between the purely hadronic models and the models with phase transitions, these can be primarily ascribed to the difference in remnant collapse time between the two. An exception is the non-thermal afterglow caused by the interaction of the fastest component of the dynamical ejecta and the interstellar medium, which is systematically boosted in the binaries with phase transition as a consequence of the more violent merger they experience.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Reverse phase transitions in binary neutron-star systems with exotic-matter cores

Multimessenger observations of binary neutron-star mergers provide a unique opportunity to constrain the dense-matter equation of state. Although it is known from quantum chromodynamics that hadronic matter will undergo a phase transition to exotic forms of matter—e.g., quark matter—the onset density of such a phase transition cannot be computed from first principles. Hence, it remains an open question if such phase transitions occur inside isolated neutron stars or during binary neutron-star mergers, or if they appear at even higher densities that are not realized in the cosmos. Here, in this article, we perform numerical relativity simulations of neutron-star mergers and investigate scenarios in which the onset density of such a phase transition is exceeded in at least one inspiraling binary component. Our simulations reveal that shortly before the merger, it is possible that such stars undergo a “reverse phase transition”—i.e., densities decrease and the quark core inside the star disappears, leaving a purely hadronic star at merger. After the merger, when densities increase once more, the phase transition occurs again and leads, in the cases considered in this work, to the rapid formation of a black hole. We compute the gravitational-wave signal and the mass ejection for our simulations of such scenarios and find clear signatures that are related to the postmerger phase transition—e.g., smaller ejecta masses due to the softening of the equation of state through the quark core formation. Unfortunately, we do not find measurable imprints of the reverse phase transition.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Detectability of QCD phase transitions in binary neutron star mergers: Bayesian inference with the next generation gravitational wave detectors

We study the detectability of postmerger QCD phase transitions in neutron star binaries with next-generation gravitational-wave detectors Cosmic Explorer and Einstein Telescope. We perform numerical relativity simulations of neutron star mergers with equations of state that include a quark deconfinement phase transition through either a Gibbs or Maxwell construction. These are followed by Bayesian parameter estimation of the associated gravitational-wave signals using the nrpmw waveform model, with priors inferred from the analysis of the inspiral signal. We assess the ability of the model to measure the postmerger peak frequency $f$$^{peak}_{2}$ and identify aspects that should be improved in the model. We show that, even at postmerger signal to noise ratios as low as 10, the model can distinguish (at the 90% level) $f$$^{peak}_{2}$ between binaries with and without a phase transition in most cases. Phase-transition induced deviations in the $f$$^{peak}_{2}$ from the predictions of equation-of-state insensitive relations can also be detected if they exceed 1.6⁢σ. Our results suggest that next-generation gravitational wave detectors can measure phase transition effects in binary neutron star mergers. Furthermore, unless the phase transition is “strong,” disentangling it from other hadronic physics uncertainties will require significant theory improvements.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

General relativistic hydrodynamic simulations of binary strange star mergers

We perform fully general-relativistic simulations of binary strange star mergers considering two different approaches for thermal effects. The first uses a cold equation of state (EOS) derived from a modified version of the MIT bag model which is then supplemented by a Γ-law correction. The second approach employs a microphysical description of the finite-temperature effects. We describe results obtained with the two treatments, highlighting the influence of thermal effects. We find that the postmerger dynamics differs significantly in the two cases, leading to quantitative differences in the postmerger gravitational-wave spectrum and ejecta mass. The peak frequency of the postmerger gravitational-wave emission is consistent with the established quasi-universal relations for binary neutron star mergers and as a result, our simulations cannot distinguish between mergers of neutron stars and those of strange stars. Our models with realistic treatment of finite-temperature effects produce a significant amount of ejecta ≳0.02 M ⊙ ​. Here, the resulting flux of strangelets near the Earth, computed assuming that all neutron star mergers are in fact strange-stars mergers and that the binary considered here is representative, is in tension with experimental upper limits. As such, our results tentatively disfavor a scenario in which strange-quark matter is the lowest energy state of matter.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Spectroscopic Orbits of 11 Nearby, Mid-to-late M-dwarf Binaries

