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
|
|---|---|---|---|---|
0.400223 | 3d53fbdec72a4ccf9866d73eb32b8430 | Observational constraints on the shape parameters of stellar halo. To fully take the possible impact of the threshold in definition of stellar into account, we add the theoretical constraints from ε≤ (0.6, 0.7, 0.8, 1.0). | 2012.12284 | sq-star.jpg |
0.517929 | dc3bd999bd4145b4b3cb0d6b6e261d2b | Two-mass-hard hexagon configuration. The dual coordinates x_i denote region momenta. | 2012.12285 | hexagon.jpg |
0.527841 | a984d28c3e914180b757d0c15c137002 | Cooper-channel fluctuation response diagrams in the normal state. The top left diagram is the AL diagram, next is the MT diagram, followed by the two DOS diagrams, and finally the diamagnetic diagram. | 2012.12287 | EtaFluctuationDiagrams.jpg |
0.520803 | 3025fd0c324d447ea6823f7cd891023d | Density-channel fluctuation response diagrams in the normal state. | 2012.12287 | ThetaFluctuationDiagrams.jpg |
0.551134 | 7fdbc0cd7eea40dea58970899ecf25aa | A diagrammatic representation of the dressed propagators appearing in the Eliashberg equations (<ref>) and (<ref>) | 2012.12287 | dressed-propagators.jpg |
0.587135 | 4af21f1879214ccaa1123cad750159a5 | A diagrammatic representation of the Dyson equation for the normal-state Eliashberg fluctuation propagators. | 2012.12287 | fluctuation-propagators.jpg |
0.534123 | 2eef55d3fac6483fa90196d736f8ba0f | A diagrammatic representation of the mean-field Eliashberg equations (<ref>) and (<ref>). The dotted lines terminated by circles denote λ, i.e., the phonon propagator and the coupling constants. The double-struck solid lines denote G_n and the bold solid lines are the full fermion propagator G. The lines at the ends of the diagram are external legs. | 2012.12287 | saddle-point-eqs.jpg |
0.408335 | c49caa5a725d4db68ff3b39d1ad960b0 | Map of sources with K_ s=16–18 mag from the VMC survey area around the SMC (the prominent dark area in the middle; the contour demarcates the boundary between `inner' and `outer' regions). The VMC tiles are labelled. The round overdensity seen at RA =5, Dec =-72 in tile SMC 5_2 is caused by the Galactic globular cluster 47 Tucanæ (another Galactic globular cluster, NGC 362, is visible in tile SMC 6_4 at RA =16, Deg =-71). The gap between tiles SMC 5_3 and SMC 5_4 (seen at RA =13.5) is caused by the absence of data. | 2012.12288 | fig1.jpg |
0.39273 | 91db28abde884f3c99b5c400551914eb | Y-K_ s colour vs. regional census sub-region density of RC stars, showing a positive correlation (albeit with considerable scatter). The colour scale quantifies the number of sources per sub-region. | 2012.12288 | fig10.jpg |
0.498257 | faebf5ecb4364c82a52d5e99d44a0037 | Histogram (bin size of 0.04 mag) comparing the extinction distribution of RC stars in the near-Wing component (turquoise, solid) to that in the far-Wing component (red, dashed). | 2012.12288 | fig11.jpg |
0.503441 | b27c712ee8924bb4ad217e61d9acc5a9 | K_ s versus Y-K_ s diagram around the RC in tile SMC 5_3, showing the paralellogram selection box inclined along the reddening vector and limited at (Y-K_ s)=1.1 mag to avoid Galactic foreground stars. The density contours are at 1% of peak (red), 5%–30% of peak with step 5% (yellow) and 50%–100% of peak with step 10% (green). | 2012.12288 | fig2.jpg |
0.489962 | 69137c2782b040599b67a51a16ae0f9e | CMDs for all tiles, each with their RC selection box and contour at 30% of the RC peak within the CMD. | 2012.12288 | fig3.jpg |
0.483474 | d5f7ef2484a840c299ea1ef8987fd5fb | Histograms of de-reddened RC star K_ s magnitudes (bin size of 0.02 mag for K_ s=16.5–18 mag range) for each SMC tile (labelled in the top–left or bottom–right of each panel). Of particular interest is the double RC seen in the eastern fields. | 2012.12288 | fig7.jpg |
