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
|
|---|---|---|---|---|
0.470613 | cf7fb5d0c6804b9aad747c7b79d0ebba | Graphical illustration of the random strategy. A player starts from any box of their choice, and then keeps opening boxes randomly among the ones that are closed until the key is found or total number of attempts are over. | 2012.13484 | random_strategy.jpg |
0.452198 | a7b59b66deb54971922014750c98c571 | Velocity profiles for Couette flow with different velocity sets and Knudsen numbers. (Left) Normalized streamwise velocity profiles. (Right) Predicted Knudsen layer defined as a deviation from a straight line profile constrained to pass through {12, 0}. | 2012.13487 | Couette_profile.jpg |
0.442535 | c01fa2f6cdcc4d3694ef563b839ac4e3 | Contours of normalized u_x at Kn = 1 for (left) GH D2Q16 and (right) NC D2Q1089. (a-b) uniform structured mesh with 200 triangle elements, (c-d) uniform structured mesh with 450 triangle elements, and (e-f) unstructured mesh with 315 triangle elements. | 2012.13487 | LDCF_contour.jpg |
0.437158 | 5f71768759c74214a3c61c8d4872f079 | Normalized u_x (red) and u_y (blue) along the cavity center for different Knudsen numbers. The DSMC data is replotted using results from reference <cit.>. | 2012.13487 | LDCF_profiles.jpg |
0.418602 | 8b1a33faffe94e5cb4cd3870c6fc1d99 | Comparison of u_x along the cavity center line for different GH and NC quadratures for (left) Kn=1, (right) Kn=8. The DSMC data is replotted using results from reference <cit.>. | 2012.13487 | NCvsGH.jpg |
0.477717 | 180a0bcda820435c988ce4f7c8fcc19e | Comparison of normalized slip velocity for Couette flow at different values of Kn. DSMC values replotted using data from Ansumali et al.<cit.> | 2012.13487 | Slip.jpg |
0.451037 | e4f2ca51c0564af3ba041b35251960b6 | L_2 error in the streamwise velocity u_x for Couette flow vs. mesh size h for different polynomial degrees p. | 2012.13487 | convergence.jpg |
0.449419 | 505a5a4fa7694a8dbb04475d54786fd9 | Unstructured mesh comprised of 315 triangle elements used in the micro-cavity simulations. | 2012.13487 | mesh.jpg |
0.462025 | eab05a855ee64654a7f8eb13b098688a | Experimental setup. (a) Single photon source preparation. A 405nm continuous wave (CW) diode laser with 20mW power pumps a periodically poled potassium titanyI phosphate (PPKTP) crystal to produce photon pairs with central wavelength of 810 nm based on SPDC. The H0 and PBS0 are used to regulate optical power. The two lenses (L1 and L2) before and after PPKTP are used for focusing and collimating beams respectively. Then, after filtering out the pumped laser with a long pass filter (LP), the photon pairs are split on PBS1 and are coupled into single-mode fibers (SMF). (b) State preparation, state evolution and state measurement. The numbers on the figure represent the position of the walker. The unlabeled blue objects in the figure are all glass plates used to compensate for the phase. Ideally, after BD6, only a HWP is needed on the path |2⟩. However, in the experiment we used HWP15 that spanned three paths. Therefore, HWP16 and HWP17 are needed to ensure that the polarization of |0⟩ and |4⟩ paths are in |V⟩ mode. Finally, we detect photons at the transmission end of PBS3 by adjusting HWP18 and HWP19. The quantum walk photons are detected by SPAD1, whereas the heralding photon is detected by SPAD0. The effective coincidence window (including the jitter of the detector) is about 2ns. | 2012.13488 | graph1.jpg |
0.449831 | bcffce61ff594812b6a102b3c4aa8d45 | Second-order correlation of a single-photon source. There is a clear dip in 160 time bins and every bin is 64ps. | 2012.13488 | graph2.jpg |
0.453349 | 9373f04351664b5eb58d2b20b3788156 | Experimental setup for measuring single observable. |post(2)⟩ (|post(2̅)⟩, which represents the orthogonal of |post(2)⟩) is the eigenstate of Π_post(2) corresponding to the eigenvalue -1(+1). The number accompanying each HWP is the angle of its optical axis relative to the horizontal polarization direction. The red line represents the |H⟩ polarization in the path, and the blue line represents the |V⟩ polarization. The red(blue) dotted line shows that there is no photon in the |H⟩ polarization(|V⟩ polarization) of this path. | 2012.13488 | graph3.jpg |
0.445992 | 0ee2b90075064548a67b163b740fd0b1 | Experimental results. (a) Coincidence photon counts. The distribution of the number of photons measured post-selection in three setups. The x-coordinate represents the three experimental setups, and the y-coordinate represents the coincidence photon counts. Blue bars represent the coincidence photon counts between SPAD0 and the output of Π_post(2). Red bars represent the coincidence photon counts between SPAD0 and the outputs of the complement of Π_post(2), i.e., Π̅_post(2). Purple bars represent the total coincidence photon counts. (b) The probability distributions for the three setups. All of them are calculated from the raw data. | 2012.13488 | graph4.jpg |
