nopperl/arxiv-image-text · Datasets at Hugging Face

dedup-isc-ft-v107-score float64 0.32 1	uid stringlengths 32 32	text stringlengths 0 32.5k	paper_id stringlengths 1 14	original_image_filename stringlengths 5 222
0.506329	ef28484d1b124e4cb5a91945a636686a	Simulation for α = 0.7.	2012.09317	path_alpha07.jpg
0.456111	1fb27033b5da40b8813830cc1b6bab34	Simulation for α = 0.8.	2012.09317	path_alpha08.jpg
0.558488	1c21a41e86944ab1b5ca97de2376eb45	Simulation for α = 0.9.	2012.09317	path_alpha09.jpg
0.443299	a734c812238944658cfbf31c39e28f89	Simulation for α = 1.	2012.09317	path_alpha10.jpg
0.436989	3d04fe5e80e44497a6b2a1703727b82d	Summary of two types of process which occur in the brain and seem to map onto Kahneman's System I and System II.	2012.09318	two_types_of_Process.jpg
0.487699	4116ba1c186341849dda5d6b37dcd615	A Comparison of the absolute bias of the R-estimator with its standard error	2012.09445	Figure0.jpg
0.467319	7eb7ecf4f608433086a919dfed32e889	Estimate of the Tail Parameter in 1,000 Years of Simulated Monthly Data	2012.09445	Figure2.jpg
0.412265	40d42e46873b44e38ae31ec35e4463d3	The Lorenz Curve for a Single Simulation	2012.09445	Figure3.jpg
0.434534	9bad90e7230640f1b9d97fe52571cec6	300 Years of Simulated Weekly Data in an Economy with One Million People	2012.09445	Figure5.jpg
0.529833	581961dcc5844a439d58df7df7880baa	The Evolution of Average Beliefs When Agents are Least-Squares Learners with Infinite Lives	2012.09445	Picture1.jpg
0.474894	1ea87597fece47dc8a3b8128ec240a44	The Evolution of Average Beliefs When Agents are Constant-Gain Learners with Finite Lives	2012.09445	Picture3.jpg
0.455372	8a1dc28f81ff49a6b01f1dbe92145802	ALMA Band-6 images before tapering of 12 targets (image size: 5”×5”). A black circle shows the beam size of each image. Black contours correspond to 4σ, 8σ, 12σ, and 16σ. A plus mark represents the centroid determined in the Ks-band image. Some of the ALMA detected sources, such as 406444 and 434585, show a larger spatial offset between the dust continuum emission and the Ks-band centroid than the other detected sources. This is probably due to the lower signal-to-noise ratios of their dust continuum emission. The stacked image of the five individually ALMA non-detected sources with log(M_*/M_⊙)= 10.0–10.4 is also shown. The central position is shown with a plus mark. The stacked emission is detected at 5σ.	2012.09447	figure1.jpg
0.487482	9329a830cb564b71900bcdb7c73532dd	Gas mass fraction versus gas-phase metallicity for our galaxy sample at z∼3.3. The horizontal bar in the bottom right corner shows an additional ± 1σ error on the gas mass fraction coming from the systematic uncertainty on M_gas. The dashed lines show how the two quantities depend on each other when fixing the dust-to-stellar mass ratio at 1×10^-3 and 3×10^-3. We find no statistically significant correlation between the gas mass fraction and gas-phase metallicity among our sample.	2012.09447	figure5.jpg
0.542845	a6c016e978e14b5a8c3592b377cd6668	Relation between gas mass fraction and gas-phase metallicity for star-forming galaxies from z=0 to z∼3.3. The horizontal bar in the left bottom corner represents an additional ± 1σ error on the gas mass fraction coming from the systematic uncertainty on M_gas for our sample. Only the molecular gas components are considered for the galaxies in <cit.> and <cit.>. The star-forming galaxies at z≳2 show an offset toward the lower gas-phase metallicity from the distribution of the local galaxies. The black lines show the model tracks from the gas regulator model of <cit.> assuming different mass-loading factors between λ= 0.5 and 2.5. The distribution of the star-forming galaxies at z∼3.3 on this diagram can be broadly explained with the model tracks with λ∼ 2–2.5, suggesting the redshift evolution of the mass-loading factor for star-forming galaxies.	2012.09447	figure6.jpg
0.488328	8ce8df4de8b24939a2c2687b91746299	Best-fit SEDs (sold line) obtained with magphys. Data points are from the COSMOS2015 catalog <cit.> and the ALMA Band-6 observation in this study. Arrows show 3σ upper limits. When fitting the ALMA non-detected sources, the 1.3 mm flux and its uncertainty are set to 1.5σ± 1σ <cit.>.	2012.09447	figure7.jpg
0.409336	e5f0782c95ff442486ce529844cacb46	Frequency histogram of weighted average of relative error of estimated ATE for eight different models: ordinary least square(OLS), LASSO, RIDGE, random forest (RF), GRU, CNN, and multilayer perception (MLP) based on the semi-synthetic data with N=160000, M=500, and α=0.05 and β=0.05 in (<ref>) of the main paper.	2012.09448	ATE_JD_160000_500_8_model.jpg
0.434653	efcf44aa6e5c4297b680f196adbff6bf	Weighted average of relative err. of estimated ATE using IoC and IwC for different models; N=160000, M=500, and α=0.05 and β=0.05 in (<ref>).	2012.09448	ATE_JD_160000_500_new.jpg
0.446368	f31af8a39de34973922c0d12c7cdffbd	Frequency histogram of weighted average of relative error of estimated ATTE for eight different models: ordinary least square(OLS), LASSO, RIDGE, random forest (RF), GRU, CNN, and multilayer perception (MLP) based on the semi-synthetic data with N=160000, M=500, and α=0.05 and β=0.05 in (<ref>) of the main paper.	2012.09448	ATTE_JD_160000_500_8_model.jpg
0.464055	a80322035b304afa88407869b05ec908	Weighted average of relative err. of estimated ATTE using IoC and IwC for different models; N=160000, M=500, and α=0.05 and β=0.05 in (<ref>).	2012.09448	ATTE_JD_160000_500_new.jpg
0.465121	2a44c23537b44e12a855b8c87becfae0	The mean relative error computed by (<ref>) vs. number of observations.	2012.09448	simu_cond_exp_consistency.jpg
0.402974	5b7bdf1b96d548718ed8f9ea60418717	The average standard deviation computed by (<ref>) vs. number of observations.	2012.09448	simu_cond_exp_std_consistency.jpg
0.428618	8b5d93c8b0d042daaa039b9c6f6d80b5	The mean relative error computed by (<ref>) vs. number of observations.	2012.09448	simu_exp_convergence.jpg
0.377465	0c97092d9dd34d8488c6aceaf2697dba	The average standard deviation computed by (<ref>) vs. number of observations.	2012.09448	simu_exp_std_convergence.jpg
0.422849	26ad506c557948ea9181629cd8b048b9	A photo of the MAFDS and its experimental test setup.	2012.09449	DSC_2177_opt_ausschnitt_cut_grau.jpg
0.512908	30786296a11c4746922a636b779d4a91	Plot compares density estimator of piezo-beam, cf. Subsection <ref>. A kernel density estimator based on all available experimental data (gray solid line), a kernel density estimator based on the computer model (black dashed line) and kernel density estimator based on an improved surrogate model (black solid line). The values of the outcome of the 10 experiments are indicated on the x-axis.	2012.09449	KeKo2019_PIEZO_plot.jpg
0.472585	6f923541b9f2485b8996dff46035aff3	Box plot without outliers of 10^6 estimates for the 0.95-quantiles of the n=100 absolute model errors for both computer models of the MAFDS.	2012.09449	KeOh2001_MAFDS_plot.jpg
0.498739	9df68f32baac46969c3bcc27fb7626e5	Three estimates of the cumulative density function of the experimental outcome of the MAFDS. An estimator based on experimental data (black solid line), an estimator based on computer experiments with computer model 1 (black dashed line) and computer model 2 (gray dashed line).	2012.09449	RoOb2011_MADFS_plot.jpg
0.476981	e0d5531ccdc745cab3cf608526334c82	Box plot without outliers for the 0.95-quantiles of the n=100 absolute model errors for both computer models of the MAFDS. The box plot contains B = 500 bootstrap 0.95-quantiles.	2012.09449	WoStLe2017_MAFDS_plot.jpg
0.484634	3de45e8e2794425c8848830ef350c1f4	A CAD model of the lateral vibration attenuation system with piezo-elastic supports and a sectional view of one of the piezo-elastic supports, cf. <cit.>.	2012.09449	piezo_beam.jpg
0.421755	6674ae92ed4745bbb0ba484a8d6db170	LS-F.	2012.09451	1872_F-eps-converted-to.jpg
0.416188	59f46f7791a84eeeb49c73e6a51f6dcc	LS-G.	2012.09451	1872_G-eps-converted-to.jpg
0.469548	d1fec0fc822b4cc9869ff97bca2f431f	The results with different values of k.	2012.09451	big_k_trim-eps-converted-to.jpg
0.444253	752ecb6225a04c2b8256ea27a2c5cf42	The number of blocks before and after applying LS-G and LS-F on METIS.	2012.09451	block_trim-eps-converted-to.jpg
0.459049	dda1c23b70a84dda95b129d4185ca50a	Predictions for the mass of the core and the amount of metallic elements in the envelope obtained using the 3rd order theory of figure. The published results are compared with the ones obtained by varying the d_0 parameter are shown in the figure and for a core constituted of pure water and of silicate and water.	2012.09454	fig6-c-arxiv.jpg
0.423496	8b847ac162e445a0bf4be4da6e45314a	Optical plasma diagnostics show strong self emission from dense narrow regions of the plasma. We see (a) raw interferogram measurement of the plasma, 1 ns after the interaction starts, measured with the Nomarsky interferometer (b) the retrieved electronic density of the plasma and (c) the raw streak camera measurement of the self emission of the same laser shot (including measurement lineout in white). x=0,y=0 are the coordinates of the jet shock point; the laser travels left to right and focalizes in the same point.	2012.09455	Image_worm.jpg
0.444622	8b0d461201064cf5bfa9eadf0ec8cb8a	RCF imprints converted to dose and alpha particle number density for three consecutive shots varying the gas jet density profile. The number density conversion presumes the appearance of alpha particles, a cut-off energy equal to the last signal layer and a step-wise flat spectral shape between layers. The typical FWHM divergence angle of the beam imprint is 9, all images have the same spatial scale. Note that number densities are normalized to one pixel, with respect to the resolution of RCF scans of 600 dpi. For shots #62 and #63, the presumed pre-aligned laser axis corresponds to the centre of the illustrated frames. For shot #64 this does not apply.	2012.09455	RCF_62_63_64_data.jpg
0.498145	b9060f6aeeaf415896fc87f94f3d2afc	Projectile spectrum for three consecutive shots varying the gas jet density profile, presuming an alpha particle beam impacting on the stack of RCF. RCF color channels R, G and B are independently analyzed.	2012.09455	RCF_62_63_64_spectrum.jpg
0.407792	0687f70542944dca8c54315cacff4f89	Projectile spectrum for three consecutive shots varying the gas jet density profile and transverse position, presuming an alpha particle beam impacting on the stack of RCF. RCF color channels R, G and B are independently analyzed.	2012.09455	RCF_79_85_87_spectrum.jpg