We present the spectroscopic orbits of 11 nearby, mid-to-late M dwarf binary systems in a variety of configurations: 2 single-lined binaries (SB1s), 7 double-lined binaries (SB2s), 1 double-lined triple (ST2), and 1 triple-lined triple (ST3). Eight of these orbits are the first published for these systems, while five are newly identified multiples. We obtained multi-epoch, high-resolution spectra with the TRES instrument on the 1.5 m Tillinghast Reflector at the Fred Lawrence Whipple Observatory located on Mt. Hopkins in AZ. Using the TiO molecular bands at 7065–7165 Å, we calculated radial velocities for these systems, from which we derived their orbits. We find LHS 1817 to have in a 7 hr period a companion that is likely a white dwarf, due to the ellipsoidal modulation we see in our MEarth-North light-curve data. We find G 123-45 and LTT 11586 to host companions with minimum masses of 41 M {sub Jup} and 44 M {sub Jup} with orbital periods of 35 and 15 days, respectively. We find 2MA 0930+0227 to have a rapidly rotating stellar companion in a 917 day orbital period. GJ 268, GJ 1029, LP 734-34, GJ 1182, G 258-17, and LTT 7077are SB2s with stellar companions with orbital periods of 10, 96, 34, 154, 5, and 84 days; LP 655-43 is an ST3 with one companion in an 18 day orbital period and an outer component in a longer undetermined period. In addition, we present radial velocities for both components of L 870-44AB and for the outer components of LTT 11586 and LP 655-43.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

BVR{sub C}I{sub C} Observations and Analysis of the UV LEO-type Binary, the Pre-WUMa, V1023 Persei

V1023 Per is a polar spotted, but well-detached dwarf binary, very likely a Pre-WUMa eclipsing binary. It was observed on 11 nights in 2015 October, November, and December at Dark Sky Observatory in North Carolina with the 0.81 m reflector of Appalachian State University. The period behavior is complex and may be increasing with a cubic or quadratic ephemeris. Its odd light curves are of high amplitude but have no totality and reach maximum light just before and following the secondary eclipse, indicating that it has polar spots similar to UV Leo. The 28 yr orbital period study calls the attention of observers to further monitor this binary to understand the complex nature of the period evolution. Its presently fixed polar spot does indicate that it must have a strong magnetic field and that it is synchronously rotating. The BVR{sub c}I{sub c} simultaneous 2016 Wilson–Devinney program (W-D) solution gives fillouts of 68% and 75% for the primary and secondary components, respectively. The polar position of the spot, its radius (24°), and T-fact (0.75) indicate that a strong magnetic field is present. The small ΔT in the components (∼289 K) show that the stars are similar in spectral type despite them being well detached. The inclination is high, ∼85.°19 ± 0.°04, but there is not a total eclipse due to the high mass ratio (∼0.739 ± 0.001). Due to the low galactic position (longitude = 150.°390, latitude = −1.°023), its reddening is addressed. The primary components temperature is estimated to be 5250 ± 250 K.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

HD 143811 AB b: A Directly Imaged Planet Orbiting a Spectroscopic Binary in Sco-Cen

We present confirmation of HD 143811 AB b, a substellar companion to spectroscopic binary HD 143811 AB through direct imaging with the Gemini Planet Imager (GPI) and Keck NIRC2. HD 143811 AB was observed as a part of the GPI Exoplanet Survey in 2016 and 2019 and is a member of the Sco-Cen star formation region. The exoplanet is detected ∼430 mas from the host star by GPI. With two GPI epochs and one from Keck/NIRC2 in 2022, we confirm through common proper motion analysis that the object is bound to its host star. We derive an orbit with a semimajor axis of $64^{+32}_{-14}$ au and eccentricity $0.23^{+0.24}_{-0.16}$. Spectral analysis of the GPI H-band spectrum and NIRC2 L′ photometry provides additional proof that this object is a substellar companion. We compare the spectrum of HD 143811 AB b to PHOENIX stellar models and Exo-Radioactive-Convective Equilibrium Model (REM) exoplanet atmosphere models and find that Exo-REM models provide the best fits to the data. From the Exo-REM models, we derive an effective temperature of $1042^{+178}_{-132}$ K for the planet and translate the derived luminosity of the planet to a mass of 5.6 ± 1.1 M Jup assuming hot-start evolutionary models. HD 143811 AB b is the first directly imaged planet around a binary that is not on an ultrawide orbit. Future characterization of this object will shed light on the formation of planets around binary star systems.