0.55259 | 9be70b79f7744d8382cb2f07f2b9c876 | Accuracy on MNIST dataset with different embedding dimension sizes. Our method performs well even when the output dimension is reduced to two, which is a desirable property for data visualization. | 2012.13489 | O_dim.jpg |
0.45306 | 3b50e5dd7b334852922e3311cb5f6bd5 | Illustration of Locality Preserving Variational Discriminative Network (LPVDN). Our model retains both global (lower part) and local (upper part) structure. Latent variables μ and σ are regularized to a mixture of Gaussian distribution. | 2012.13489 | net.jpg |
0.365835 | 10f307b683dc4e25892f618a0fea09fa | The 10 nearest neighbors retrieved using our proposed model, VaDE and raw data on MNIST dataset. The picture on the left is the query picture of digit “4”, which looks similar to “9”. | 2012.13489 | nn_crop.jpg |
0.454787 | da89b78caf2342679c2b0117c97e4356 | Experimental scheme. a. Illustration of f–2f interferometry using octave-spanning soliton microcombs and second-harmonic generators in a nanophotonic platform harboring simultaneous χ^(3) and χ^(2) susceptibilities. The offset frequency f_ceo is accessible from the beatnotes of δ_1 and δ_2, and f_p is the pump laser frequency. b. Top: sketch of a hexagonal AlN layer (lattice constants: a and c) epitaxially grown on a c-plane sapphire substrate. Bottom: unit cell of an AlN crystal (left) and photograph image of a 2-inch AlN wafer featuring a broad transparency window from ultraviolet to mid-infrared regimes (right). c. Spectroscopic ellisometer mapping of the AlN film thickness in a region of 4 × 9 mm^2, showing a minor variation of 1000±5 nm denoted by the right color bar. | 2012.13496 | Figure_1.jpg |
0.417829 | 098f4800d3c64712ad7c3eb318963650 | Octave soliton microcombs at hundreds of gigahertz repetition rates. a. Top: integrated dispersion (D_int) of a 50 μm-radius AlN resonator (cross section: 1.0 × 2.3 μm^2), where the anomalous and normal GVD regimes are shaded with light blue and orange colors respectively. Insets: zoom-in view of the measured (red dots) and simulated (blue curve) values and the resonator modal profile shown in the right. Bottom: soliton microcomb spectra from the experiment (blue) and simulation (red) at an on-chip pump power of ∼390 mW. The resonator Q_int is 1.6 million and the f_rep is estimated to be around 433 GHz. b. Soliton microcomb spectrum from a 100 μm-radius AlN resonator (cross section: 1.0 × 3.5 μm^2) with a decreased f_rep of ∼220 GHz. The applied pump power is ∼1 W at a resonator Q_int of 3.0 million. | 2012.13496 | Figure_2.jpg |
0.414416 | e324fb4a4513497c9f0a7f863bceac7f | Octave soliton microcombs with agilely tunable spectra. a. Engineered D_int curves of 60 μm-radius AlN resonators at varied widths of 2.3–2.5 μm revealed by the colored shadow regime. b. Corresponding soliton microcomb spectra at resonator widths of 2.3, 2.4, and 2.5 μm from the top to bottom panel, respectively. The f_rep is ∼360 GHz, while the vertical arrows in spectral wings indicate the emergence of DWs. Akin to Fig. <ref>a, high-frequency DWs here also exhibit an evident blue shift from the D_int = 0 position. From a sech^2 fit, the corresponding temporal pulse duration is estimated to be ∼23, 22 and 19 fs (from top to bottom), respectively. | 2012.13496 | Figure_3.jpg |
0.39578 | 5512cfd6454f46a9a98fc179c6d4d46e | Second-harmonic generators. a. Colored scanning electron microscope images of fabricated AlN nanophotonic chips composed of octave microcomb generators (microring resonators) and SHG waveguides (total length of 6 cm, not fully shown). b. SHG spectra collected from a modal phase-matched waveguide (width of 1.395 μm) at an on-chip 1f power of 355 mW. Insets: modal profiles of the pump (TM_00) and SHG (TM_20) waves as well as the wavelength-dependent SHG power (pink dots), where a sinc^2-function fit (blue curve) is applied. The error bars reflect the SHG power variation from continuous three measurements. | 2012.13496 | Figure_4.jpg |
0.426057 | 6045cd89f68a41b6b7a73f3a94c9a9c5 | f–2f heterodyne measurement. a. Schematic diagram for assessing the f_ceo. The symbol “×2” indicates a RF frequency doubler. The experimental details are introduced in Supplementary Section III. b. Free-running f–2f beatnotes after the down-conversion process, suggesting a signal-to-noise ratio of 10 dB at a resolution bandwidth of 1 MHz. The local oscillator frequencies f_LO1 and f_LO2 are chosen to be 11.8 and 9.1 GHz, respectively. Inset: the equivalent curve of f_ceo = 2f_LO1 + f_LO2 – Δ f_2. | 2012.13496 | Figure_5.jpg |
