Learning a task induces connectivity changes in neural circuits, thereby changing their dynamics. To elucidate task-related neural dynamics, we study trained recurrent neural networks. We develop a mean field theory for reservoir computing networks trained to have multiple fixed point attractors. Our main result is that the dynamics of the network’s output in the vicinity of attractors is governed by a low-order linear ordinary differential equation. The stability of the resulting equation can be assessed, predicting training success or failure. As a consequence, networks of rectified linear units and of sigmoidal nonlinearities are shown to have diametrically different properties when it comes to learning attractors. Furthermore, a characteristic time constant, which remains finite at the edge of chaos, offers an explanation of the network’s output robustness in the presence of variability of the internal neural dynamics. Finally, the proposed theory predicts state-dependent frequency selectivity in the network response.

We consider the problem of characterizing the set of input-output correlations that can be generated by an arbitrarily given quantum measurement. Our main result is to provide a closed-form, full characterization of such a set for any qubit measurement, and to discuss its geometrical interpretation. As applications, we further specify our results to the cases of real and complex symmetric, informationally complete measurements and mutually unbiased bases of a qubit, in the presence of isotropic noise. Our results provide the optimal device-independent tests of quantum measurements.

We compute next-to-leading order QCD corrections to the top-bottom interference contribution to $H+j$ production at the LHC. To achieve this, we combine the recent computation of the two-loop amplitudes for $gg\to Hg$ and $qg\to Hq$, performed in the approximation of a small $b$-quark mass, and the numerical calculation of the squared one-loop amplitudes for $gg\to Hgg$ and $qg\to Hqg$, performed within OpenLoops. We find that QCD corrections to the interference are large and similar to the QCD corrections to the top-mediated Higgs production cross section. We also observe a significant reduction in the mass-renormalization scheme uncertainty once the next-to-leading order QCD prediction for the interference is employed.

We experimentally demonstrate the heralded generation of bichromatic single photons from an atomic collective spin excitation (CSE). The photon arrival times display collective quantum beats, a novel interference effect resulting from the relative motion of atoms in the CSE. A combination of velocity-selective excitation with strong laser dressing and the addition of a magnetic field allows for exquisite control of this collective beat phenomenon. The present experiment uses a diamond scheme with near-IR photons that can be extended to include telecommunications wavelengths or modified to allow storage and retrieval in an inverted-$Y$ scheme.

We introduce a new coupled mode theory to model nonlinear Schrödinger equations for counterpropagating Bloch modes that include disorder-induced multiple scattering effects on nonlinear soliton propagation in photonic crystal waveguides. We also derive subunit-cell coupling coefficients and use these to introduce a generalized length scale associated with each coupling effect. In particular, we define a multiple-scattering length scale that quantifies the spatial extent of a disorder-induced cavity mode. Our numerical simulations of nonlinear pulse propagation are in excellent qualitative agreement with recent experiments and provide insight into how structural disorder inhibits soliton propagation and other nonlinear propagation effects in photonic crystal waveguides.

We demonstrate that the key to realizing arbitrary control over pairs of polarization states of light, i.e., transforming an arbitrarily polarized pair of input states to an arbitrarily polarized pair of output states, is the ability to generate pairs of states with orthogonal polarizations from nonorthogonal pairs of initial states. Then, we develop a new class of non-Hermitian metamaterials, termed complex birefringent metamaterials, which are able to do exactly this. Such materials could facilitate the detection of small polarization changes in scattering experiments as well as enable new polarization multiplexing schemes in communications networks.

Novel spatial, temporal, and energetically resolved measurements of bremsstrahlung hard-x-ray (HXR) emission from runaway electron (RE) populations in tokamaks reveal nonmonotonic RE distribution functions whose properties depend on the interplay of electric field acceleration with collisional and synchrotron damping. Measurements are consistent with theoretical predictions of momentum-space attractors that accumulate runaway electrons. RE distribution functions are measured to shift to a higher energy when the synchrotron force is reduced by decreasing the toroidal magnetic field strength. Increasing the collisional damping by increasing the electron density (at a fixed magnetic and electric field) reduces the energy of the nonmonotonic feature and reduces the HXR growth rate at all energies. Higher-energy HXR growth rates extrapolate to zero at the expected threshold electric field for RE sustainment, while low-energy REs are anomalously lost. The compilation of HXR emission from different sight lines into the plasma yields energy and pitch-angle-resolved RE distributions and demonstrates increasing pitch-angle and radial gradients with energy.