0.51122 | cefeec0cc0f94b97932fea3b8153bafb | Histograms of residuals from the plane fits (bin size of 0.2 kpc) for `outer' (red), `all' (black) and `census' (turquoise) SMC selections. The first (leftmost) and fourth (rightmost) quartiles cover a wide range of distances. Location of the boundaries between the quartiles are indicated with vertical lines and labelled as `1', `M' and `3', respectively. Cumulative distributions are overlain (monotonically increasing curves, with values on the right hand side of the y-axis). | 2012.12288 | fig8.jpg |
0.473036 | 5c5ce45348da443f834d9d9134fbd724 | The figure on the left shows a representative example of the robot navigating through the crowd of pedestrians over time using both the baseline and the group aware policy. The baseline policy chooses actions that cut through the group of pedestrians and influences the group formation, while the group aware policy chooses actions that move around the group with minimal disturbance. The figure on the right shows the average distance between the pedestrian and the robot (top) and the average pedestrian velocity (bottom) over time. Here, we show the group aware policy results in increased distance to the pedestrians while allowing the pedestrians to maintain faster speeds. | 2012.12291 | case-study-draft.jpg |
0.429144 | 6bf745f04ebe4f1a8f4f8d31f44915af | During training, we compute the convex hull of pedestrians that are members of a group. We encourage the robot to maintain separation between the robot and the group, which, in turn, leads to a more socially appropriate group navigation policy. | 2012.12291 | convex_hull.jpg |
0.42458 | 28b9a4f8d05142ad868468d9651e03c7 | On the left is the CrowdNav Simulation Environment <cit.> that provides pedestrian and robot state information as well as the reward to the policy. The right figure represents the neural network architecture used for our attention-based, actor-critic policy. The pedestrian and robot state vectors are concatenated to represent a pairwise combined state vector; the output of the network are the policy π over potential actions and value V of the current state. The gray and green blocks indicate features from individual pedestrians. The blue blocks indicate aggregate features across pedestrians. The argmax of the policy is chosen as the action, which subsequently is sent to the CrowdNav Simulation Environment to control the robot. | 2012.12291 | network_arch.jpg |
0.467652 | 48e5aeb87c2a42f6a4a942a70c312be2 | The objective of this work is to learn a navigation policy that allows the robot to safely reach its goal while minimizing impact to individual and groups of pedestrians. | 2012.12291 | teaser.jpg |
0.379556 | e74f39e6d5d9463382c813be7940b633 | BSC and SSSC model parameters learned from artificial data using EEM (for more details see the description of the experiment in Section <ref>). | 2012.12294 | bars_appendix.jpg |
0.477349 | 2f0c3dcc0cd94017aaae8c8a2aa97ed5 | Illustration of the variational optimization using genetic operators. | 2012.12294 | eem_illustration.jpg |
0.372931 | 3d6212a3ff204a769171e15f2d5a3327 | Dictionary elements learned by noisy-OR (top), BSC (middle) and SSSC (bottom) models on natural image patches. The models were trained on separate datasets that had been obtained using different preprocessing schemes (see text). For noisy-OR, a random selection of 50 out of 100 fields is displayed; for BSC the fields corresponding to the 60 out of 300 most active hidden units are shown; for SSSC a random selection of 60 out of 512 fields is displayed. The full dictionaries learned by the individual models are depicted in Figure <ref> in Appendix <ref>. | 2012.12294 | norbscssscHateren.jpg |
0.415077 | 1ef172067a8a487b8736da03a23008c2 | Dictionaries learned from natural image patches using noisy-OR, BSC and SSSC models (see Section <ref>). For noisy-OR, we considered raw image patches and trained a model with H=100 components. For BSC and SSSC, image patches were preprocessed using a whitening procedure. For BSC and SSSC, H=300 and H=512 generative fields were learned, respectively. In B and C, the fields are ordered according to their activation, starting with the fields corresponding to the most active hidden units. | 2012.12294 | norbscssscHateren_full.jpg |