Astronomy and AstroPhysics↗
DOE OSTI2025

Phase equilibria and microstructure evolution in the Ni-rich region of binary Ni-Ce and ternary Ni-Ce-X (X=Al, Nb, Cr, Ti) systems at 900 °C

The equilibrium compositions of all phases in the Ni-rich region of binary Ni-Ce and ternary Ni-Ce-X (X=Al, Nb, Cr, Ti) systems at 900 °C were determined experimentally through isothermal annealing for up to 500 h and compared with currently available computational thermodynamic descriptions. Additionally, the study examined how adding a third element to the binary system affects the microstructure by analyzing changes in the volume fraction of the intermetallic phase in the eutectic region. It was observed that Nb and Ti addition lowered the intermetallic fraction in the eutectic region while Al and Cr addition did the opposite. The fraction of intermetallic phase in the eutectic region is very important because it contributes to extreme brittleness of these alloys. Also, Al, Nb and Ti addition promoted different precipitates upon annealing when added in sufficient amounts. The collected data on composition and phase equilibria provides crucial information for improving the thermodynamic assessments of the binary and ternary alloy systems.

energy dispersive spectrometry↗
DOE OSTI2022

Binary Superlattices of Gold Nanoparticles in Two Dimensions

Here, we have created two-dimensional (2D) binary superlattices by cocrystallizing gold nanoparticles (AuNPs) of two distinct sizes into √3 x √3 and 2 x 2 complex binary superlattices, derived from the hexagonal structures of the single components. The building blocks of these binary systems are AuNPs that are functionalized with different chain lengths of poly(ethylene glycol) (PEG). The assembly of these functionalized NPs at the air–water interface is driven by the presence of salt, causing PEG-AuNPs to migrate to the aqueous surface and assemble into a crystalline lattice. We have used liquid surface X-ray reflectivity (XR) and grazing incidence small-angle X-ray scattering (GISAXS) to examine the assembly and crystallization at the liquid interface.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Phase Diagram and Structure Map of Binary Nanoparticle Superlattices from a Lennard-Jones Model

A first-principles prediction of the binary nanoparticle phase diagram assembled by solvent evaporation has eluded theoretical approaches. In this paper, we show that a binary system interacting through the Lennard-Jones (LJ) potential contains all experimental phases in which nanoparticles are effectively described as quasi hard spheres. We report a phase diagram consisting of 53 equilibrium phases, whose stability is quite insensitive to the microscopic details of the potentials, thus giving rise to some type of universality. Furthermore, we show that binary lattices may be understood as consisting of certain particle clusters, i.e., motifs, that provide a generalization of the four conventional Frank–Kasper polyhedral units. Our results show that metastable phases share the very same motifs as equilibrium phases. Finally, we discuss the connection with packing models, phase diagrams with repulsive potentials, and the prediction of likely experimental superlattices.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Spatially Aligned Binary Single-Site Catalyst on Defective SiO 2 for Cascading Reactions

Capitalizing on the success of single-atom catalysts (SACs), dual-atom catalysts (DACs) have emerged as a new frontier in heterogeneous catalysis. However, most SACs and DACs studies seek to uniformly distribute the catalytic sites on the support material, which can hinder their effectiveness in intricate multistep cascading reactions. Particularly, it is a grand challenge to precisely control the spatial distribution of two different single sites forming binary sites so that reactants and intermediates contact the catalytic sites in the exact sequence required by the reaction steps. Here, in this work, we report a new type of binary single-site catalyst, Cu 1 –Zr 1 @SiO 2 , with Cu 1 and Zr 1 sites spatially aligned with the reaction sequence of the cascade reactions. The catalyst is synthesized by a modified reverse microemulsion approach, with single Cu sites anchored by nonbridging oxygen hole centers, which were induced by doping single Zr sites into SiO 2 . Low-energy ion scattering spectroscopy (LEIS) reveals that the outermost surface of the catalyst contains only Cu single sites, while the Zr sites are dispersed in the bulk. The catalytic performance is demonstrated in ethanol conversion to butenes, a model cascade reaction which includes ethanol dehydrogenation and aldol condensation steps. The precisely spatially controlled binary sites enable ethanol to first undergo dehydrogenation to acetaldehyde on Cu sites, followed by aldol condensation of acetaldehyde on Zr sites. As a result, C 3+ olefins selectivity as high as 77.0% (56.0% selectivity of butenes) is achieved by suppressing ethylene formation.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Discovery of two-dimensional binary nanoparticle superlattices using global Monte Carlo optimization