0.445226	a0c847948322455fa1ff670d95909772	Electronic temperature chart in MeV of the n_at,max = 4× 10^19cm^-3 (0.036 n_c) simulation.	2012.09455	Te_TUCAN.jpg
0.458839	f8093faaaebf4b37b436f9a1f8cc5337	Diameter evolution of an individual etch pit on a TasTrak CR-39 with strioscopic snapshots at 2, 3, 9 and 15 of etching with 6.25 N etchant at 70. The diameter evolves considerably and is not detected for 2 of etching as it may be underneath the resolution of the microscope system.	2012.09455	centerion_pit_strio_sketch.jpg
0.396265	687afba797c34bd9b54809575e447ff1	Summary of results.	2012.09455	comp.jpg
0.429131	24ed3b27bb40441fa37c7c8f1a1be7e0	Comparison between the detector RAW signal comprising photopeak and all accelerated particle species (blue full line) and a signal with photopeak only (orange dashed line).	2012.09455	comp_photopeak_norm.jpg
0.382355	5463e70af6cf41d0b9206be144dbc96e	Photodiode detector impulse response (left) and signal after deconvolution (right).	2012.09455	deconv_subplot.jpg
0.426681	228c3dbf390649698a2ffb08618323d9	E_y field chart resulting from the simulation with a density profile with n_at,max = 4× 10^20cm^-3. The blue dashed line marks the density peak.	2012.09455	ey_TUCAN.jpg
0.389623	fbf45ad50ee140e4962492ad934ed9b2	(a) Interferometric images overlaid by self-emission line-outs for shots with variation of the transverse position of the laser-gas interaction. The longest plasma channel is recorded for shot #81 with positive transverse displacement – the length of the channel decreases for the series of shots towards negative transverse displacements. The last shot of the series #83 does not agree with this assessment and shows a longer channel again. Note that (b) un-driven gas density profiles acquired prior to corresponding shots show visual changes from shot-to-shot, indicating changes to the nozzle. This agrees with a qualitative difference of the nozzle surface between shot #77 and #83, presented with fig. <ref>.	2012.09455	gas_77_80_81_82_83.jpg
0.44898	96bdf814c360425e8b0569270869b82f	Shot-to-shot variation of the gas jet density profile using a mix of Helium and Nitrogen in a ratio 1 to 9, (top row) longitudinal un-driven gas jet density profile, and (bottom row) interferometric image superposed to aligned laser beam and plasma self-emission in arbitrary units as red line, 1ns after the interaction starts. The laser is focused to x=0,y=0, the coordinates of the jet shock point, but transverse displaced by 0;±50.	2012.09455	gas_79_85_87_overview.jpg
0.507553	138f4487b1674706b8dd157debac2d39	Diameter evolution for similar groups of etch pits on a TasTrak CR-39 with snapshots at 2, and 9 of etching with 6.25 N etchant at 70. Highlighted are groups (a,b,d,e) with circles concentric to the etch pit position for the short etching time and (c) with circles concentric to the pit position for long etching time. The dashed line delimits a scratched area, we do not see the edge of the plastic sheet.	2012.09455	monstermark_pit_brightfield_sketch.jpg
0.41038	b0ac722be9ce44bbac75db5749106150	Overview on RCF results for shots #79, #85 and #87 converted from RAW to deposited dose maps.	2012.09455	qulitative_RCF_stack_Al4UEBT32EBT3.jpg
0.414494	53ee9e3a33954c2ba9449350b06cd95b	Nitrogen ions density chart (left) and zoom on one of the ion acoustic waves travelling in the down-ramp of the density profile (right) for the first simulation run with a_0 = 6.8.	2012.09455	qx_TUCAN_and_zoom.jpg
0.496641	2e9c4e4dbbf449f6adc6c707b2a5c8a1	Z-scan of etched pits on a TasTrak CR-39 after 15 of etching with 6.25 N NaOH etchant at 70. The bulk etching rate was quantified to 1.32. Experimental data points are underlaid with a relative point number density color map in grayscale. The simulated spectrally resolving species specific L(D) lobes are for alpha particles on the left hand side and for fully ionized Nitrogen ions on the right hand side. Specific impact energies are highlighted with red marker arrows (1–4). Experimental data points that overlap with the lobes in the range of their uncertainty are denoted with the closest corresponding energy on the lobe. Simulated with the CR-39 plug-in of PySTarT.	2012.09455	track_diameter_energy_15.jpg
0.393433	f384dd6127914218b1a195583a6d9dad	An analytical investigation of the critical transmissibilities of both strains for increasing clustering coefficients for the ER networks defined in Fig <ref>. The curves are generated from Eqs <ref> and <ref> at the value of T in which the outbreak fraction becomes large than ϵ. The critical points reduce as C increases, indicating that contact clustering helps to spread an emergent epidemic. The interval [T_1^,T_2^] is the transmissibility window in which strain-1 exists as an epidemic on the network without strain-2. We observe that increased clustering reduces the interval and thus increases the extent of coinfection in this model, at the expense of the mono-infection equilibrium. We note, however, that clustering can never reduce the coinfection critical point below the single strain threshold due to the strict conditions of perfect coinfection in the premise of the model.	2012.09457	Delta_transmissibilities.jpg
0.406831	cf355fd736d649b9b393e7f911db2b9a	(Top) The outbreak fractions of both strains for increasing triangle probabilities within the CHCN degree distribution defined in Eq <ref>. Clustering reduces the epidemic threshold, in agreement with ER experiments; however, coinfection decreases with increasing θ. (Bottom) The expected epidemic sizes of both strains for a p^CHCN(k) network with θ=0.0 and a degree-δ CHCN distribution network defined by Eq <ref>. Clustering increases the epidemic threshold in this model. Markers are the average of 100 repetitions of bond percolation on CHCN networks with N=1e^5 nodes and T_2=0.6.	2012.09457	PerfCoinfFig5PlotB.jpg
0.428417	9e21e7b4c97a421984bbdb01424cc315	The tree-like edge topologies found in the GCC following the first strain and their probabilities.	2012.09457	TREES_coinf.jpg
0.449054	32d42d02a2424f7eb789ca98b5ea1df0	The 3-cycle edge-topologies found in the GCC following the first strain and their probabilities. Node colours and patters are defined according to Fig <ref>. Triangles D, E and F consist of inhomogeneous neighbour-states and hence, due to the symmetry of the shape, their reflection about a vertical axis bisecting the focal node can also occur with equal probability. The curved arrows in triangles D and F indicate the additional pathway through the cycle that the infected neighbour could infect the focal node.	2012.09457	TRIANGLES_coinf.jpg
0.541495	ae746598d94a4951a87ab0f3d011efa2	The outbreak fractions for the 2-strain coinfection model on ER random graphs with T_2=0.6, N = 25000, for varying clustering coefficients and fixed mean degree ⟨ s⟩+2⟨ t ⟩=2. It is clear that clustering reduces the epidemic thresholds of both diseases and reduces the outbreak size of the first strain; however, it increases the coinfected fraction of the network. Points are the average of 500 repeats of Monte Carlo simulation while solid lines are the theoretical predictions from Eqs <ref> and <ref>.	2012.09457	coinfection.jpg
0.443824	eb01c3e963fd431a9eae58660704fde1	The expected number of bars of length ≥, N^_B, as a function of .	2012.09459	BMCalculated.jpg
0.445585	e5e90015606e4f589e11585d2272602b	N^_B as approximated with simulations of random Rademacher walks.	2012.09459	NvepsBMnothing.jpg
0.405588	c40d5ac8926d49b8b25d83705a18a0cc	A function f: [0,1] → and its associated tree T_f in dashed lines.	2012.09459	TreeFunction.jpg
0.433145	65a32d573bf44ee38c0d1d3885863f75	A depiction of the first steps of the algorithm which assigns a barcode (f) to a tree T_f.	2012.09459	algorithm.jpg
0.433826	1a2dcfb0306f4ad3a89510733afe1eda	Comparison between theoretical predictions and finite element simulation for twist (a) and stretch (b) as a function of the spontaneous deformation parameter λ_s. For these calcuations, we used a fibre with aspect ratio R_0/L=1/15 represented by 5400 hex-8 elements.	2012.09461	TwistStretchComparison.jpg
0.451198	4e80121ea0574c56bf57822187014d96	Finite element simulation of a fibre with Young modulus E=1, Poisson ratio ν=0.45, 4800 elements, aspect ratio of R_0/L=0.07 and a R independent director-field with α=π/4 and β=π/2. The fibre is fixed at one end and subject to a tension T=0.2 πμ R_0^2 while preventing rotation at the other end. As the value of λ_s is decreased, the fibre wants to shrink and twist but cannot do the latter as the ends are not allowed to rotate. The fibre thus coils to release some of the twist energy, resulting in a greater stroke amplitude then that of classical linear actuation.	2012.09461	coil.jpg
0.40009	2a9c956b6e814e7bb1dfecf4b3f8bfa1	Comparison between Finite element simulations and theoretical data for the case α=π/4 and β=π/4. a) An example of a deformed cross section directly from FE analysis. In red are theoretical lines for the deformed radii while dots are node positions. b) Comparison between predicted and theoretical coning. c) Comparison between theory and FE on the predicted rotation angle difference as a function of radius. A fibre with aspect ratio R_0/L=1/5, and 54000 hex-8 elements was used for this simulation.	2012.09461	coning2.jpg
0.407857	4d702e0350634fe4b7804395911f467c	a) 2D plot of twist as a function of both angles α and β. The twist is maximised when β=π/2 and α∼π/4. b) – c) Variation of the twist and stretch as a function of α for different values of λ_s.	2012.09461	costant2.jpg
0.401871	543a1573c7c14894a3347b74cded426a	Relationship between output twist a) and stretch b) as a function of the twist imposed at genesis. Shaded in grey, the region for known experimental results <cit.>.	2012.09461	genesis.jpg
0.444126	8bb867071cb64e048c5453869f4940bb	a) Schematic of a fibre being drawn and twisted from a drop of LC monomer while being cured with UV light. b) An example of the resulting director field in a fibre of radius r=2/τ_0. The two integral curves highlight the change in azimuthal component as a function of the radius.	2012.09461	genesis2.jpg