0.452989 | 5683e18451c3457a826a30eb1b067c54 | Nanophotonic device fabrication. a. Wafer-scale thickness mapping of a 2-inch AlN wafer using a spectroscopic ellipsometer. The right color bar indicates the corresponding thickness variation. b. Surface roughness characterization of the AlN film using an atomic force microscope. c. Transmittance of a 50 μm-radius AlN resonator (width = 2.3 μm) for octave-soliton generation. A zoom-in view of the resonance (indicated by dashed lines) around 1542 nm is shown in the right side, revealing a Q_int of ∼1.6 million. d. Resonance from an 100 μm-radius AlN resonator (width = 3.5 μm), highlighting a reduced resonant linewidth and an improved Q_int of ∼3.0 million. | 2012.13496 | Figure_S1.jpg |
0.444883 | c7926f32b36a4a47bc1ea2b441787416 | Dispersion engineering for octave soliton generation. a. Height-dependent integrated dispersion (D_int) of 50 μm-radius AlN resonators at a fixed width of 2.3 μm. b. Top: width-dependent D_int curves (radius of 50 μm, height of 1.00 μm). Bottom: numerically simulated soliton comb spectra without (orange) or with (blue) the influence of Raman effects for the D_int curve at a resonator width of 2.3 μm (height of 1.0 μm). c. Experimentally recorded chaotic comb spectra at a varied resonator width of 2.3, 2.4 and 2.5 μm. The vertical arrows indicate the emergence of dispersive wave-like envelopes, and the corresponding D_int curves are shown in the top panel of b. | 2012.13496 | Figure_S2.jpg |
0.506996 | eceb87cfc6624deb9f8a6acda04ea987 | Dispersion engineering for repetition rate-detectable octave solitons. a. Top: D_int curves of engineered AlN resonators at a radius of 100 μm and varied widths of 3.3 and 3.5 μm. Bottom: captured noise-state comb spectra with a reduced repetition rate of ∼220 GHz. The vertical arrows indicate the emergence of DW-like peaks. b. Top: D_int curve of AlN resonators at an increasing radius of 200 μm and an optimal width of 5.0 μm. Bottom: simulated soliton spectrum at an on-chip pump power of 100 mW, highlighting a reduced repetition rate of 109 GHz. | 2012.13496 | Figure_S3.jpg |
0.395073 | 8c2ff0a26e144b3f8d21a830d6426a29 | Octave-soliton characterization. a. Sketch of the experimental configuration. The SC-SSBM produces a blue-shifted sideband (λ_p) relative to incident light (λ_c) from a continuous-wave (CW) laser. The sideband is then boosted by an erbium-doped fiber amplifier (EDFA) and aligned to the vertical polarization before entering the AlN chip via an aspherical lens pair. b. Comb power trace recorded in the OSC, indicating elongated soliton duration time beyond 2 s. Inset: initial soliton lifetime of ∼20 ns. c. Free-running soliton power trace as a function of the time. d. Radio-frequency (RF) beating properties of chaotic (red) and soliton microcombs (blue) comparing with the PD background (black). | 2012.13496 | Figure_S4.jpg |
0.429189 | a9e646bc90d6436186ec36d0b7e1ac1c | SHG conversion efficiency. a. On-chip SHG power versus the pump power as well as the applied second-order polynomial fit (orange solid line). Inset: calculated effective indices (n_eff) of TM_00 (wavelength of 1970 nm) and second-order TM_20 (wavelength of 985 nm) modes versus the waveguide width. b. Calculated SHG efficiency (indicated by right color bar) as a function of the TM_20 mode propagation loss and the waveguide length. The "red star" symbol corresponds to our experimentally estimated SHG efficiency. | 2012.13496 | Figure_S5.jpg |
0.394163 | 5c9a048a25e940558e2b88f6dd784865 | Implementation of f–2f interferometry. a Octave soliton spectrum after suppressing the pump by a FBG (see full spectrum in Fig. 4b of the main text). b. Sketch of the involved frequencies (monitored by the OSAs) around the f and 2f bands for the f_ceo measurement. c. Illustration of the experimental setup for electronically accessing the f_ceo based on a down-conversion frequency process. PD1: 830–2150 nm, 0–12.5 GHz; PD2: 800–1700 nm, 0–26 GHz; LNA1: 6–18 GHz, gain ≈ 27 dB; LNA2: 10–800 MHz, gain ≈ 60 dB; Mixer1: L/R port (7.5–20 GHz), I port (0–7.5 GHz); BPF1: 20–1000 MHz; RF doubler: 10–1000 MHz; BPF2: 1.5–2 GHz; Mixer2: L/R port (1–2700 MHz), I port (1–2000 MHz). | 2012.13496 | Figure_S6.jpg |
0.417074 | c599f968871c497484a4d77d2f5b09f6 | The generated training data. The first column is from source domain. The second column is from SPGAN-generated data and the other columns are from CamStyle data. | 2012.13498 | camstyle.jpg |
0.431991 | 394d2d62c7f7489fac3dc90b59c436dc | The pipeline for domain adaptation stage | 2012.13498 | pipeline_new_new.jpg |
0.392747 | 0b2734edf649476280d88239313bee27 | A schematic representation of the two models analyzed: (left) model 1, a solution of polyions composed of different numbers of charged monomers and (right) model 2, a solution of polyions and simple monovalent salt in an aqueous solvent in both cases. The large particles are charged polymers of N_+ and N_- monomers treated as Flory-Huggins particles. | 2012.13503 | Figure1.jpg |