Binary nanoparticle (NP) superlattices exhibit distinct collective plasmonic, magnetic, optical, and electronic properties. Here, we computationally demonstrate how fluid-fluid interfaces could be used to self-assemble binary systems of NPs into 2D superlattices when the NP species exhibit different miscibility with the fluids forming the interface. We develop a basin-hopping Monte Carlo (BHMC) algorithm tailored for interface-trapped structures to rapidly determine the ground-state configuration of NPs, allowing us to explore the repertoire of binary NP architectures formed at the interface. By varying the NP size ratio, interparticle interaction strength, and difference in NP miscibility with the two fluids, we demonstrate the assembly of an array of exquisite 2D periodic architectures, including AB-, AB 2 -, and AB 3 -type monolayer superlattices as well as AB-, AB 2 -, A 3 B 5 -, and A 4 B 6 -type bilayer superlattices. Our results suggest that the interfacial assembly approach could be a versatile platform for fabricating 2D colloidal superlattices with tunable structure and properties.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Stratified wind from a super-Eddington X-ray binary is slower than expected

Accretion disks in strong gravity ubiquitously produce winds, seen as blueshifted absorption lines in the X-ray band of both stellar mass X-ray binaries (black holes and neutron stars) and supermassive black holes. Some of the most powerful winds (termed Eddington winds) are expected to arise from systems in which radiation pressure is sufficient to unbind material from the inner disk (L ≳ L Edd ). These winds should be extremely fast and carry a large amount of kinetic power, which, when associated with supermassive black holes, would make them a prime contender for the feedback mechanism linking the growth of those black holes with their host galaxies. Here we show the XRISM Resolve spectrum of the galactic neutron star X-ray binary, GX 13+1, which reveals one of the densest winds ever seen in absorption lines. This Compton-thick wind significantly attenuates the flux, making it appear faint, although it is intrinsically more luminous than usual (L ≳ L Edd ). However, the wind is extremely slow, more consistent with the predictions of thermal-radiative winds launched by X-ray irradiation of the outer disk than with the expected Eddington wind driven by radiation pressure from the inner disk. This puts new constraints on the origin of winds from bright accretion flows in binaries, but also highlights the very different origin required for the ultrafast (v ~ 0.3c) winds seen in recent Resolve observations of a supermassive black hole at a similarly high Eddington ratio.

Audard, Marc [Univ. of Geneva, Versoix (Switzerlan↗
DOE OSTI2023

Quantitative relationships between film morphology, charge carrier dynamics, and photovoltaic performance in bulk-heterojunction binary vs. ternary acceptor blends

Addressing pertinent and perplexing questions regarding why nonfullerene acceptors (NFAs) promote higher power conversion efficiencies (PCEs) than traditional fullerenes and how photoactive bulk heterojunction (BHJ) film morphology, charge photogeneration, and recombination dynamics dictate solar cell performance have stimulated many studies of polymer solar cells (PSCs), yet quantitative relationships remain limited. Better understanding in these areas offers the potential to advance materials design and device engineering, afford higher PCEs, and ultimate commercialization. Here we probe quantitative relationships between BHJ film morphology, charge carrier dynamics, and photovoltaic performance in model binary and ternary blend systems having a wide bandgap donor polymer, a fullerene, and a promising NFA. We show that optimal PC 71 BM incorporation in a PBDB-TF:ITIC-Th binary system matrix retains the original π-face-on orientation, ITIC-Th crystallinity and BHJ film crystallite dimensions, and reduces film upper surface ITIC-Th segregation. Such morphology changes together simultaneously increase hole (μ h ) and electron (μ e ) mobilities, facilitate light-activated ITIC-Th to PC 71 BM domain electron delocalization, reduce free charge carrier (FC) bimolecular recombination (BR) within PBDB-TF:ITIC-Th mixed regions, and increase FC extraction pathways viaPBDB-TF:PC 71 BM pairs. The interplay of these effects yields significantly enhanced inverted cell short-circuit current density (J SC ), fill factor (FF), and PCE. Unexpectedly, we also find that excessive PC 71 BM in the PBDB-TF:ITIC-Th binary system alters the PBDB-TF orientation to π-edge-on, increases large scale PC 71 BM-rich aggregations and BHJ upper surface PC 71 BM composition. Further, these morphology changes increase parasitic decay processes such as intersystem crossing from photoexcited PC 71 BM, compromising the J SC , FF, and PCE metrics. ITIC-Th X-ray diffraction reveals a unique sidechain-dominated molecular network with previously unknown sidechain-end group stacking, rationalizing the STEM and GIWAXS results, photophysics, and the high μ e . DFT computation reveals charge transfer networks within ITIC-Th crystallites, supporting excited-state electron delocalization from ITIC-Th to PC 71 BM. This structure–property understanding leads to a newly reported NFA blend with PCE near 17%.