0.419085	1e886a40c1fe412da0ec9f645961dc06	(a) The optical dipole potential energy for right-handed circular polarization. Solid line represents the case where ℓ=+1, dashed curve represents the case where ℓ=-1; (b) The optical dipole potential energy for left-handed circular polarization. Solid line represents the case where ℓ=-1, dashed curve represents the case where ℓ=+1. In both (a) and (b) the dotted line shows the optical dipole potential energy without taking into account the contribution of the spin-orbit term. The insets to the figures show the potential energies of the solid curves and their projections in the focal plane. These insets should be interchanged for the dashed curves. In all plots the potential energy is in recoil energy units while the radial distances are in LG beam waist w_0 units.	2012.09465	Fig1.jpg
0.537845	6d19c5b4b60143ae834f65a3ace93ccb	(a) The optical dipole potential energy for right-handed circular polarization. Solid line represents the case where ℓ=+5, dashed curve is for ℓ=-5. (b) The optical dipole potential energy for left-handed circular polarization. Solid line represents the case where ℓ=-5, dashed curve is for ℓ=+5. In both (a) and (b) the dotted line shows the optical dipole potential energy without taking into account the contribution of the spin-orbit term. In both plots the potential energy is in recoil energy units while the radial distances are in LG beam waist w_0 units.	2012.09465	Fig2.jpg
0.498272	1eb48bec68b0485dba9fe6ea2581c29d	(a) The optical dipole potential energy for a bi-chromatic field with ℓ=1 (solid line). The plot is inverted if ℓ=-1 (dashed line). (b) Optical dipole potential energy for a bi-chromatic field with ℓ=5 (solid line). This plot is inverted if ℓ=-5. In both plots the potentials are given in recoil energy units while the radial distances are in LG beam waist w_0 units.	2012.09465	Fig3.jpg
0.442721	f26e0411027d4b1482a0e43a5e96c2d5	(a) The optical dipole potential energy, in the radial and axial directions, for a bi-chromatic field with ℓ=5 (left). (b) The plot is inverted for ℓ=-5 (right). The potential energy is given in recoil energy units, the radial distances are in LG beam waist w_0 units, while the axial distances are in Rayleigh range (z_R) units. The arrows in the plot (b) indicate the trapping regions.	2012.09465	Fig4.jpg
0.495479	748b4a0b327e4a6d859e8d55f1d7437a		2012.09467	1.jpg
0.434933	62f10df832834a91af5196337b1c9927	Comparison of the analytical (exact) and annealed result of Eq. (<ref>) using the 6th order Gauss-Legendre collocation method with a timestep Δ t = 0.5. The numerical result is visually indistinguishable from the analytical (exact) solution.	2012.09469	annealing_result.jpg
0.429897	af6bb72085b249cda2a0835ca8c71e76	Comparison of the simulated annealed, analytical, and Newton method for u(0) = 0.1. The analytical solution was obtained by using the 4th Runge-Kutta method with much smaller time step sizes. The simulated annealed solution does not match the result of the Newton method.	2012.09469	nonlinear_annealing_0.1.jpg
0.454794	2b41d3e6f4c24d68a208be3a3f23a61a	Comparison of the scaling of the energy gap between the ground and first excited state for the adiabatic Hamiltonian for the Crank-Nicolson scheme with changing Δ x. It can be seen that g_min is proportional to Δ x and higher time steps result in a lower energy gap.	2012.09469	scaling_crank_nicolson_dx.jpg
0.511214	dbdd7da906f44890b6a5e1761ab1563c	Comparison of the scaling of the energy gap between the ground and first excited state for the adiabatic Hamiltonian for the explicit Euler. For sufficient small Δ x, the expected scaling and numerical determined scaling match.	2012.09469	scaling_explicit_euler_dx.jpg
0.404099	c02ea58f53e44749ae494148d0d66c53	Error of the annealed Gauss-Legendre 6th order Runge-Kutta step using our variational approach, compared to a numerical evaluation of the same Runge-Kutta step. The dashed lines are obtained on a D-Wave quantum annealer, and the solid lines show an exact annealing reference solution calculated on a classical computer. The parameter c is the exponent shift of the variational approach in Eq. (<ref>). For all cases, two qubits per number register are used.	2012.09469	variational_convergence.jpg
0.45143	76c07e9a50834f3ca7e011d726f4eada	(Color online) Pressure as a function of baryon density ρ/ρ_0 for hadronic phase (solid line), quark phase (dash line) and mixed phase (dash-dotted line). The fitted EoSs with first-order (a) and crossover (b) phase transitions use respectively B= 150 and 90 MeV/fm^3 in the MIT model.	2012.09471	fig1.jpg
0.468463	b6db6d87f0c94daaafaabb523187fff0	(Color online) Maximum compression density reached with hadronic EoS (thick lines) and EoSs including first-order phase transition and crossover (thin lines) in Au + Au collisions at beam energies of 2-8 GeV/nucleon.	2012.09471	fig2.jpg
0.431901	72e92f1bdbb440ea9f93bc6dd74261b2	(Color online) Proton average transverse momentum ⟨ p_x⟩ as a function of rapidity in Au + Au collisions using EoSs with and without phase transition at beam energies of 2-8 GeV/nucleon. The experimental data are taken from Ref. <cit.>.	2012.09471	fig3.jpg
0.441927	14f6c0ee578c464291e8b78fdcdaa249	(Color online) Proton directed flow v_1 in Au + Au collisions using three kinds of EoSs at beam energies of 2-8 GeV/nucleon. The experimental data are taken from Ref. <cit.>.	2012.09471	fig4.jpg
0.434658	799c2f4c81354682a93d1ae739600160	(Color online) Same as Fig. <ref>, but with different phase-transition boundaries.	2012.09471	fig5.jpg
0.46452	5b559b27ce2649809c18090df3efc9b2	The formation of the total number of bonds 𝒩_ b with time t for different bond lengths ℓ_ b. The data is shown for n=3 tri-functional (dashed lines) and n=4 tetra-functional (solid lines) monomers. Parts (a) and (b) show the data for the FENE and harmonic bonds, respectively.	2012.09473	Figure1.jpg
0.532836	0e82cbb822524b78bfe51683e8130628	Number density of bonds ρ_ b (panel a) and monomers ρ_ m (panel b) as a function of bond length ℓ_ b. The data is shown for the systems after the isobaric equilibration. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.	2012.09473	Figure2.jpg
0.43542	3b7b20708601466e850311e20932f694	Panels (a-d) show the simulation snapshots of a 2σ thick layer along the z direction after the network cure in the canonical ensemble. The lateral dimensions of the snapshots are 67.03 σ. Snapshots are shown for different bond lengths ℓ_ b and functionalities n, as mentioned in the figure headings. Panels (a) and (c) are for ℓ_ b = 0.90σ and panels (b) and (d) are for ℓ_ b = 0.97σ. Panel (e) shows the fraction of the void volume v_ void as a function of ℓ_ b. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.	2012.09473	Figure3.jpg
0.476856	9316b3c0cf8c4b0a9c1368152ada4c77	The normalized thermal transport coefficient κ/κ_ linear as a function of the bond length ℓ_ b. The data is shown for different network functionality n and for the both bonds. The data is normalized with respect to the linear chain connected by the harmonic springs giving κ_ linear = 5.4±0.3 k_ B/τ m. Here the linear chain lengths are chosen as N_ℓ=50. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.	2012.09473	Figure4.jpg
0.434515	270df98126f041498b71653f06d03963	The normalized thermal transport coefficient κ/κ_ linear as a function of the volumetric specific heat c_ v (part a) and the sound wave velocity v (panel b). The data is shown for different functionality n, bond length ℓ_ b and for both bonds. The data is normalized with respect to the linear chain connected by the harmonic springs giving κ_ linear = 5.4±0.3 k_ B/τ m. Here the linear chain lengths are chosen as N_ℓ=50. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.	2012.09473	Figure5.jpg
0.536148	891b015e5dd54ce88f031664dc0d560a	Vowel-wise acoustic feature comparisons between COVID-19 negative (neg.) and COVID-19 positive (pos.) participants in form of boxplots for features with a differentiation effect r>.4 ordered from left to right according to a decreasing r, respectively. The effect size r as well as the p-value of the Mann-Whitney U difference test are given above each boxplot. p is rounded to three decimal places. r is rounded to two decimal places. Outliers (marked with red plus symbols) are defined as value that are more than 1.5 times the interquartile range away from the bottom or top of the respective box. # = number of, F0 = fundamental frequency, F1 = first vowel formant, F3 = third vowel formant, len. = length, pctlrg = percentile range, RS = rising slope, seg. = segment, ST = semitone from 27.5 Hz, SDnorm = standard deviation normalised by the arithmetic mean (coefficient of variation), slp = slope; VR = voiced regions	2012.09478	boxplots.jpg
0.434778	541175090317427499416a9170371b58	Dynamical formation of topological synchronized states (TSSs). (a) Schematic of the considered nonlinear system. Nonlinear oscillators are placed on a square lattice. Each oscillator is linearly coupled to other oscillators at the nearest neighbor sites. The linear coupling corresponds to the Hamiltonian of a topological insulator laser, which exhibits amplified edge oscillations. (b) TSS observed under the open boundary condition. The nonlinear oscillators near the boundary synchronize, while those away from the boundary exhibit chaotic dynamics instead of being synchronized. (c),(d) Frequency of the first component of oscillators at each site. We numerically calculate the dynamics of the first model of the TSS (Eqs. (<ref>), (<ref>)). Panels (c) and (d) are the snapshots at the time t=100 and t=200 for each. Empty sites represent the oscillators oscillating around their natural frequencies. We confirm that the frequencies of the edge oscillators are almost homogeneous and constant, which indicates their frequency-synchronization. Meanwhile, the oscillators inside exhibit inhomogeneous frequencies. One can also confirm that their frequencies vary over time, thus indicating the desynchronization and instability of the bulk oscillators. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2. (e) Amplitudes of the oscillators. Here, we numerically simulate the dynamics of the model as in panels (c) and (d). We check that the amplitudes of most of the edge oscillators are larger than those of the bulk oscillators. The same parameters as panels (c) and (d) are used in the calculation.	2012.09479	figure1.jpg