0.449124 | d1ca5265217f47738c4bb0e64b5a6903 | Dependence of the ratio κ̃^2(N_3, N_4, ϕ_s, ϕ_M) on ϕ_s, ϕ_M, Eq. <ref>, for N_3 = 10 and N_4 = 50 as indicated in the figure . The symmetric case of N_3 = N_4 reduces back to the standard Debye length, coinciding with κ̃^2(N_3, N_4, ϕ_s, ϕ_M) = 1 on the figure. The dependence of the screening length on the two volume fractions deviates significantly from the Debye screening length for non-vanishing volume fractions ϕ_s, ϕ_M, and shows opposite trends as a function of ϕ_s and ϕ_M, i.e., one is increasing and the other one is decreasing. | 2012.13503 | Figure4.jpg |
0.502853 | 27e3a90ce85b427f9c8766b3bbdd0dc2 | Gender gap in investment across treatments | 2012.13514 | Figure2aBarPlot.jpg |
0.451745 | 2ea746d0153140bb81ab4f83c73368ec | Gender gap in investment across treatment over round | 2012.13514 | Figure4.jpg |
0.424346 | 126c6bcb3ba44fcd90fabd89c31b0523 | Distribution of Total Investment across Treatments | 2012.13514 | FigureA3.jpg |
0.5213 | e2a62d9cbde64029871cfbc9ce458895 | At each x ∈E the extremal outward directions satisfy ξ_1 = - ξ_2, while the set Ξ_x(E_) spans a half-circle. | 2012.13515 | Basic_Cases_01_v03.jpg |
0.508398 | 3901d08a7dc847ee80493fe355651e4e | For a wedge x there are two extremal contributors y_1, y_2 and two extremal outward directions ξ_1, ξ_2, forming an angle θ = ∢(ξ_1, ξ_2). | 2012.13515 | Basic_Cases_02_v03.jpg |
0.492145 | c34f3489b41b4f39a9c97c9244a16fb1 | For x = {y_1, y_2} with y_1 - x = - (y_2 - x), the set of extremal outward directions satisfies x⊂{ξ_1, ξ_2}. | 2012.13515 | Basic_Cases_03_v03.jpg |
0.408489 | 03efd1c0b8fc44159f330bc35577002f | A singularity x ∈∂ E_ with x = {y_1, y_2} and x = {ξ_1, ξ_2}. | 2012.13515 | Basic_Concepts_03-01.jpg |
0.449084 | 03de5676a3514d7f9bdeaa04b947cc2c | The set Ξ_x(E_) ⊂ S^1 of outward directions is geodesically convex with boundary ∂_S^1Ξ_x(E_) = {ξ_1, ξ_2}. | 2012.13515 | Basic_Concepts_03-02.jpg |
0.41471 | de6882a924144c4493b9effaae129308 | The boundaries ∂ B_(d_k, ℓ^n) for d_k,ℓ^n ∈ D^n (red dots) provide an approximation (red curve) for ∂ E_. | 2012.13515 | Finite_Approximating_Sets_01_01.jpg |
0.410705 | afc6cc9dcbfd4960bcccfd242f186c66 | For D^n+1 the approximation improves, and as n →∞, it converges uniformly to ∂ E_. | 2012.13515 | Finite_Approximating_Sets_02_01.jpg |
0.441294 | 80cdb6dbfe9f436190a61ae2c772a7ae | The boundaries ∂ B_(y_1) and ∂ B_(y_2) give an approximation for the local geometry of the boundary ∂ E_ inside the ball B_r(x). | 2012.13515 | LBR_Motivation_01_01.jpg |
0.483364 | 1aab9a0a7da54c1e97cc870cb87cbc44 | For an exact representation, one needs to consider all the contributors within some radius δ > 0 from the extremal contributors y_1 and y_2. | 2012.13515 | LBR_Motivation_02_02.jpg |
0.468137 | 79936f05ea8742fd8b5318f9bd130fe1 | Schematic illustration of Proposition <ref>. For a sufficiently small r>0, the balls B_δ(y_1) and B_δ(y_2) contain all the contributors y ∈Π_E(z) (red dotted lines) of those boundary points z ∈∂ E_ that lie inside the ball B_r(x) (red solid lines). For ρ > r the boundary segment generated by the point y^* ∈ E would lie inside the larger ball B_ρ(x) and would need to be accounted for separately. Note that the geometry of the boundary ∂ E_ inside B_r(x) is not affected by whether or not the points y on the blue dotted line satisfy y ∈ E. | 2012.13515 | Local_Boundary_Newnew_03.jpg |
0.426093 | 4673707bcf9f444aa0cf0df7b710e34e | n = 1 | 2012.13515 | Local_Boundary_Repr_1-1.jpg |
0.452183 | 5f29c91cbc804ad18517616319371593 | n = 2 | 2012.13515 | Local_Boundary_Repr_2-2.jpg |
0.408092 | 351a4880243e4a8dabb815b887df537f | The geometric picture. Here r_n = - cos( arcsin (h_n/)) = - √(^2 - h_n^2), and one can solve for these values so that H_n = φ_n is satisfied. | 2012.13515 | Outward_is_tangent_1_v4.jpg |
0.478446 | d5e5ae735bde4ca289aa10742325b022 | For each n ∈, the point z_n ∈∂ E_ lies on a geodesic arc-segment on ∂ B_φ_n(x), which connects the points x_n and x + H_n. | 2012.13515 | Outward_is_tangent_2_v4.jpg |
0.465558 | dc56c00daa054a2e8f43946501972bec | Schematic illustration of Theorem <ref>. The grey area represents the -neighbourhood E_, the white area the complement ^2 ∖ E_. Every boundary point x ∈∂ E_ either is a smooth point or belongs to exactly one of eight categories of singularities. At a wedge (S1) the one-sided tangents form an angle 0 < θ < π. A sharp singularity (S2) and a sharp-sharp singularity (S3) can be thought of as extremal cases of a wedge, with θ = 0. A shallow singularity (S4) and a shallow-shallow singularity (S5) have a well-defined tangent, but they are accumulation points (from one or two directions, respectively) of sequences of increasingly obtuse wedges (black dots). A chain singularity (S6), a chain-chain singularity (S7) and a sharp-chain singularity (S8) share the geometric property of being accumulation points of sequences of increasingly acute wedges (black dots). This turns out to be equivalent (see Proposition <ref>) to the topological property of being the limit with respect to Hausdorff distance (see Definition <ref>) of a sequence of disjoint connected components of the complement E_^c. See also Figure <ref>. | 2012.13515 | Types_of_Boundary_Points_v03.jpg |