0.428029 | ae1d971d98c94d45a47eebecaa5c2041 | Variational bound and PSNR value pairs obtained from applying ES3C to the noisy “House” image (AWG noise with σ=50). 2,000 EEM iterations were performed and PSNR values were measured at iterations 1, 2, 5, 10, 20, 50, 100, 200, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, 2000. The experiment was repeated five times using the same noise realization in each run. The PSNR values at the last iteration were 28.94, 28.95, 28.94, 28.96, 28.95 (avg. rounded 28.95±0.01; all values in dB). | 2012.12294 | psnr_fenergy_ssscHouse50.jpg |
0.494284 | 73043470e4124e89b94db020590ebc69 | SSSC generative fields learned from the noisy “Barbara” image (compare Section <ref>). The fields are ordered according to their activation, starting with the fields corresponding to the most active hidden units. | 2012.12294 | ssscBarbaraParameters.jpg |
0.411243 | 4acce91ae57b487a8ae2a9fa286058da | Megno map for circumbinary system, as obtained from the numerical integration of a grid of 250 × 100 initial conditions in the (a,e) plane. All angles are initially set to zero. Integration time was set to 10^4 binary periods, and escapes during this timespan are highlighted in brown. The color code shows values of Y^*=log_10(⟨ Y ⟩-2). Dark-blue colors correspond to more regular orbits while lighter tones of green indicate increasingly chaotic motion. The libration widths of the most relevant MMRs determined with the model are plotted in white lines. Top scale indicates the positions of N/1 resonances with respect of the mean motion of the binary. See text for details. | 2012.12296 | Kepler-aemap1.jpg |
0.56811 | 0b2d11f708e94867b7adda03320e2c87 | Level curves of a fictitious Hamiltonian showing with an "O" a stable equilibrium point at (θ_s=60, I_0=2) and with an "X" an unstable equilibrium point at (θ_u=240, I_0=2). The arrows show the total resonance's width which is 2 (I_sep-I_0) being I_sep=2.44. | 2012.12296 | fictham.jpg |
0.503987 | aea2d0016a674bcfb34ec08445382d52 | Hamiltonian level curves for a fictitious planetary system composed by a star with m_0= 1 M_⊙ and two planets both with mass 0.001 M_⊙, mutual inclination of 20, e_1=0.5 and e_2=0.9 located in the resonance 3/2 with a_1=1 au. The stable equilibrium points are showed with an "O" and the unstable one is showed with an "X". Using Eq. (<ref>) our model predicts a total width of 0.048 au in a_2. | 2012.12296 | hamtwojup3to2.jpg |
0.39778 | 5858ce645742436cb0761793015d5aa1 | Resonant structure determined by the semi-analytical model in the (a_2,e_2) plane surrounding the middle planet (top) and in the (a_3,e_3) plane (bottom). Two-planet commensurabilities between planets m_1 and m_2 are shown in red while those between the middle and outer planets are presented in blue. In both plots the nominal position of the planets are identified by black filled circles, while the collision curves are depicted in gray. | 2012.12296 | hd31527_m2_m3.jpg |
0.391866 | f9b58f2308f14c38862ca546cd869ccc | Results of a long-term N-body simulation of the nominal solution for the HD31527 system. Top rows show the semimajor axis a_3 and eccentricity e_3 of the outer planet, as a function of time, for the first 10^5 years of the integration. The lower left-hand plot shows, in the green, the difference in longitude of pericenter ϖ_3-ϖ_2, while the red curve presents the behavior of the resonant angle σ_2 = 3 λ_2 - 16 λ_3 + 13ϖ_3. The lower right-hand graph shows the maximum Lyapunov exponent (LCE) for the system during the complete simulation. | 2012.12296 | hd31527_orbita.jpg |