0.418368	39e44e4776f3406e8d41bb48c9db67fe	On-demand pattern designing of the synchronized oscillators. We numerically demonstrate that the on-demand pattern designing of the synchronized oscillators is possible by utilizing the TSS. To arrange the synchronized oscillators in the shape of characters, UT, we damp the oscillators encircled by the green boxes and numerically calculate the dynamics of the first model of topological synchronization (cf. Eqs. (<ref>), (<ref>) and Fig. <ref>) under the periodic boundary condition. The parameters used are u=-1, b=0.3, α=0.5, α_d=-100, β=1, ω_0=0.1, and Δω=0.2. In panel (a), the frequencies of the oscillators are shown. The oscillators around the damped ones exhibit a homogeneous frequency, indicating the frequency-synchronization of those oscillators. We obtain the desired shape of the collection of the synchronized oscillators. In panel (b), the indicator of the local order parameter M(x,y) in Eq. (<ref>) at each site is shown. We also confirm the appearance of the desired synchronization pattern from this result.	2012.09479	figure10.jpg
0.429163	fc4a26f4352842a99190af33d0a264d4	Applications of the TSS to the defect detection. (a),(b) Schematics of detection mechanism. We assume that the dynamics of nonlinear oscillators are detected by sensors. Information of the oscillators' state from the sensors is processed to judge the synchronization of oscillators and the existence of defects. If oscillators are broken and stop their self-oscillation, the oscillators around them are synchronized as shown in panel (a). By observing such synchronization, we can detect the breakdown of oscillators. When sensors are disordered and send no signals, the oscillators remain desynchronized as shown in panel (b). Therefore, this kind of breakdown of sensors is distinguishable from the breakdown of oscillators. (c) The indicator of the local order parameter M(x,y) in Eq. (<ref>) at each site under the existence of the broken oscillators. Large values of the indicator show the synchronization of the oscillators at the corresponding site. One can judge that the oscillators surrounded by synchronized ones are broken. The parameters used are u=-1, b=0.5, α=0.5, α_d=-100, β=1, ω_0=0.1, and Δω=0.2. (d) The indicator M(x,y) at each site under the existence of the broken sensors. The indices are small everywhere, which shows that there are no synchronized oscillators, and thus no oscillators are broken. The parameters used are the same as panel (c).	2012.09479	figure11.jpg
0.484298	b2adbf3b25844666b6d573e15aed92d2	Proposed electrical circuit to realize topological synchronization. (a) Linear system described by the effective Hamiltonian of topological insulator laser constructed by an electrical circuit using capacitors, inductors, resistors, and negative impedance converters with current inversion (INICs). We consider two layers of square lattices, and each lattice point (a red dashed circle) has two sites (black dots). The dynamics of the voltages at the sites follows the Hamiltonian of topological insulator laser in Eq. (<ref>). The left inset represents the detail of the circuit in each layer (corresponding to the enlarged view of the sites encircled by the blue dot curve in the middle panel). The middle panel shows how to connect the sites of the different layers. The legend shows the correspondence of the figures and the circuit elements. (b) Substitution rules to construct the electrical circuit of TSS. We can construct the model of the TSS by replacing the circuit elements in panel (a). We substitute each site (a black dot) into two van-del-Pol circuits (blue dashed squares) coupled by an INIC. Green-filled boxes represent the nonlinear resistors. The other panels represent the substitution rules of capacitors, inductors, resistors, and INICs. We only use resistors and INICs to avoid frequency dependence of impedance. Red dashed squares represent the coupled van der Pol circuits in the left panel.	2012.09479	figure12.jpg
0.48385	3b2cc527a5334af1aba33a2fa08585e2	Dispersion relation of the lasing modes localized at the right edge. We calculate the dispersion relation of the Hamiltonian of topological lasing modes (<ref>). We consider the open (periodic) boundary condition in the x (y) direction and arrange 20 sites in the x direction. We obtain a gapless edge band with positive imaginary parts of eigenvalues. This gapless band corresponds to lasing edge modes localized at the right side. The parameters used are u=-1 and γ=0.5.	2012.09479	figure13.jpg
0.43492	9d7da78c609a425aa39ae42cd7ff1ccc	Cluster synchronization from a topological Hamiltonian. We calculate the dynamics of the Stuart-Landau oscillators with linear coupling described by a Hamiltonian of topological lasing modes that is different from the Hamiltonians considered in the main text (<ref>). Panels (a) and (b) show the snapshots of the frequency distributions of the first component of oscillators at the time t=1000 and t = 2000 for each. The bulk and left-side oscillators exhibit homogeneous and constant frequency, which is different from that of the right-side oscillators. Therefore, the bulk oscillators form a synchronized cluster, indicating the absence of the TSS. The parameters used are u=-1, γ=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.	2012.09479	figure14.jpg
0.412836	8e06ea0dbbc7499c9df52a314b296428	Coexistence of synchronized edge oscillators and damped bulk oscillators. (a) Frequency of the first component of oscillators at each site in a damping parameter region. We numerically simulate the first model of the TSS (Eqs. (<ref>), (<ref>)) at the parameters u=-1, b=0.5, α=-0.2, β=1, ω_0=1, and Δω=0. The figure shows the snapshot at time t=100. The edge oscillators exhibit homogeneous frequencies and thus synchronize. (b) Amplitudes of the first components of oscillators at each site at time t=100. We can confirm that the bulk oscillators exhibit almost zero amplitudes. The parameters used are the same as in panel (a).	2012.09479	figure15.jpg
0.433622	9eb08df2845f4dd6af510e654fbbd246	TSS without fluctuations of the natural frequencies. We simulate the first model of the TSS (Eqs. (<ref>), (<ref>)) under the condition of the homogeneous natural frequencies. Panels (a) and (b) show the snapshots of the frequency distributions of the first component of oscillators at the time t=100 and t=200 for each. As in the case of the fluctuating natural frequencies (Fig. <ref>(c),(d)), the edge oscillators exhibit homogeneous and constant frequencies, while the bulk ones vibrate at space- and time-varying frequencies. Therefore, the edge oscillators synchronize, while the bulk ones desynchronize, which indicates the emergence of the TSS. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.	2012.09479	figure16.jpg
0.523824	bfc96dbb14e34a64b1ae46240d377ed2	Dispersion relation of the linearized equation of one-dimensional chain model. We numerically calculate the dispersion relation of the coefficient matrix obtained from the linearization of the one-dimensional chain model around the stationary solution (Eqs. (<ref>), (<ref>)). The upper (lower) panel shows the real (imaginary) part of the eigenvalues. All the eigenvalues have nonpositive imaginary parts, which implies the linear stability of the stationary solution. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=0, and Δω=0.	2012.09479	figure17.jpg
0.400378	c0abf1318b6c425e8c3c79a0da4b72bf	Robustness of TSS against disorders in linear couplings. We numerically simulate the model of the TSS (Eqs. (<ref>), (<ref>)) with the inhomogeneous linear couplings. Panels (a) and (b) show the frequency distributions at time t=100 and t=200 for each. The edge oscillators exhibit homogeneous and constant frequencies, while the bulk ones oscillate at time- and space-varying frequencies. Therefore, we obtain the TSS robustly against the noise in linear couplings. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.	2012.09479	figure18.jpg
0.41893	20a68d9534c84bf6b3135232ed98c7aa	Dispersion relation of the effective Hamiltonian. We calculate the dispersion relation of the effective Hamiltonian obtained from the linearization of the first model (cf. Eqs. (<ref>), (<ref>)). We consider the open (periodic) boundary condition in the x (y) direction and arrange 20 sites in the x direction. While the bulk bands are gapless as opposed to conventional topological systems, the edge localized modes still exist in this dispersion relation. The edge modes exhibit negative imaginary parts of the eigenvalues. In contrast, some bulk modes possess positive imaginary parts of the eigenvalues depicted by the red dashed circle, which indicates instability of the bulk oscillators. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.	2012.09479	figure19.jpg
0.428422	56a8574d23c944a3b2f0273bf4e63efd	Phase diagram of the model of the TSS. The ratio of frequency fluctuations of the edge oscillators to those of the bulk ones is depicted at each parameter. We calculate the dynamics of the model of the TSS (Eqs. (<ref>), (<ref>)) under different strengths of the linear coupling and the inhomogeneity of the natural frequencies. The red region (right bottom) shows parameter regimes where the TSS emerges. The circles and squares represent the classification of data of frequency fluctuations of the edge and bulk oscillators obtained by a numerical clustering method. The circles correspond to the parameters for which the presence of the TSS is identified. The phase boundary obtained here corresponds to the parameter points at which the bandgap of the effective Hamiltonian closes and thus the bulk band can change its topology.	2012.09479	figure2.jpg
0.445635	22ff9df41b854dbbb395b5a61973e7f6	Amplitude distribution indicating amplitude chaos. The amplitudes of oscillators at the first components of lattice points are shown. To obtain the amplitude distribution, we numerically calculate the dynamics of the first model of TSS (cf. Eqs. (<ref>), (<ref>)). The red-filled circles represent the site with small amplitudes. The existence of such small-amplitude oscillators indicates amplitude chaos. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.	2012.09479	figure20.jpg
0.48224	81abffe99d6d4bf4836010abd8472d9e	Inverse participation ratios of the Lyapunov vectors in the wavenumber space. We numerically calculate and plot the inverse participation ratios of the Fourier components of the Lyapunov vectors obtained from the model of the TSS (Eqs. (<ref>), (<ref>)). The relative indices correspond to those in Fig. <ref>. We confirm the large IPRs at small relative indices (smaller than about 0.05) and large relative indices (larger than about 0.95). The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=0.2, and Δω=0.2.	2012.09479	figure21.jpg