14 SOLAR ENERGY↗
DOE OSTI2024

Cepheids with giant companions: II. Spectroscopic confirmation of nine new double-lined binary systems composed of two Cepheids

Context.Binary Cepheids with giant companions are crucial for studying the physical properties of Cepheid variables, in particular providing the best means to measure their masses. Systems composed of two Cepheids are even more important, but to date, only one such system has been identified, in the Large Magellanic Cloud (LMC). Aims.Our current aim is to increase the number of these systems known tenfold and to provide their basic characteristics. The final goal is to obtain the physical properties of the component Cepheids, including their masses and radii, and to learn about their evolution in the multiple systems, also revealing their origin. Methods.We started a spectroscopic monitoring campaign of nine unresolved pairs of Cepheids from the OGLE catalog to check if they are gravitationally bound. Two of these so-called double Cepheids are located in the LMC, five are in the Small Magellanic Cloud (SMC), and two are in the Milky Way (MW). Results.We report a spectroscopic detection of the binarity of all nine of these double Cepheids with orbital periods ranging from 2 to 18 years. This increases the number of known binary double (BIND) Cepheids from 1 to 10 and triples the number of all confirmed double-lined binary (SB2) Cepheids. For five BIND Cepheids, the disentangled pulsational light curves of the components show anti-correlated phase shifts due to orbital motion. We show the first empirical evidence that typical period–luminosity relations (PLRs) are rather binary Cepheid PLRs, as they include light of the companion. Conclusions.The statistics of pulsation period ratios of BIND Cepheids do not agree with those expected for pairs of Cepheids of the same age. These ratios together with the determined mass ratios far from unity suggest a merger origin of at least one component for about half of the systems. The SMC and MW objects are the first found in SB2 systems composed of giants in their host galaxies. The Milky Way BIND Cepheids are also the closest such systems, being located at about 11 and 26 kpc.

Astronomy & Astrophysics↗
DOE OSTI2021

Source term method for binary neutron stars initial data

The initial condition problem for a binary neutron star system requires a Poisson equation solver for the velocity potential with a Neumann-like boundary condition on the surface of the star. Difficulties that arise in this boundary value problem are: (a) the boundary is not known a priori, but constitutes part of the solution of the problem; (b) various terms become singular at the boundary. In this work, we present a new method to solve the fluid Poisson equation for irrotational/spinning binary neutron stars. The advantage of the new method is that it does not require complex fluid surface fitted coordinates and it can be implemented in a Cartesian grid, which is a standard choice in numerical relativity calculations. This is accomplished by employing the source term method proposed by Towers, where the boundary condition is treated as a jump condition and is incorporated as additional source terms in the Poisson equation, which is then solved iteratively. The issue of singular terms caused by vanishing density on the surface is resolved with an additional separation that shifts the computation boundary to the interior of the star. We present two-dimensional tests to show the convergence of the source term method, and we further apply this solver to a realistic three-dimensional binary neutron star problem. By comparing our solution with the one coming from the initial data solver cocal, we demonstrate agreement to approximately 1%. We report our method can be used in other problems with non-smooth solutions like in magnetized neutron stars.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Compressing deep neural networks on FPGAs to binary and ternary precision with hls4ml

We present the implementation of binary and ternary neural networks in the hls4ml library, designed to automatically convert deep neural network models to digital circuits with FPGA firmware. Starting from benchmark models trained with floating point precision, we investigate different strategies to reduce the network's resource consumption by reducing the numerical precision of the network parameters to binary or ternary. We discuss the trade-off between model accuracy and resource consumption. In addition, we show how to balance between latency and accuracy by retaining full precision on a selected subset of network components. As an example, we consider two multiclass classification tasks: handwritten digit recognition with the MNIST data set and jet identification with simulated proton-proton collisions at the CERN Large Hadron Collider. The binary and ternary implementation has similar performance to the higher precision implementation while using drastically fewer FPGA resources.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