0.451466 | a8e8d47d89d5412db3f8caeb54f7de99 | Megno map for the HD31527 system, as obtained from the numerical integration of a grid of 300 × 300 initial conditions in the (a_3,e_3) plane. All other orbital elements were set to the nominal values. Integration time was set to 10^4 years, and escapes during this timespan are highlighted in red. The color code shows values of (⟨ Y ⟩ - 2). Dark blue colors correspond to more regular orbits while lighter tones of green indicate increasingly chaotic motion. The location of the outer planet is shown with a yellow filled circle and the collision curve in gray. The libration widths of the most relevant MMRs determined analytically are plotted in white lines. See text for details. | 2012.12296 | hd31527_rhill.jpg |
0.447262 | 8ea1af263a84460a91016006794eb1df | Left: Top and bottom graphs show a Δ e dynamical map for the HD74156 system, as obtained from the numerical integration of a grid of 900 × 900 initial conditions in the (a_2,e_2) plane. Upper plot corresponds to semimajor axis a_2 ∈ [0.80,0.95] au, while the bottom frame for a_2 ∈ [0.95,1.10] au. Integration time was set to 10^3 years, and escapes during this timespan are highlighted in light yellow. The location of the middle planet is shown with a white filled circle. The libration widths of the most relevant MMRs determined analytically are highlighted in blue (resonances with outer planet) and red (resonance with inner planet). Right: Megno map for the same plane. The set of initial conditions was reduced to a 300 × 300 grid, but the integration time extended to 10^4 yrs. | 2012.12296 | hd74156_modelo_deltae_900_megno_300.jpg |
0.494708 | 058a8c3831394e6992a11d909c3ab294 | Resonance N/1 and N/2 widths setting different mass ratio between components q={0.2,1} in the plane (a,i) calculated with the semi-analytical model. For q=0.2 we use green (N/1) and magenta (N/2) while for q=1 we use orange (N/1) and blue (N/2) resonances. Horizontal lines indicates the separatrix location for nodal libration. See the text for more detail. | 2012.12296 | inclinedall.jpg |
0.421635 | e7e15a53fbc547ee82ccd74c64b286b3 | As in Fig. <ref>, but for the scale-dependent bias b(k) of the high-z galaxy samples (z̅≈0.52). The grey regions depict 68% credible regions, based on the source subsamples BACKa and BACKb high-z (z̅≈0.93,1.14). The green data points are SAM predictions. | 2012.12299 | bk_highz.jpg |
0.413814 | c6b69b6e1f134017850de12de53bcee8 | Posterior scale-dependent bias b(k) for galaxy samples at redshift z̅≈0.35. The galaxy sample is indicated in the top left corner. The orange filled regions and the dashed lines show the median, 68%, and 95% credible region for the measurement using the full source sample BACK low-z. The overlapping transparent grey regions show 68% regions that are based on the source subsamples BACKa and BACKb low-z. These noisier measurements at mean source redshifts z̅≈0.86 and z̅≈1.12 test the robustness of the measurement. The filled green diamonds and triangles indicate the SAM predictions by H15 and L12, respectively. | 2012.12299 | bk_lowz.jpg |
0.488764 | 1545219b5bc549aca7c1caf4bb6dce6e | As in Fig. <ref>, but for the scale-dependent correlation factor r(k) of the high-z samples (z̅≈0.52). The grey 68% regions are based on two source subsamples. The green data points are SAM predictions. | 2012.12299 | rk_highz.jpg |
0.445677 | 4e497b15ce074981940cb0e2a9e9c1d2 | As in Fig. <ref>, but for the scale-dependent correlation factor r(k) of the low-z galaxy samples (z̅≈0.35). The grey 68% regions are based on two source subsamples. The green data points are SAM predictions. | 2012.12299 | rk_lowz.jpg |
0.465213 | 2f82cabce2b64740911ee982124bf1fc | Schematic representation of Ising-Heisenberg tetrahedral chain. Small balls (σ_i) corresponds to Ising spins, and large balls (S_a(b),i) correspond to Heisenberg spins. | 2012.12300 | fig-fnc-1.jpg |