0.506329

ef28484d1b124e4cb5a91945a636686a

Simulation for α = 0.7.

2012.09317

path_alpha07.jpg

0.456111

1fb27033b5da40b8813830cc1b6bab34

Simulation for α = 0.8.

2012.09317

path_alpha08.jpg

0.558488

1c21a41e86944ab1b5ca97de2376eb45

Simulation for α = 0.9.

2012.09317

path_alpha09.jpg

0.443299

a734c812238944658cfbf31c39e28f89

Simulation for α = 1.

2012.09317

path_alpha10.jpg

0.436989

3d04fe5e80e44497a6b2a1703727b82d

Summary of two types of process which occur in the brain and seem to map onto Kahneman's System I and System II.

2012.09318

two_types_of_Process.jpg

0.487699

4116ba1c186341849dda5d6b37dcd615

A Comparison of the absolute bias of the R-estimator with its standard error

2012.09445

Figure0.jpg

0.467319

7eb7ecf4f608433086a919dfed32e889

Estimate of the Tail Parameter in 1,000 Years of Simulated Monthly Data

2012.09445

Figure2.jpg

0.412265

40d42e46873b44e38ae31ec35e4463d3

The Lorenz Curve for a Single Simulation

2012.09445

Figure3.jpg

0.434534

9bad90e7230640f1b9d97fe52571cec6

300 Years of Simulated Weekly Data in an Economy with One Million People

2012.09445

Figure5.jpg

0.529833

581961dcc5844a439d58df7df7880baa

The Evolution of Average Beliefs When Agents are Least-Squares Learners with Infinite Lives

2012.09445

Picture1.jpg

0.474894

1ea87597fece47dc8a3b8128ec240a44

The Evolution of Average Beliefs When Agents are Constant-Gain Learners with Finite Lives

2012.09445

Picture3.jpg

0.455372

8a1dc28f81ff49a6b01f1dbe92145802

ALMA Band-6 images before tapering of 12 targets (image size: 5”×5”). A black circle shows the beam size of each image. Black contours correspond to 4σ, 8σ, 12σ, and 16σ. A plus mark represents the centroid determined in the Ks-band image. Some of the ALMA detected sources, such as 406444 and 434585, show a larger spatial offset between the dust continuum emission and the Ks-band centroid than the other detected sources. This is probably due to the lower signal-to-noise ratios of their dust continuum emission. The stacked image of the five individually ALMA non-detected sources with log(M_*/M_⊙)= 10.0–10.4 is also shown. The central position is shown with a plus mark. The stacked emission is detected at 5σ.

2012.09447

figure1.jpg

0.487482

9329a830cb564b71900bcdb7c73532dd

Gas mass fraction versus gas-phase metallicity for our galaxy sample at z∼3.3. The horizontal bar in the bottom right corner shows an additional ± 1σ error on the gas mass fraction coming from the systematic uncertainty on M_gas. The dashed lines show how the two quantities depend on each other when fixing the dust-to-stellar mass ratio at 1×10^-3 and 3×10^-3. We find no statistically significant correlation between the gas mass fraction and gas-phase metallicity among our sample.

2012.09447

figure5.jpg

0.542845

a6c016e978e14b5a8c3592b377cd6668

Relation between gas mass fraction and gas-phase metallicity for star-forming galaxies from z=0 to z∼3.3. The horizontal bar in the left bottom corner represents an additional ± 1σ error on the gas mass fraction coming from the systematic uncertainty on M_gas for our sample. Only the molecular gas components are considered for the galaxies in <cit.> and <cit.>. The star-forming galaxies at z≳2 show an offset toward the lower gas-phase metallicity from the distribution of the local galaxies. The black lines show the model tracks from the gas regulator model of <cit.> assuming different mass-loading factors between λ= 0.5 and 2.5. The distribution of the star-forming galaxies at z∼3.3 on this diagram can be broadly explained with the model tracks with λ∼ 2–2.5, suggesting the redshift evolution of the mass-loading factor for star-forming galaxies.

2012.09447

figure6.jpg

0.488328

8ce8df4de8b24939a2c2687b91746299

Best-fit SEDs (sold line) obtained with magphys. Data points are from the COSMOS2015 catalog <cit.> and the ALMA Band-6 observation in this study. Arrows show 3σ upper limits. When fitting the ALMA non-detected sources, the 1.3 mm flux and its uncertainty are set to 1.5σ± 1σ <cit.>.

2012.09447

figure7.jpg

0.409336

e5f0782c95ff442486ce529844cacb46

Frequency histogram of weighted average of relative error of estimated ATE for eight different models: ordinary least square(OLS), LASSO, RIDGE, random forest (RF), GRU, CNN, and multilayer perception (MLP) based on the semi-synthetic data with N=160000, M=500, and α=0.05 and β=0.05 in (<ref>) of the main paper.

2012.09448

ATE_JD_160000_500_8_model.jpg

0.434653

efcf44aa6e5c4297b680f196adbff6bf

Weighted average of relative err. of estimated ATE using IoC and IwC for different models; N=160000, M=500, and α=0.05 and β=0.05 in (<ref>).

2012.09448

ATE_JD_160000_500_new.jpg

0.446368

f31af8a39de34973922c0d12c7cdffbd

Frequency histogram of weighted average of relative error of estimated ATTE for eight different models: ordinary least square(OLS), LASSO, RIDGE, random forest (RF), GRU, CNN, and multilayer perception (MLP) based on the semi-synthetic data with N=160000, M=500, and α=0.05 and β=0.05 in (<ref>) of the main paper.

2012.09448

ATTE_JD_160000_500_8_model.jpg

0.464055

a80322035b304afa88407869b05ec908

Weighted average of relative err. of estimated ATTE using IoC and IwC for different models; N=160000, M=500, and α=0.05 and β=0.05 in (<ref>).

2012.09448

ATTE_JD_160000_500_new.jpg

0.465121

2a44c23537b44e12a855b8c87becfae0

The mean relative error computed by (<ref>) vs. number of observations.

2012.09448

simu_cond_exp_consistency.jpg

0.402974

5b7bdf1b96d548718ed8f9ea60418717

The average standard deviation computed by (<ref>) vs. number of observations.

2012.09448

simu_cond_exp_std_consistency.jpg

0.428618

8b5d93c8b0d042daaa039b9c6f6d80b5

The mean relative error computed by (<ref>) vs. number of observations.

2012.09448

simu_exp_convergence.jpg

0.377465

0c97092d9dd34d8488c6aceaf2697dba

The average standard deviation computed by (<ref>) vs. number of observations.

2012.09448

simu_exp_std_convergence.jpg

0.422849

26ad506c557948ea9181629cd8b048b9

A photo of the MAFDS and its experimental test setup.

2012.09449

DSC_2177_opt_ausschnitt_cut_grau.jpg

0.512908

30786296a11c4746922a636b779d4a91

Plot compares density estimator of piezo-beam, cf. Subsection <ref>. A kernel density estimator based on all available experimental data (gray solid line), a kernel density estimator based on the computer model (black dashed line) and kernel density estimator based on an improved surrogate model (black solid line). The values of the outcome of the 10 experiments are indicated on the x-axis.

2012.09449

KeKo2019_PIEZO_plot.jpg

0.472585

6f923541b9f2485b8996dff46035aff3

Box plot without outliers of 10^6 estimates for the 0.95-quantiles of the n=100 absolute model errors for both computer models of the MAFDS.

2012.09449

KeOh2001_MAFDS_plot.jpg

0.498739

9df68f32baac46969c3bcc27fb7626e5

Three estimates of the cumulative density function of the experimental outcome of the MAFDS. An estimator based on experimental data (black solid line), an estimator based on computer experiments with computer model 1 (black dashed line) and computer model 2 (gray dashed line).

2012.09449

RoOb2011_MADFS_plot.jpg

0.476981

e0d5531ccdc745cab3cf608526334c82

Box plot without outliers for the 0.95-quantiles of the n=100 absolute model errors for both computer models of the MAFDS. The box plot contains B = 500 bootstrap 0.95-quantiles.

2012.09449

WoStLe2017_MAFDS_plot.jpg

0.484634

3de45e8e2794425c8848830ef350c1f4

A CAD model of the lateral vibration attenuation system with piezo-elastic supports and a sectional view of one of the piezo-elastic supports, cf. <cit.>.

2012.09449

piezo_beam.jpg

0.421755

6674ae92ed4745bbb0ba484a8d6db170

LS-F.

2012.09451

1872_F-eps-converted-to.jpg

0.416188

59f46f7791a84eeeb49c73e6a51f6dcc

LS-G.

2012.09451

1872_G-eps-converted-to.jpg

0.469548

d1fec0fc822b4cc9869ff97bca2f431f

The results with different values of k.

2012.09451

big_k_trim-eps-converted-to.jpg

0.444253

752ecb6225a04c2b8256ea27a2c5cf42

The number of blocks before and after applying LS-G and LS-F on METIS.

2012.09451

block_trim-eps-converted-to.jpg

0.459049

dda1c23b70a84dda95b129d4185ca50a

Predictions for the mass of the core and the amount of metallic elements in the envelope obtained using the 3rd order theory of figure. The published results are compared with the ones obtained by varying the d_0 parameter are shown in the figure and for a core constituted of pure water and of silicate and water.

2012.09454

fig6-c-arxiv.jpg

0.423496

8b847ac162e445a0bf4be4da6e45314a

Optical plasma diagnostics show strong self emission from dense narrow regions of the plasma. We see (a) raw interferogram measurement of the plasma, 1 ns after the interaction starts, measured with the Nomarsky interferometer (b) the retrieved electronic density of the plasma and (c) the raw streak camera measurement of the self emission of the same laser shot (including measurement lineout in white). x=0,y=0 are the coordinates of the jet shock point; the laser travels left to right and focalizes in the same point.

2012.09455

Image_worm.jpg

0.444622

8b0d461201064cf5bfa9eadf0ec8cb8a

RCF imprints converted to dose and alpha particle number density for three consecutive shots varying the gas jet density profile. The number density conversion presumes the appearance of alpha particles, a cut-off energy equal to the last signal layer and a step-wise flat spectral shape between layers. The typical FWHM divergence angle of the beam imprint is 9, all images have the same spatial scale. Note that number densities are normalized to one pixel, with respect to the resolution of RCF scans of 600 dpi. For shots #62 and #63, the presumed pre-aligned laser axis corresponds to the centre of the illustrated frames. For shot #64 this does not apply.