We report the experimental observation of spin-orbit torque induced switching of perpendicularly magnetized $\mathrm{Pt}/\mathrm{Co}$ elements in a time resolved stroboscopic experiment based on high resolution Kerr microscopy. Magnetization dynamics is induced by injecting subnanosecond current pulses into the bilayer while simultaneously applying static in-plane magnetic bias fields. Highly reproducible homogeneous switching on time scales of several tens of nanoseconds is observed. Our findings can be corroborated using micromagnetic modeling only when including a fieldlike torque term as well as the Dzyaloshinskii-Moriya interaction mediated by finite temperature.

The generation of short spin current pulses is the basis for fast spintronic devices. In thin bilayer systems consisting of a nonmagnetic metal and a ferromagnet, a pure spin current is induced by a precessing magnetization into the nonmagnetic layer by spin pumping. This effect has been experimentally demonstrated at ferromagnetic resonance at GHz frequencies. Here, it is theoretically shown that transient magnetization dynamics efficiently generates short spin current pulses that exhibit two transient contributions. An effective coherent spin current generation is found far above the ferromagnetic resonance up to THz frequencies although dynamic magnetization amplitudes are very small in this regime. The results provide a concept for coherent THz spin current generation.

Carbon nanotubes provide a rare access point into the plasmon physics of one-dimensional electronic systems. By assembling purified nanotubes into uniformly sized arrays, we show that they support coherent plasmon resonances, that these plasmons couple to nanotube and substrate phonons, and that the resulting phonon-plasmon resonances have quality factors as high as 10. Because nanotube plasmons intensely strengthen electromagnetic fields and light-matter interactions, they provide a compelling platform for surface-enhanced spectroscopy and tunable optical devices at deep-subwavelength scales.

Multiphoton entangled states such as “N00N states” have attracted a lot of attention because of their possible application in high-precision, quantum enhanced phase determination. So far, N00N states have been generated in spontaneous parametric down-conversion processes and by mixing quantum and classical light on a beam splitter. Here, in contrast, we demonstrate superresolving phase measurements based on two-photon N00N states generated by quantum dot single-photon sources making use of the Hong-Ou-Mandel effect on a beam splitter. By means of pulsed resonance fluorescence of a charged exciton state, we achieve, in postselection, a quantum enhanced improvement of the precision in phase uncertainty, higher than prescribed by the standard quantum limit. An analytical description of the measurement scheme is provided, reflecting requirements, capability, and restraints of single-photon emitters in optical quantum metrology. Our results point toward the realization of a real-world quantum sensor in the near future.

We introduce the first multiorbital effective tight-binding model to describe the effect of electron-electron interactions in this system. Upon fixing all the effective hopping parameters in the normal state against an ab initio band structure, and with only the overall scale of the interactions as the sole adjustable parameter, we find that a self-consistent Hartree-Fock solution reproduces extremely well the experimental behavior of the charge density wave (CDW) order parameter in the full range $0<T<T_c$, as well as the precise reciprocal space locations of the partial gap opening and Fermi arc development. The interaction strengths extracted from fitting to the experimental CDW gap are consistent with those derived from an independent Stoner-type analysis.

A search for the rare decays $B_s^0\to {\tau }^{+}{\tau }^{-}$ and $B^0\to {\tau }^{+}{\tau }^{-}$ is performed using proton–proton collision data collected with the LHCb detector. The data sample corresponds to an integrated luminosity of $3{\mathrm{fb}}^{-1}$ collected in 2011 and 2012. The $\tau$ leptons are reconstructed through the decay ${\tau }^{-}\to {\pi }^{-}{\pi }^{+}{\pi }^{-}{\nu }_{\tau }$. Assuming no contribution from $B^0\to {\tau }^{+}{\tau }^{-}$ decays, an upper limit is set on the branching fraction $\mathcal{B}(B_s^0\to {\tau }^{+}{\tau }^{-})<6.8\times {10}^{-3}$ at the 95% confidence level. If instead no contribution from $B_s^0\to {\tau }^{+}{\tau }^{-}$ decays is assumed, the limit is $\mathcal{B}(B^0\to {\tau }^{+}{\tau }^{-})<2.1\times {10}^{-3}$ at the 95% confidence level. These results correspond to the first direct limit on $\mathcal{B}(B_s^0\to {\tau }^{+}{\tau }^{-})$ and the world’s best limit on $\mathcal{B}(B^0\to {\tau }^{+}{\tau }^{-})$.