0.474159 | 6d8e8d0007cf48b18d6b7011aea61177 | A schematic representation of the spin-1/2 Ising-Heisenberg ladder model with alternating Ising and Heisenberg inter-leg interactions, thick vertical lines correspond to the Heisenberg coupling (J,J_z), whereas thin vertical and horizontal lines correspond to the Ising exchange interactions J_0 and J_1. | 2012.12300 | ladder-chain.jpg |
0.547995 | 4f30df5220834daca29a13a26dfdf4b2 | Log_10(S_I(E)) for symmetric states of the harmonic oscillator; I = 5,8,11,…, 32. | 2012.12303 | Fig.1HarmonicOscillator.jpg |
0.480754 | 817b01b468ef4f669ab6660480755a55 | Nesting of the partial sums Log_10(S_I(E)) centered around the ground state (i.e. E_gr = 1) for the harmonic oscillator, where I = 3,4,5,…, 12. Note that all curves share the same, fixed, minimum. | 2012.12303 | Fig.2HarmonicOscillator_gr.jpg |
0.392628 | f98a1c2f985c474197a3e7be79d5344e | Nesting of the partial sums Log_10(S_I(E)) centered around the 2^nd excited state, E_2 = 5, for the harmonic oscillator, where I = 3,4,5,…, 12. Note that their respective minima, in the energy variable, monotonically increase to E= 5; and they all have the same derivative at that point. | 2012.12303 | Fig.3HarmonicOscillator_2nd.jpg |
0.427634 | 184b554f390e4af48216a49ba8538545 | Examples of variation of the jet and soft scales to estimate perturbative uncertainty. The e_J_0 parameter is related to a `trumpet' variation in the jet scale that turns off in the fixed-order region. For the soft scale we vary μ_0, the perturbative value that the soft scale freezes to in the nonperturbative region. | 2012.12304 | AllVariation.jpg |
0.508455 | b12277c99fda4a1c8c74458bd1e520ed | Perturbative convergence of the 2-jettiness cross section, with all curves normalized with the Born cross section σ_0. The peak location shows excellent convergence, while the normalization corrections are significant and exhibit slower convergence. | 2012.12304 | eetop-MSR-Hm-MSbar-Mass-sub-On-Gap-run-On-GapSub-On-Q-700-MaxEvol-Off.jpg |
0.463742 | e18edb1edf2a4e518afb3ef09a803d6d | A high-level overview of some of the ways ethical considerations can be incorporated throughout the pipeline of abusive language detection. | 2012.12305 | pipeline9-3.jpg |
0.419056 | eeb1bd842c754199bfe8b1c02263bafd | Seed selection in a sample network (weight is 1 for all edges; barricade factor is 2 for all nodes; v_1 must be selected as a seed node for full influenceability). | 2012.12309 | SampleNetwork.jpg |
0.448623 | 979ce894d6b84dfb895a7d8f0018c407 | Bayesian Information Criterion (BIC) and Negative Log-Likelihood (NLL) curves for model orders 1-10. The NLL continues to decrease with increased model order, indicating improved model fit. However, the BIC is lowest for the model with order |Z|=3. | 2012.12310 | bic.jpg |
0.445212 | 3e910048d4204318bc3b9fc9499063dc | Number of patients (ordinate) with at least a given number of missing values (abscissa) in TRACK-Pilot. | 2012.12310 | missing.jpg |
0.42844 | 11598e3189ed498e80c7c099d536b228 | Cumulative expected error as a function of score threshold. For a given confidence score, the expected errors of all subjects with scores greater than this threshold are shown for six outcome measures. | 2012.12310 | paper_rev_results_conf_perc_3_thresh.jpg |
0.456323 | 87dc1ae27f1445bc809888ee0356c0af | Distributions of the Expected Absolute Error (EAE) of the outcome variable for the baseline and |Z|=3 models. One EAE is computed for each subject using each model in a leave-one-out procedure. | 2012.12310 | paper_rev_results_errdist_base_GOSE_OverallScore12M.jpg |
0.506362 | 98770d85886843a69522048b5969d025 | Mass attenuation coefficient (μ_m) for borate-based glasses by using Geant4 simulations in the photon energy region from 0.001 MeV to 10 MeV. | 2012.12315 | Graph1.jpg |
0.444689 | f74b4a5641d049b1b7b5281a9200f3da | Comparison between the values of μ_m obtained by Geant4 and XCOMfor borate-based glasses. | 2012.12315 | Graph2.jpg |