2012.09455

RCF_62_63_64_data.jpg

0.498145

b9060f6aeeaf415896fc87f94f3d2afc

Projectile spectrum for three consecutive shots varying the gas jet density profile, presuming an alpha particle beam impacting on the stack of RCF. RCF color channels R, G and B are independently analyzed.

2012.09455

RCF_62_63_64_spectrum.jpg

0.407792

0687f70542944dca8c54315cacff4f89

Projectile spectrum for three consecutive shots varying the gas jet density profile and transverse position, presuming an alpha particle beam impacting on the stack of RCF. RCF color channels R, G and B are independently analyzed.

2012.09455

RCF_79_85_87_spectrum.jpg

0.445226

a0c847948322455fa1ff670d95909772

Electronic temperature chart in MeV of the n_at,max = 4× 10^19cm^-3 (0.036 n_c) simulation.

2012.09455

Te_TUCAN.jpg

0.458839

f8093faaaebf4b37b436f9a1f8cc5337

Diameter evolution of an individual etch pit on a TasTrak CR-39 with strioscopic snapshots at 2, 3, 9 and 15 of etching with 6.25 N etchant at 70. The diameter evolves considerably and is not detected for 2 of etching as it may be underneath the resolution of the microscope system.

2012.09455

centerion_pit_strio_sketch.jpg

0.396265

687afba797c34bd9b54809575e447ff1

Summary of results.

2012.09455

comp.jpg

0.429131

24ed3b27bb40441fa37c7c8f1a1be7e0

Comparison between the detector RAW signal comprising photopeak and all accelerated particle species (blue full line) and a signal with photopeak only (orange dashed line).

2012.09455

comp_photopeak_norm.jpg

0.382355

5463e70af6cf41d0b9206be144dbc96e

Photodiode detector impulse response (left) and signal after deconvolution (right).

2012.09455

deconv_subplot.jpg

0.426681

228c3dbf390649698a2ffb08618323d9

E_y field chart resulting from the simulation with a density profile with n_at,max = 4× 10^20cm^-3. The blue dashed line marks the density peak.

2012.09455

ey_TUCAN.jpg

0.389623

fbf45ad50ee140e4962492ad934ed9b2

(a) Interferometric images overlaid by self-emission line-outs for shots with variation of the transverse position of the laser-gas interaction. The longest plasma channel is recorded for shot #81 with positive transverse displacement – the length of the channel decreases for the series of shots towards negative transverse displacements. The last shot of the series #83 does not agree with this assessment and shows a longer channel again. Note that (b) un-driven gas density profiles acquired prior to corresponding shots show visual changes from shot-to-shot, indicating changes to the nozzle. This agrees with a qualitative difference of the nozzle surface between shot #77 and #83, presented with fig. <ref>.

2012.09455

gas_77_80_81_82_83.jpg

0.44898

96bdf814c360425e8b0569270869b82f

Shot-to-shot variation of the gas jet density profile using a mix of Helium and Nitrogen in a ratio 1 to 9, (top row) longitudinal un-driven gas jet density profile, and (bottom row) interferometric image superposed to aligned laser beam and plasma self-emission in arbitrary units as red line, 1ns after the interaction starts. The laser is focused to x=0,y=0, the coordinates of the jet shock point, but transverse displaced by 0;±50.

2012.09455

gas_79_85_87_overview.jpg

0.507553

138f4487b1674706b8dd157debac2d39

Diameter evolution for similar groups of etch pits on a TasTrak CR-39 with snapshots at 2, and 9 of etching with 6.25 N etchant at 70. Highlighted are groups (a,b,d,e) with circles concentric to the etch pit position for the short etching time and (c) with circles concentric to the pit position for long etching time. The dashed line delimits a scratched area, we do not see the edge of the plastic sheet.

2012.09455

monstermark_pit_brightfield_sketch.jpg

0.41038

b0ac722be9ce44bbac75db5749106150

Overview on RCF results for shots #79, #85 and #87 converted from RAW to deposited dose maps.

2012.09455

qulitative_RCF_stack_Al4UEBT32EBT3.jpg

0.414494

53ee9e3a33954c2ba9449350b06cd95b

Nitrogen ions density chart (left) and zoom on one of the ion acoustic waves travelling in the down-ramp of the density profile (right) for the first simulation run with a_0 = 6.8.

2012.09455

qx_TUCAN_and_zoom.jpg

0.496641

2e9c4e4dbbf449f6adc6c707b2a5c8a1

Z-scan of etched pits on a TasTrak CR-39 after 15 of etching with 6.25 N NaOH etchant at 70. The bulk etching rate was quantified to 1.32. Experimental data points are underlaid with a relative point number density color map in grayscale. The simulated spectrally resolving species specific L(D) lobes are for alpha particles on the left hand side and for fully ionized Nitrogen ions on the right hand side. Specific impact energies are highlighted with red marker arrows (1–4). Experimental data points that overlap with the lobes in the range of their uncertainty are denoted with the closest corresponding energy on the lobe. Simulated with the CR-39 plug-in of PySTarT.

2012.09455

track_diameter_energy_15.jpg

0.393433

f384dd6127914218b1a195583a6d9dad

An analytical investigation of the critical transmissibilities of both strains for increasing clustering coefficients for the ER networks defined in Fig <ref>. The curves are generated from Eqs <ref> and <ref> at the value of T in which the outbreak fraction becomes large than ϵ. The critical points reduce as C increases, indicating that contact clustering helps to spread an emergent epidemic. The interval [T_1^*,T_2^*] is the transmissibility window in which strain-1 exists as an epidemic on the network without strain-2. We observe that increased clustering reduces the interval and thus increases the extent of coinfection in this model, at the expense of the mono-infection equilibrium. We note, however, that clustering can never reduce the coinfection critical point below the single strain threshold due to the strict conditions of perfect coinfection in the premise of the model.

2012.09457

Delta_transmissibilities.jpg

0.406831

cf355fd736d649b9b393e7f911db2b9a

(Top) The outbreak fractions of both strains for increasing triangle probabilities within the CHCN degree distribution defined in Eq <ref>. Clustering reduces the epidemic threshold, in agreement with ER experiments; however, coinfection decreases with increasing θ. (Bottom) The expected epidemic sizes of both strains for a p^CHCN(k) network with θ=0.0 and a degree-δ CHCN distribution network defined by Eq <ref>. Clustering increases the epidemic threshold in this model. Markers are the average of 100 repetitions of bond percolation on CHCN networks with N=1e^5 nodes and T_2=0.6.

2012.09457

PerfCoinfFig5PlotB.jpg

0.428417

9e21e7b4c97a421984bbdb01424cc315

The tree-like edge topologies found in the GCC following the first strain and their probabilities.

2012.09457

TREES_coinf.jpg

0.449054

32d42d02a2424f7eb789ca98b5ea1df0

The 3-cycle edge-topologies found in the GCC following the first strain and their probabilities. Node colours and patters are defined according to Fig <ref>. Triangles D, E and F consist of inhomogeneous neighbour-states and hence, due to the symmetry of the shape, their reflection about a vertical axis bisecting the focal node can also occur with equal probability. The curved arrows in triangles D and F indicate the additional pathway through the cycle that the infected neighbour could infect the focal node.

2012.09457

TRIANGLES_coinf.jpg

0.541495

ae746598d94a4951a87ab0f3d011efa2

The outbreak fractions for the 2-strain coinfection model on ER random graphs with T_2=0.6, N = 25000, for varying clustering coefficients and fixed mean degree ⟨ s⟩+2⟨ t ⟩=2. It is clear that clustering reduces the epidemic thresholds of both diseases and reduces the outbreak size of the first strain; however, it increases the coinfected fraction of the network. Points are the average of 500 repeats of Monte Carlo simulation while solid lines are the theoretical predictions from Eqs <ref> and <ref>.

2012.09457

coinfection.jpg

0.443824

eb01c3e963fd431a9eae58660704fde1

The expected number of bars of length ≥, N^_B, as a function of .

2012.09459

BMCalculated.jpg

0.445585

e5e90015606e4f589e11585d2272602b

N^_B as approximated with simulations of random Rademacher walks.

2012.09459

NvepsBMnothing.jpg

0.405588

c40d5ac8926d49b8b25d83705a18a0cc

A function f: [0,1] → and its associated tree T_f in dashed lines.

2012.09459

TreeFunction.jpg

0.433145

65a32d573bf44ee38c0d1d3885863f75

A depiction of the first steps of the algorithm which assigns a barcode (f) to a tree T_f.

2012.09459

algorithm.jpg

0.433826

1a2dcfb0306f4ad3a89510733afe1eda

Comparison between theoretical predictions and finite element simulation for twist (a) and stretch (b) as a function of the spontaneous deformation parameter λ_s. For these calcuations, we used a fibre with aspect ratio R_0/L=1/15 represented by 5400 hex-8 elements.

2012.09461

TwistStretchComparison.jpg

0.451198

4e80121ea0574c56bf57822187014d96

Finite element simulation of a fibre with Young modulus E=1, Poisson ratio ν=0.45, 4800 elements, aspect ratio of R_0/L=0.07 and a R independent director-field with α=π/4 and β=π/2. The fibre is fixed at one end and subject to a tension T=0.2 πμ R_0^2 while preventing rotation at the other end. As the value of λ_s is decreased, the fibre wants to shrink and twist but cannot do the latter as the ends are not allowed to rotate. The fibre thus coils to release some of the twist energy, resulting in a greater stroke amplitude then that of classical linear actuation.

2012.09461

coil.jpg

0.40009

2a9c956b6e814e7bb1dfecf4b3f8bfa1

Comparison between Finite element simulations and theoretical data for the case α=π/4 and β=π/4. a) An example of a deformed cross section directly from FE analysis. In red are theoretical lines for the deformed radii while dots are node positions. b) Comparison between predicted and theoretical coning. c) Comparison between theory and FE on the predicted rotation angle difference as a function of radius. A fibre with aspect ratio R_0/L=1/5, and 54000 hex-8 elements was used for this simulation.

2012.09461

coning2.jpg

0.407857

4d702e0350634fe4b7804395911f467c

a) 2D plot of twist as a function of both angles α and β. The twist is maximised when β=π/2 and α∼π/4. b) – c) Variation of the twist and stretch as a function of α for different values of λ_s.