0.427331 | 00ba4dda75f64aadbc2e98fadfbac8fc | Purcell factor. (a) Then integrand of Eq. (A1) as a function of in-plane wavevector for the case h=10 nm. (b) The dependence of χ with distance h. (c) The total Purcell factor, together with P_1 and P_2, as function of the distance between the line of dipoles and one of the graphene layers. | 2012.13519 | S1.jpg |
0.431269 | 7b5e5fdef70c466496ad048325fc2ce3 | Symmetry of graphene plasmons. (a) Schematic illustration of a graphene double-layer system. A horizontally polarized Hertzian dipole is placed at a distance of h from the upper graphene layer located at z=0 (b) The plasmon dispersion of the double-layer system. 'SPE' stands for the single particle excitation. (c) The Fourier spectra of graphene plasmons excited with a single emitter placed 10 nm away from the upper graphene layer. | 2012.13519 | fig1.jpg |
0.491464 | 16aa04ac67f44f008ed275850eacbbcd | Plasmon excitation with a single emitter. The symmetry of the graphene plasmons as a function of the distance of (a) a horizontally and (b) vertically oriented dipole from the top graphene layer. Spatial distributions of electric fields in the x direction when h=0 excited with a (c) horizontally and (d) vertically oriented dipole. | 2012.13519 | fig2.jpg |
0.531247 | 0a95735ab0814848b08c171ffcc0aad7 | Plasmon excitation with multiple emitters (a) The symmetry of the graphene plasmons as a function of the spacing between two dipoles. (b) The symmetry as a function of the number of dipoles separated by 120 nm. Spatial distributions of electric fields in the x direction when (c) N=2 and (d) N=21. | 2012.13519 | fig3.jpg |
0.456948 | f571fd7c2d8b438195c19a2abd7f1367 | Far-field excitation of acoustic plasmons (a) (Upper) Resonator design based on the double graphene ribbon arrays. (Lower) New resonator design based on a continuous graphene sheet with a graphene ribbon array. (b) Numerical results for far-field spectra obtained from the two designs. The red curve shows the case of the new design based on a continuous graphene layer. The spatial distributions of electric fields in the x direction for (c) double ribbon arrays and (d) asymmetric double-layer design. | 2012.13519 | fig4.jpg |
0.427165 | a6d21dc304924e2da3a4419f32a36f41 | Ultimate monoatomic plasmon confinement limit. (a) Schematic illustration showing two graphene layers separated by a monoatomic spacer. (b) Schematic illustration of the graphene-on-a-mirror system with a monoatomic spacer in between. This configuration is effectively equivalent to that of double-layer graphene separated by a double-atomic-layer spacer. (c) The lateral plasmon confinement as a function of g at λ= 8 μm. | 2012.13519 | fig5.jpg |
0.50291 | 4aa6307975444566aa6cb6244acac5e1 | Schematic of the silicene constriction. The zigzag direction is parallel to the x axis as it's shown in the picture. The orange region shows a central symmetrical scattering region for the silicene constriction. The blue part gives a different geometry structure with Δ y referring to the relative distance of the top edge to that of the orange geometry. | 2012.13520 | figure1.jpg |
0.426646 | 54aace22d28049c6a6d237b43b808a2c | Schematic diagram of the silicene constriction with two narrow ribbons in central scattering region. The zigzag direction is parallel to x axis as shown in the picture. The variety of blue and orange region shows two different distances of the two narrow ribbons. d=2Mξ is defined as the distance between two ribbons. | 2012.13520 | figure7.jpg |
0.426725 | 0b44867f54ef494b9f32c7cfe6f134e3 | An IRS-assisted uplink OFDM system. | 2012.13521 | CEE.jpg |
0.472456 | 38b9f5f1898941458a90c61b7daeba61 | NMSE versus transmit power for low-resolution IRS cases and high-resolution IRS cases. | 2012.13521 | NMSE5_8.jpg |
0.436399 | ec12618276ab4084a8a90d2b852554e0 | NMSE versus transmit power under existing reflection pattern design schemes (b=5). | 2012.13521 | frequency4_8.jpg |
0.418427 | a16926305e364d68821c3e75fe1ef0ec | Top: The value of (14a) versus the number of iterations based on Algorithm 1, b=1. Bottom: The value of (14a) versus the resolution b based on Algorithm 2. | 2012.13521 | iteration6.jpg |