We report a first, complete lattice QCD calculation of the long-distance contribution to the $K^{+}\to {\pi }^{+}\nu \overline{\nu }$ decay within the standard model. This is a second-order weak process involving two four-Fermi operators that is highly sensitive to new physics and being studied by the NA62 experiment at CERN. While much of this decay comes from perturbative, short-distance physics, there is a long-distance part, perhaps as large as the planned experimental error, which involves nonperturbative phenomena. The calculation presented here, with unphysical quark masses, demonstrates that this contribution can be computed using lattice methods by overcoming three technical difficulties: (i) a short-distance divergence that results when the two weak operators approach each other, (ii) exponentially growing, unphysical terms that appear in Euclidean, second-order perturbation theory, and (iii) potentially large finite-volume effects. A follow-on calculation with physical quark masses and controlled systematic errors will be possible with the next generation of computers.

Smectic liquid crystals are characterized by layers that have a preferred uniform spacing and vanishing curvature in their ground state. Dislocations in smectics play an important role in phase nucleation, layer reorientation, and dynamics. Typically modeled as possessing one line singularity, the layer structure of a dislocation leads to a diverging compression strain as one approaches the defect center, suggesting a large, elastically determined melted core. However, it has been observed that for large charge dislocations, the defect breaks up into two disclinations [C. E. Williams, Philos. Mag. 32, 313 (1975)]. Here we investigate the topology of the composite core. Because the smectic cannot twist, transformations between different disclination geometries are highly constrained. We demonstrate the geometric route between them and show that despite enjoying precisely the topological rules of the three-dimensional nematic, the additional structure of line disclinations in three-dimensional smectics localizes transitions to higher-order point singularities.

We introduce an exact replica method for the study of critical systems with quenched bond randomness in two dimensions. For the $q$-state Potts model, we show that a line of renormalization group fixed points interpolates from weak to strong randomness as $q-2$ grows from small to large values. This theory exhibits a $q$-independent sector, and allows at the same time for a correlation length exponent which keeps the Ising value and continuously varying magnetization exponent and effective central charge. These findings appear to solve long-standing numerical and theoretical puzzles, and to illustrate the peculiarities which may characterize the conformal field theories of random fixed points.

A renormalization group flow of Hamiltonians for two-dimensional classical partition functions is constructed using tensor networks. Similar to tensor network renormalization [G. Evenbly and G. Vidal, Phys. Rev. Lett. 115, 180405 (2015); S. Yang, Z.-C. Gu, and X.-G. Wen, Phys. Rev. Lett. 118, 110504 (2017)], we obtain approximate fixed point tensor networks at criticality. Our formalism, however, preserves positivity of the tensors at every step and hence yields an interpretation in terms of Hamiltonian flows. We emphasize that the key difference between tensor network approaches and Kadanoff’s spin blocking method can be understood in terms of a change of the local basis at every decimation step, a property which is crucial to overcome the area law of mutual information. We derive algebraic relations for fixed point tensors, calculate critical exponents, and benchmark our method on the Ising model and the six-vertex model.

We report the observation of recoil inversion of the CO ($v=0$, $J_{\mathrm{CO}}=66$) state in the UV dissociation of lab-frame oriented carbonyl sulfide (OCS). This state is ejected in the opposite direction with respect to all other ($>30$) states and in absence of any OCS rotation, thus resulting in spatial filtering of this particular high-$J$ rovibrational state. This inversion is caused by resonances occurring in shallow local minima of the molecular potential, which bring the sulfur closer to the oxygen than the carbon atom, and is a striking example where such subtleties severely modify the photofragment trajectories. The resonant behavior is observed only in the photofragment trajectories and not in their population, showing that stereodynamic measurements from oriented molecules offer an indispensable probe for exploring energy landscapes.

We study locking phenomena of two strongly coupled, high quality factor nanomechanical resonator modes to a common parametric drive at a single drive frequency in different parametric driving regimes. By controlled dielectric gradient forces we tune the resonance frequencies of the flexural in-plane and out-of-plane oscillation of the high stress silicon nitride string through their mutual avoided crossing. For the case of the strong common parametric drive signal-idler generation via nondegenerate parametric two-mode oscillation is observed. Broadband frequency tuning of the very narrow linewidth signal and idler resonances is demonstrated. When the resonance frequencies of the signal and idler get closer to each other, partial injection locking, injection pulling, and complete injection locking to half of the drive frequency occurs depending on the pump strength. Furthermore, satellite resonances, symmetrically offset from the signal and idler by their beat note, are observed, which can be attributed to degenerate four-wave mixing in the highly nonlinear mechanical oscillations.