2012.09461

costant2.jpg

0.401871

543a1573c7c14894a3347b74cded426a

Relationship between output twist a) and stretch b) as a function of the twist imposed at genesis. Shaded in grey, the region for known experimental results <cit.>.

2012.09461

genesis.jpg

0.444126

8bb867071cb64e048c5453869f4940bb

a) Schematic of a fibre being drawn and twisted from a drop of LC monomer while being cured with UV light. b) An example of the resulting director field in a fibre of radius r=2/τ_0. The two integral curves highlight the change in azimuthal component as a function of the radius.

2012.09461

genesis2.jpg

0.419085

1e886a40c1fe412da0ec9f645961dc06

(a) The optical dipole potential energy for right-handed circular polarization. Solid line represents the case where ℓ=+1, dashed curve represents the case where ℓ=-1; (b) The optical dipole potential energy for left-handed circular polarization. Solid line represents the case where ℓ=-1, dashed curve represents the case where ℓ=+1. In both (a) and (b) the dotted line shows the optical dipole potential energy without taking into account the contribution of the spin-orbit term. The insets to the figures show the potential energies of the solid curves and their projections in the focal plane. These insets should be interchanged for the dashed curves. In all plots the potential energy is in recoil energy units while the radial distances are in LG beam waist w_0 units.

2012.09465

Fig1.jpg

0.537845

6d19c5b4b60143ae834f65a3ace93ccb

(a) The optical dipole potential energy for right-handed circular polarization. Solid line represents the case where ℓ=+5, dashed curve is for ℓ=-5. (b) The optical dipole potential energy for left-handed circular polarization. Solid line represents the case where ℓ=-5, dashed curve is for ℓ=+5. In both (a) and (b) the dotted line shows the optical dipole potential energy without taking into account the contribution of the spin-orbit term. In both plots the potential energy is in recoil energy units while the radial distances are in LG beam waist w_0 units.

2012.09465

Fig2.jpg

0.498272

1eb48bec68b0485dba9fe6ea2581c29d

(a) The optical dipole potential energy for a bi-chromatic field with ℓ=1 (solid line). The plot is inverted if ℓ=-1 (dashed line). (b) Optical dipole potential energy for a bi-chromatic field with ℓ=5 (solid line). This plot is inverted if ℓ=-5. In both plots the potentials are given in recoil energy units while the radial distances are in LG beam waist w_0 units.

2012.09465

Fig3.jpg

0.442721

f26e0411027d4b1482a0e43a5e96c2d5

(a) The optical dipole potential energy, in the radial and axial directions, for a bi-chromatic field with ℓ=5 (left). (b) The plot is inverted for ℓ=-5 (right). The potential energy is given in recoil energy units, the radial distances are in LG beam waist w_0 units, while the axial distances are in Rayleigh range (z_R) units. The arrows in the plot (b) indicate the trapping regions.

2012.09465

Fig4.jpg

0.495479

748b4a0b327e4a6d859e8d55f1d7437a

2012.09467

1.jpg

0.434933

62f10df832834a91af5196337b1c9927

Comparison of the analytical (exact) and annealed result of Eq. (<ref>) using the 6th order Gauss-Legendre collocation method with a timestep Δ t = 0.5. The numerical result is visually indistinguishable from the analytical (exact) solution.

2012.09469

annealing_result.jpg

0.429897

af6bb72085b249cda2a0835ca8c71e76

Comparison of the simulated annealed, analytical, and Newton method for u(0) = 0.1. The analytical solution was obtained by using the 4th Runge-Kutta method with much smaller time step sizes. The simulated annealed solution does not match the result of the Newton method.

2012.09469

nonlinear_annealing_0.1.jpg

0.454794

2b41d3e6f4c24d68a208be3a3f23a61a

Comparison of the scaling of the energy gap between the ground and first excited state for the adiabatic Hamiltonian for the Crank-Nicolson scheme with changing Δ x. It can be seen that g_min is proportional to Δ x and higher time steps result in a lower energy gap.

2012.09469

scaling_crank_nicolson_dx.jpg

0.511214

dbdd7da906f44890b6a5e1761ab1563c

Comparison of the scaling of the energy gap between the ground and first excited state for the adiabatic Hamiltonian for the explicit Euler. For sufficient small Δ x, the expected scaling and numerical determined scaling match.

2012.09469

scaling_explicit_euler_dx.jpg

0.404099

c02ea58f53e44749ae494148d0d66c53

Error of the annealed Gauss-Legendre 6th order Runge-Kutta step using our variational approach, compared to a numerical evaluation of the same Runge-Kutta step. The dashed lines are obtained on a D-Wave quantum annealer, and the solid lines show an exact annealing reference solution calculated on a classical computer. The parameter c is the exponent shift of the variational approach in Eq. (<ref>). For all cases, two qubits per number register are used.

2012.09469

variational_convergence.jpg

0.45143

76c07e9a50834f3ca7e011d726f4eada

(Color online) Pressure as a function of baryon density ρ/ρ_0 for hadronic phase (solid line), quark phase (dash line) and mixed phase (dash-dotted line). The fitted EoSs with first-order (a) and crossover (b) phase transitions use respectively B= 150 and 90 MeV/fm^3 in the MIT model.

2012.09471

fig1.jpg

0.468463

b6db6d87f0c94daaafaabb523187fff0

(Color online) Maximum compression density reached with hadronic EoS (thick lines) and EoSs including first-order phase transition and crossover (thin lines) in Au + Au collisions at beam energies of 2-8 GeV/nucleon.

2012.09471

fig2.jpg

0.431901

72e92f1bdbb440ea9f93bc6dd74261b2

(Color online) Proton average transverse momentum ⟨ p_x⟩ as a function of rapidity in Au + Au collisions using EoSs with and without phase transition at beam energies of 2-8 GeV/nucleon. The experimental data are taken from Ref. <cit.>.

2012.09471

fig3.jpg

0.441927

14f6c0ee578c464291e8b78fdcdaa249

(Color online) Proton directed flow v_1 in Au + Au collisions using three kinds of EoSs at beam energies of 2-8 GeV/nucleon. The experimental data are taken from Ref. <cit.>.

2012.09471

fig4.jpg

0.434658

799c2f4c81354682a93d1ae739600160

(Color online) Same as Fig. <ref>, but with different phase-transition boundaries.

2012.09471

fig5.jpg

0.46452

5b559b27ce2649809c18090df3efc9b2

The formation of the total number of bonds 𝒩_ b with time t for different bond lengths ℓ_ b. The data is shown for n=3 tri-functional (dashed lines) and n=4 tetra-functional (solid lines) monomers. Parts (a) and (b) show the data for the FENE and harmonic bonds, respectively.

2012.09473

Figure1.jpg

0.532836

0e82cbb822524b78bfe51683e8130628

Number density of bonds ρ_ b (panel a) and monomers ρ_ m (panel b) as a function of bond length ℓ_ b. The data is shown for the systems after the isobaric equilibration. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.

2012.09473

Figure2.jpg

0.43542

3b7b20708601466e850311e20932f694

Panels (a-d) show the simulation snapshots of a 2σ thick layer along the z direction after the network cure in the canonical ensemble. The lateral dimensions of the snapshots are 67.03 σ. Snapshots are shown for different bond lengths ℓ_ b and functionalities n, as mentioned in the figure headings. Panels (a) and (c) are for ℓ_ b = 0.90σ and panels (b) and (d) are for ℓ_ b = 0.97σ. Panel (e) shows the fraction of the void volume v_ void as a function of ℓ_ b. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.

2012.09473

Figure3.jpg

0.476856

9316b3c0cf8c4b0a9c1368152ada4c77

The normalized thermal transport coefficient κ/κ_ linear as a function of the bond length ℓ_ b. The data is shown for different network functionality n and for the both bonds. The data is normalized with respect to the linear chain connected by the harmonic springs giving κ_ linear = 5.4±0.3 k_ B/τ m. Here the linear chain lengths are chosen as N_ℓ=50. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.

2012.09473

Figure4.jpg

0.434515

270df98126f041498b71653f06d03963

The normalized thermal transport coefficient κ/κ_ linear as a function of the volumetric specific heat c_ v (part a) and the sound wave velocity v (panel b). The data is shown for different functionality n, bond length ℓ_ b and for both bonds. The data is normalized with respect to the linear chain connected by the harmonic springs giving κ_ linear = 5.4±0.3 k_ B/τ m. Here the linear chain lengths are chosen as N_ℓ=50. Open and solid symbols are for the FENE and the harmonic bonds, respectively. Lines are drawn to guide the eye.

2012.09473

Figure5.jpg

0.536148

891b015e5dd54ce88f031664dc0d560a

Vowel-wise acoustic feature comparisons between COVID-19 negative (neg.) and COVID-19 positive (pos.) participants in form of boxplots for features with a differentiation effect r>.4 ordered from left to right according to a decreasing r, respectively. The effect size r as well as the p-value of the Mann-Whitney U difference test are given above each boxplot. p is rounded to three decimal places. r is rounded to two decimal places. Outliers (marked with red plus symbols) are defined as value that are more than 1.5 times the interquartile range away from the bottom or top of the respective box. # = number of, F0 = fundamental frequency, F1 = first vowel formant, F3 = third vowel formant, len. = length, pctlrg = percentile range, RS = rising slope, seg. = segment, ST = semitone from 27.5 Hz, SDnorm = standard deviation normalised by the arithmetic mean (coefficient of variation), slp = slope; VR = voiced regions

2012.09478

boxplots.jpg

0.434778

541175090317427499416a9170371b58

Dynamical formation of topological synchronized states (TSSs). (a) Schematic of the considered nonlinear system. Nonlinear oscillators are placed on a square lattice. Each oscillator is linearly coupled to other oscillators at the nearest neighbor sites. The linear coupling corresponds to the Hamiltonian of a topological insulator laser, which exhibits amplified edge oscillations. (b) TSS observed under the open boundary condition. The nonlinear oscillators near the boundary synchronize, while those away from the boundary exhibit chaotic dynamics instead of being synchronized. (c),(d) Frequency of the first component of oscillators at each site. We numerically calculate the dynamics of the first model of the TSS (Eqs. (<ref>), (<ref>)). Panels (c) and (d) are the snapshots at the time t=100 and t=200 for each. Empty sites represent the oscillators oscillating around their natural frequencies. We confirm that the frequencies of the edge oscillators are almost homogeneous and constant, which indicates their frequency-synchronization. Meanwhile, the oscillators inside exhibit inhomogeneous frequencies. One can also confirm that their frequencies vary over time, thus indicating the desynchronization and instability of the bulk oscillators. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2. (e) Amplitudes of the oscillators. Here, we numerically simulate the dynamics of the model as in panels (c) and (d). We check that the amplitudes of most of the edge oscillators are larger than those of the bulk oscillators. The same parameters as panels (c) and (d) are used in the calculation.