0.530615 | 3876edfb27fb4ecfad3a32c856ae70e4 | (a,b) Evolution of the band gap with the angles θ_s and ϕ_s in the Heisenberg-DMI model (K=0) (a), and Heisenberg-Kitaev model (D=0) (b). (a, b) share the same color map and the units are meV. (c) Thermal Hall conductivity κ^xy_TH as a function of θ_s and ϕ_s at T=100K (same parameters as in (b)). (d) κ^xy_TH as a function of temperature and ϕ_s at θ_s=60^∘ (e-f) κ^xy_TH as a function of temperature and θ_s for different values of D assuming ϕ_s=0^∘ and K=5.2 meV. (c-f) share the same color map in units of W/T. | 2012.13729 | B_effect_update.jpg |
0.465776 | 9e9f993442ea4480b44cd181a4855c04 | (a) The comparison of magnon dispersions with different Kitaev and DMI parameters specified in the legend (in the units of meV and degrees). The magnon dispersions represented with black, blue, green and red lines have almost the same band gap between the two branches. (b) Band gap as a function of D and K at θ=54.74^∘. (c) Dependence of thermal Hall conductivity κ^xy_TH on K and D at T=100 K and θ=54.74^∘. The T-dependence of thermal Hall conductivity κ^xy_TH is shown in (d-g). (d) κ^xy_TH as a function of T and D, with K and θ corresponding to the case of CrI_3. (e-f) The corresponding BZ distribution of the Berry curvature (in arb. units) for the first band with D=+0.4 meV (e) and -0.4 meV (f) and the Kitaev interaction is the same of (d). (g) κ^xy_TH as a function of T and K for θ=54.74^∘ and D=-0.2 meV. (h) κ^xy_TH as a function of θ and D at T=100 K and K=5.2 meV. The corresponding color map is in the units of W/K. | 2012.13729 | D_effect_update.jpg |
0.499777 | 8aec42e0414749e89e1e0f2ba3191e51 | (a) Sketch of the structure of honeycomb CrI_3 monolayer. The unit cell is outlined with a thin black line, where blue balls represent Cr^+3 ions and pink balls are iodide ions. The Kitaev bonds x (red), y (dark blue), z (yellow) are indicated with thick colored lines. The arrows mark the second-nearest-neighbor bond orientations along black dotted lines that share a common sign of out-of-plane DM vector. (b) Schematic diagram of the influence of an in-plane magnetic field 𝐁 on the spin direction 𝐒 whose polar angle is defined as θ_s. (c) The perspective view of Cr_2I_2 plane, where its normal vector is marked as γ̂ and the Kitaev angle is defined as the polar angle of γ̂. | 2012.13729 | structure_update.jpg |
0.504066 | 39f0f059121947d1b07a2220510fe4c1 | (a) The comparison of magnon dispersions along high-symmetry lines for different values of the Kitaev parameters. Grey, blue and red lines correspond to the K values of 0, 5.2 and 3 meV. The dashed and solid lines correspond to the Kitaev angles θ of 45^∘ and 54.74^∘, respectively. The corresponding energy-resolved Chern number is shown in (b). The Berry curvature distribution of the first magnon branch in the first Brillouin zone for different K-values with θ=54.74^∘ is shown in the inset of (b). The color map ranges from -40 to 40 in arb. units, and the exceeding values are marked with black. The temperature dependence of thermal Hall conductivity κ^xy_TH and magnon Nernst conductivity κ^xy_N is shown in (c) and (d), in units of 10^-11 W/K and k_B/2π, respecitively. (e) The map of κ^xy_TH/T as a function of temperature T and parameter K at the Kitaev angle of θ=54.74^∘. (f) The map of κ^xy_TH/T as a function of T and θ at the constant K-value of 5.2 meV. The units of the color maps in (e)-(f) are chosen as W/K^2. The magnon band gap as a function of K (at θ=54.74^∘) and θ (at K=5.2 meV) is shown in (g) and (h), respectively. | 2012.13729 | without_DB.jpg |
0.39749 | 45732be71c7c4d08a0820b3cf5f22e30 | (a) The sketch of SL-TMD nano-bubble. The strain-induced bandgap modulation gives rise to radial funneling force, which is partially compensated by the counteracting diffusive propagation. The piezoelectricity-induced dipolar interaction and the emergent pseudomagnetic field generate highly anisotropic forces, leading to spatially inhomogenous drift of excitons. (b) The electric field streamlines on top of bubble profile colormap. The horizontal (vertical) colorbar corresponds to piezo-induced electric field, E_ piezo/E_0 with E_0 = 10^7 V/m (height h/h_ max). | 2012.13730 | Fig1.jpg |
0.480072 | 8d5c866d28e54b3a97a749c4ed192228 | (a-c) The spatial colormap of strain-induced potential energies and the corresponding streamlines of drifting forces: (a) scalar funneling potential; (b) pseudo - magnetic field; (c) dipole-dipole interactions. The vertical (horizontal) colorbar corresponds to potential energy in meV (drifting force in F_0 = 100 meV/nm). The black circle denotes the bubble boundary with radius R =100 nm. (d) Potential energy peak value plotted versus the bubble radius. For macroscopic bubbles (R > 100 nm) the exciton dynamics is solely governed by the radial funneling term while for nano-bubbles (R < 100 nm) the dipolar and magnetic sources rapidly increase by reducing the bubble radius. | 2012.13730 | Fig2.jpg |
0.414369 | 6a8fa452df284bceba730beef728b6ba | (a), (b) and (c) snapshots of exciton density (in a.u.) for T=10 K and R=100 nm bubble radius at t= 0, 2, 10 ps, respectively. The thin lines indicate cross-sections along x and y directions. (d) The difference (dashed curves amplified by a factor of 10) and the average (solid curve) of exciton density along x and y cross-sections at different time steps, corresponding to panels (b) and (c). (e), (f) and (g) snapshots of exciton density (in a.u.) for T=10 K and R=75 nm bubble radius at t= 0, 2, 10 ps, respectively. In smaller bubble the asymmetric hexagon-shaped is pronounced and persists for longer time. (h) The difference (dashed curves amplified by a factor of 10) and the average (solid curve) of exciton density in x and y cross-sections at different time steps, corresponding to panels (f) and (g). (i) exciton density at the bubble center in the absence of XX-interactions at several temperatures. The solid lines correspond to dissipative dynamics, and the dashed lines correspond to the absence of dissipation, i.e. τ_d→∞. Vertical lines indicate the respective lifetime value τ_d=(0.1+0.9T[ K])ps. (j), (k) The snapshots of exciton density at t= 10, 20 ps with elevated temperature T=40 K. (l) The difference (dashed curves amplified by a factor of 10) and the average (solid curve) of exciton density in x and y cross-sections at different time steps, corresponding to panels (j) and (k). | 2012.13730 | Fig3.jpg |
0.44429 | 62fa6f67de2846c58444285ded7b3735 | Cat to other species translation results by <cit.>. | 2012.13736 | animalchange.jpg |
0.493232 | f60285bd865f44ccbf0cf21709488e82 | Dynamic results of Gaussian kernel for regular GANs and LSGANs on cmr CIFAR-10 dataset. Red portions represent real samples and Blue portions signify generated ones. GAN almost in all the steps generated the synthetic samples that are very close to their corresponding original (target) samples. | 2012.13736 | fig11-1.jpg |
0.403498 | 381ac42218d240db96191397e58d5ddf | Performance of single-path and dual-path network. Left: standard single path model to learn the latent representation. Right: Two-pathway network combined with self-supervised learning, which can learn complete representations. | 2012.13736 | fig13.jpg |
0.419706 | 4f405293155c44e89e2b129230e0e3fb | Face style transferring samples with CelebA dataset by <cit.> . | 2012.13736 | fig14.jpg |
0.429014 | d719fd633b564d289616aa5f1f7b3bb0 | Generated samples of flowers by StackGAN <cit.> . | 2012.13736 | fig15.jpg |
0.445159 | a0607c771a1d4e82a8e166c609802ccc | Sample generated results by BEGAN, CGAN, LSGAN, StarGAN, and DA-GAN on four image datasets. | 2012.13736 | fig16.jpg |
0.478928 | a305358d73c94ce8ac7987dd8d24e767 | 3D completion results on real-world scans. 3D-ED-GAN shows the low-resolution completion result when going through LRCN. | 2012.13736 | fig17.jpg |
0.446619 | 6cd33b0370a64a658e04e8fcbb0889dd | Reconstruction error of the various 3D GANs models on 3D CAD and MoCap datasets at various training epochs. Reconstruction measures the closeness of the synthetic results to the training test ground truth distributions. | 2012.13736 | fig19.jpg |
0.472751 | e57ae9e7b996453c9d1a55bf562a430e | An overview of image synthesis methods since 2014, we tried to reference the original research articles, later discuss their proposed variations in the content. | 2012.13736 | fig1a-1.jpg |
0.469969 | c1b678c3f4b04467b05f6804d7a0a7cf | It shows the most frequent keywords in CVPR and ICCV conferences from 2017 to 2020. The size of each word indicates the frequency of that keywords. The, "synthetic" and "adversarial" keywords are often have been used. | 2012.13736 | fig1b.jpg |
0.525414 | d548b7e42c6d4fd0b9177b970a47ae6a | GAN architectures and its variations.(a) Shows the basic architecture of a GAN <cit.>; (b) Shows the conditional GAN, introduced in <cit.> that perform class-conditional image synthesis; (c) Shows the architecture of InfoGAN <cit.> and ACGAN <cit.>; and (d) Shows the BAGAN architecture <cit.>. | 2012.13736 | fig2-1.jpg |
0.387654 | 17fdf23af0d04f52a2fafa89e9e92722 | Comparison of generated results and ground truth for face aging on CelebA datasets by using <cit.>. | 2012.13736 | fig21.jpg |
0.415063 | b5277231e6cb4fe593005e3e63e63d60 | Street scene image translation sample results by <cit.>. For each pair, left is input and right is the translated image. | 2012.13736 | fig22.jpg |
0.412024 | 5c36c78c9d53437b88b0005d8e4f899b | Samples of face to cartoon conversion using XGAN <cit.> on CelebA dataset. | 2012.13736 | fig23.jpg |
0.46806 | b12b3dbf8def4ff7bfe1d9ff073b0579 | Throughout the training, the G is tries to generate a distribution of samples which have similar distribution to real data. Generally, the GANs are represented by the learned parameters (weights) that have been captured from G and D. | 2012.13736 | fig3.jpg |