2012.09479

figure1.jpg

0.418368

39e44e4776f3406e8d41bb48c9db67fe

On-demand pattern designing of the synchronized oscillators. We numerically demonstrate that the on-demand pattern designing of the synchronized oscillators is possible by utilizing the TSS. To arrange the synchronized oscillators in the shape of characters, UT, we damp the oscillators encircled by the green boxes and numerically calculate the dynamics of the first model of topological synchronization (cf. Eqs. (<ref>), (<ref>) and Fig. <ref>) under the periodic boundary condition. The parameters used are u=-1, b=0.3, α=0.5, α_d=-100, β=1, ω_0=0.1, and Δω=0.2. In panel (a), the frequencies of the oscillators are shown. The oscillators around the damped ones exhibit a homogeneous frequency, indicating the frequency-synchronization of those oscillators. We obtain the desired shape of the collection of the synchronized oscillators. In panel (b), the indicator of the local order parameter M(x,y) in Eq. (<ref>) at each site is shown. We also confirm the appearance of the desired synchronization pattern from this result.

2012.09479

figure10.jpg

0.429163

fc4a26f4352842a99190af33d0a264d4

Applications of the TSS to the defect detection. (a),(b) Schematics of detection mechanism. We assume that the dynamics of nonlinear oscillators are detected by sensors. Information of the oscillators' state from the sensors is processed to judge the synchronization of oscillators and the existence of defects. If oscillators are broken and stop their self-oscillation, the oscillators around them are synchronized as shown in panel (a). By observing such synchronization, we can detect the breakdown of oscillators. When sensors are disordered and send no signals, the oscillators remain desynchronized as shown in panel (b). Therefore, this kind of breakdown of sensors is distinguishable from the breakdown of oscillators. (c) The indicator of the local order parameter M(x,y) in Eq. (<ref>) at each site under the existence of the broken oscillators. Large values of the indicator show the synchronization of the oscillators at the corresponding site. One can judge that the oscillators surrounded by synchronized ones are broken. The parameters used are u=-1, b=0.5, α=0.5, α_d=-100, β=1, ω_0=0.1, and Δω=0.2. (d) The indicator M(x,y) at each site under the existence of the broken sensors. The indices are small everywhere, which shows that there are no synchronized oscillators, and thus no oscillators are broken. The parameters used are the same as panel (c).

2012.09479

figure11.jpg

0.484298

b2adbf3b25844666b6d573e15aed92d2

Proposed electrical circuit to realize topological synchronization. (a) Linear system described by the effective Hamiltonian of topological insulator laser constructed by an electrical circuit using capacitors, inductors, resistors, and negative impedance converters with current inversion (INICs). We consider two layers of square lattices, and each lattice point (a red dashed circle) has two sites (black dots). The dynamics of the voltages at the sites follows the Hamiltonian of topological insulator laser in Eq. (<ref>). The left inset represents the detail of the circuit in each layer (corresponding to the enlarged view of the sites encircled by the blue dot curve in the middle panel). The middle panel shows how to connect the sites of the different layers. The legend shows the correspondence of the figures and the circuit elements. (b) Substitution rules to construct the electrical circuit of TSS. We can construct the model of the TSS by replacing the circuit elements in panel (a). We substitute each site (a black dot) into two van-del-Pol circuits (blue dashed squares) coupled by an INIC. Green-filled boxes represent the nonlinear resistors. The other panels represent the substitution rules of capacitors, inductors, resistors, and INICs. We only use resistors and INICs to avoid frequency dependence of impedance. Red dashed squares represent the coupled van der Pol circuits in the left panel.

2012.09479

figure12.jpg

0.48385

3b2cc527a5334af1aba33a2fa08585e2

Dispersion relation of the lasing modes localized at the right edge. We calculate the dispersion relation of the Hamiltonian of topological lasing modes (<ref>). We consider the open (periodic) boundary condition in the x (y) direction and arrange 20 sites in the x direction. We obtain a gapless edge band with positive imaginary parts of eigenvalues. This gapless band corresponds to lasing edge modes localized at the right side. The parameters used are u=-1 and γ=0.5.

2012.09479

figure13.jpg

0.43492

9d7da78c609a425aa39ae42cd7ff1ccc

Cluster synchronization from a topological Hamiltonian. We calculate the dynamics of the Stuart-Landau oscillators with linear coupling described by a Hamiltonian of topological lasing modes that is different from the Hamiltonians considered in the main text (<ref>). Panels (a) and (b) show the snapshots of the frequency distributions of the first component of oscillators at the time t=1000 and t = 2000 for each. The bulk and left-side oscillators exhibit homogeneous and constant frequency, which is different from that of the right-side oscillators. Therefore, the bulk oscillators form a synchronized cluster, indicating the absence of the TSS. The parameters used are u=-1, γ=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.

2012.09479

figure14.jpg

0.412836

8e06ea0dbbc7499c9df52a314b296428

Coexistence of synchronized edge oscillators and damped bulk oscillators. (a) Frequency of the first component of oscillators at each site in a damping parameter region. We numerically simulate the first model of the TSS (Eqs. (<ref>), (<ref>)) at the parameters u=-1, b=0.5, α=-0.2, β=1, ω_0=1, and Δω=0. The figure shows the snapshot at time t=100. The edge oscillators exhibit homogeneous frequencies and thus synchronize. (b) Amplitudes of the first components of oscillators at each site at time t=100. We can confirm that the bulk oscillators exhibit almost zero amplitudes. The parameters used are the same as in panel (a).

2012.09479

figure15.jpg

0.433622

9eb08df2845f4dd6af510e654fbbd246

TSS without fluctuations of the natural frequencies. We simulate the first model of the TSS (Eqs. (<ref>), (<ref>)) under the condition of the homogeneous natural frequencies. Panels (a) and (b) show the snapshots of the frequency distributions of the first component of oscillators at the time t=100 and t=200 for each. As in the case of the fluctuating natural frequencies (Fig. <ref>(c),(d)), the edge oscillators exhibit homogeneous and constant frequencies, while the bulk ones vibrate at space- and time-varying frequencies. Therefore, the edge oscillators synchronize, while the bulk ones desynchronize, which indicates the emergence of the TSS. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.

2012.09479

figure16.jpg

0.523824

bfc96dbb14e34a64b1ae46240d377ed2

Dispersion relation of the linearized equation of one-dimensional chain model. We numerically calculate the dispersion relation of the coefficient matrix obtained from the linearization of the one-dimensional chain model around the stationary solution (Eqs. (<ref>), (<ref>)). The upper (lower) panel shows the real (imaginary) part of the eigenvalues. All the eigenvalues have nonpositive imaginary parts, which implies the linear stability of the stationary solution. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=0, and Δω=0.

2012.09479

figure17.jpg

0.400378

c0abf1318b6c425e8c3c79a0da4b72bf

Robustness of TSS against disorders in linear couplings. We numerically simulate the model of the TSS (Eqs. (<ref>), (<ref>)) with the inhomogeneous linear couplings. Panels (a) and (b) show the frequency distributions at time t=100 and t=200 for each. The edge oscillators exhibit homogeneous and constant frequencies, while the bulk ones oscillate at time- and space-varying frequencies. Therefore, we obtain the TSS robustly against the noise in linear couplings. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.

2012.09479

figure18.jpg

0.41893

20a68d9534c84bf6b3135232ed98c7aa

Dispersion relation of the effective Hamiltonian. We calculate the dispersion relation of the effective Hamiltonian obtained from the linearization of the first model (cf. Eqs. (<ref>), (<ref>)). We consider the open (periodic) boundary condition in the x (y) direction and arrange 20 sites in the x direction. While the bulk bands are gapless as opposed to conventional topological systems, the edge localized modes still exist in this dispersion relation. The edge modes exhibit negative imaginary parts of the eigenvalues. In contrast, some bulk modes possess positive imaginary parts of the eigenvalues depicted by the red dashed circle, which indicates instability of the bulk oscillators. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.

2012.09479

figure19.jpg

0.428422

56a8574d23c944a3b2f0273bf4e63efd

Phase diagram of the model of the TSS. The ratio of frequency fluctuations of the edge oscillators to those of the bulk ones is depicted at each parameter. We calculate the dynamics of the model of the TSS (Eqs. (<ref>), (<ref>)) under different strengths of the linear coupling and the inhomogeneity of the natural frequencies. The red region (right bottom) shows parameter regimes where the TSS emerges. The circles and squares represent the classification of data of frequency fluctuations of the edge and bulk oscillators obtained by a numerical clustering method. The circles correspond to the parameters for which the presence of the TSS is identified. The phase boundary obtained here corresponds to the parameter points at which the bandgap of the effective Hamiltonian closes and thus the bulk band can change its topology.

2012.09479

figure2.jpg

0.445635

22ff9df41b854dbbb395b5a61973e7f6

Amplitude distribution indicating amplitude chaos. The amplitudes of oscillators at the first components of lattice points are shown. To obtain the amplitude distribution, we numerically calculate the dynamics of the first model of TSS (cf. Eqs. (<ref>), (<ref>)). The red-filled circles represent the site with small amplitudes. The existence of such small-amplitude oscillators indicates amplitude chaos. The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=1, and Δω=0.2.

2012.09479

figure20.jpg

0.48224

81abffe99d6d4bf4836010abd8472d9e

Inverse participation ratios of the Lyapunov vectors in the wavenumber space. We numerically calculate and plot the inverse participation ratios of the Fourier components of the Lyapunov vectors obtained from the model of the TSS (Eqs. (<ref>), (<ref>)). The relative indices correspond to those in Fig. <ref>. We confirm the large IPRs at small relative indices (smaller than about 0.05) and large relative indices (larger than about 0.95). The parameters used are u=-1, b=0.5, α=0.5, β=1, ω_0=0.2, and Δω=0.2.

2012.09479

figure21.jpg