0.410428 | befed5bde0c24e6588d67dcdd9a4bebd | Architecture of (a) auto-encoders GAN; (b) BiGAN (c) Adversarial Generator-Encoder Network (AGE) <cit.>; (d) BEGAN architecture <cit.>. z is the random noise for G and x is input image. BEGAN deploys an auto-encoder architecture for the discriminator. | 2012.13736 | fig4-1.jpg |
0.421831 | e9670e0bacec4293b40911c17a0f3272 | Architecture of: (a) GAN with auxiliary classifier <cit.>, in which 'y' is the conditional input label and C is the classifier; (b) CycleGAN <cit.>. | 2012.13736 | fig6.jpg |
0.435686 | 9dfe61c5bf09432989a067007ba8c070 | Architecture of; (a) class-condition GAN <cit.>; (b) optimized GAN <cit.>. The model contains two different optimization networks. | 2012.13736 | fig7.jpg |
0.451794 | 6cf0f9deff094fc49dd4e9916180c8e2 | Sample generated results using CycleGAN <cit.> | 2012.13736 | fig8.jpg |
0.47666 | 4fbb86a0bb6f459a9af4bc2d3f0a5058 | Sample training results of different iterations on FashionMNIST dataset using DCGAN <cit.> . | 2012.13736 | fig8b.jpg |
0.459337 | 74d167bc033347938dd832d43374047d | Performance of CycleGAN <cit.>and DCGAN <cit.> based on their training loss. The models are evaluated on flickr dataset . (a) Training loss of CycleGAN on different setting; (b) DCGAN training loss of G and D. | 2012.13736 | fig9.jpg |
0.405481 | 61b04e5805884fd48856a3e56ce3ca1d | Sample generated results by <cit.>. | 2012.13736 | g-43.jpg |
0.399538 | ea3a83f322eb4088bddcc0e77fa66b9a | Male to female image translation sample results by <cit.>. For each row, left is input and right is the translated samples. | 2012.13736 | genderchange.jpg |
0.391046 | 11767bb8d3db437aad53107306b14a98 | Sample generated results by <cit.>. | 2012.13736 | image-completion.jpg |
0.460593 | 877a5ef035ff4322abf7effea72432f7 | (Color online) Qualitative view of the dependence of the subband bottom energy (E1 and E2 are the energies of the electron subband bottom, H1, H2, and H3 are the energies of the hole subband bottom) of dimensional quantization of the HgTe QW on its thickness dw. | 2012.13738 | f1.jpg |
0.564035 | b473dffbd1134274a9b6b333cecf574f | (Color online)(a) Magnetoresistance of a 2D TI based on an 8 nm HgTe QW in local and nonlocal geometry. (b) Normalized magnetoconductivity as a function of the magnetic field in the range |B|<1.2 T. Dashed curves show linear approximations of the dependences. | 2012.13738 | f10.jpg |
0.553445 | 4303c4f2a1614332bfca39ae4aa2b6f9 | (Color online) Local and nonlocal resistance of a 2D TI depending on the magnetic field applied in the QW plane: (a) local magnetoresistance, (b) nonlocal magnetoresistance, (c) theoretical calculation of the fraction of the dielectric state in the 2D plane with and without taking impurity disorder into account. | 2012.13738 | f11.jpg |
0.427957 | e0d2d6d9060e476bac6ba8eb8c890283 | (Color online) (a) Photo of the microstructure of the special Hall geometry. (b) Local resistance at B = 0 and the Hall resistance at B = 1 T as a function of the effective gate voltage . (c) Dependence of the nonlocal resistance at B = 0. | 2012.13738 | f12.jpg |
0.401639 | e3d5754dcccc41d48cf0f256ed0e6292 | (Color online)(a) Local photoresistance and the resistance as functions of the effective gate voltage. (b) Nonlocal photoresistance and the resistance as functions of the effective gate voltage. Red curves are shown to guide the eye. | 2012.13738 | f13.jpg |
0.485585 | d11380ca7ef34b46a185811d035121f8 | (Color online)Qualitative view of the dispersion law of the bulk and edge states of 2D TI in a 14 nm HgTe QW: (a) spectrum without taking the mixing of bulk and edge states into account (S1 and S2 are the dispersion laws of electron subbands); (b) resulting spectrum for the two upper hole branches and edge states. | 2012.13738 | f14.jpg |
0.42487 | 379049e693e345c3bcbf74cf06156165 | (Color online)Temperature dependence of the edge transport of a macroscopic sample based on a 14 nm HgTe QW. (a) Local and (b) nonlocal resistance as a function of the gate voltage at various temperatures. (c) Temperature dependences of the local and nonlocal resistance at the CNP. | 2012.13738 | f15.jpg |
0.478964 | ffd97cc0442a41f9bf82968b9b26430a | (Color online)The measured local and nonlocal resistances of micrometer-size samples based on a 14 nm HgTe QW. | 2012.13738 | f16.jpg |
0.417835 | ab5046356cd34b16827f385ed76caa47 | (Color online)(a) Cross section of the studied structures. (b) Energy spectrum of a strained 80 nm mercury telluride film. (c) Dependences ρ_xx and ρ_xy. (d) Electron and hole densities as functions of the gate voltage. (e) Average electron and hole mobility. | 2012.13738 | f17.jpg |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.