0.466627 | b57462aca0b04f8d85d3762670dd2ab7 | Effective atomic number (Z_eff) for borate-based glasses in the energy range 0.001 MeV to 10 MeV. | 2012.12315 | Graph3.jpg |
0.493582 | 953894f45bec412f8306cd5978df83c2 | Half value layer (HVL) as a function of photon energy for borate-based glasses. | 2012.12315 | Graph5.jpg |
0.397176 | 34164beacee84c058c2349c956a7f215 | Effective removal cross section of fast neutrons for borate-based glasses. | 2012.12315 | Graph9.jpg |
0.492672 | 7399f2c8c4d84a659e697b7af9aee141 | Euclidean distance (absolute) between values generated by each of the LCOM1–5 and the proposed | 2012.12324 | euclidean_abs.jpg |
0.465669 | 4a7e6f095041486eb7eda28f4d0d3601 | Euclidean distance (normalized) between values generated by each of the LCOM1–5 and the proposed | 2012.12324 | euclidean_norm.jpg |
0.441371 | 6a6326ba531544889887afe052c02a2a | Class methods, fields, and relationship among them for class belonging to AWS Dynamodb Encryption repository | 2012.12324 | example2.jpg |
0.427402 | df36a5905c02465d85b85d861f82cb7b | Class methods, fields, and relationship among them for class belonging to Apache Metamodel repository | 2012.12324 | example4.jpg |
0.398099 | ee6121faa7b348d498df0a1ba78194fd | Responses to the survey for each case considered for establishing ground truth | 2012.12324 | ground_truth.jpg |
0.427451 | 5551dfee3be7449fbdd70683df2f6158 | Software development experience in years of the survey participants | 2012.12324 | lcom_participants_experience.jpg |
0.420089 | ac8175990b2b4e9586130ef2361ba5fc | Survey responses toward assessing sufficiency of the commonly used metrics | 2012.12324 | metrics_responses.jpg |
0.430286 | 3afe690e225d4227b962a1e468b24b6f | Commonly used metrics by the survey participants | 2012.12324 | metrics_used_new.jpg |
0.39612 | 20dfada301c94eea96f4dba974e608f8 | Overview of the method | 2012.12324 | overview.jpg |
0.420776 | 8cc144ac81694ed0a95d9c79e1402719 | Software developer experience of the survey participants | 2012.12324 | survey_participants.jpg |
0.442638 | c26f838076034cd0bae8b837396d1c8e | The spatial geometry of a near-extremal black hole consists of a long throat whose length diverges in the extremal limit. The throat for non-extremal black holes is bounded from both ends. At the near-horizon end the AdS geometry gives way to a Rindler region and at the far end the throat is terminated at a boundary. The coordinates σ and ρ represent proper distance from the horizon and the boundary respectively. | 2012.12326 | ads2.jpg |
0.414056 | 2dab51fa54754849b605815e9418d36b | Euclidean AdS(2) can be represented by the Poincare disc. The geometry at finite temperature is cut off by a boundary shown in red. The bulk geometry is the shaded portion of the disc. | 2012.12326 | boundary.jpg |
0.513648 | a0c7765ee69040b1abc29c124885e759 | Branching diffusion. The diagrams controlling the large q behavior of operator growth in SYK are tree like. They grow exponentially with depth. The number of leaves of the tree is proportional to the number of vertices. The tree-like behavior and exponential growth breaks down at the scrambling time which only happens well after the particle has left the Throat and passed through the Rindler region. | 2012.12326 | branching.jpg |
0.421062 | 1451ba6e12f54b7b949ee8f58257a048 | ladder | 2012.12326 | ladder.jpg |
0.412412 | b232e7e5271043418f6425bfddc23db7 | q=8 “pitchfork diagrams" representing the basic unit of operator growth in the U(1)-SYK model.The expressions under each vertex indicate its weighting. | 2012.12326 | pitchforks.jpg |
0.461307 | 5445c895cbd34fc9b6fe51324d9c4cd6 | Splitting and redistributing e^ν. A propagator representing a positive fermion is shown weighted by a factor e^-ν in the left panel. In the right panel that same factor is redistributed to the vertices connected by the propagator. | 2012.12326 | splitmu.jpg |
0.433765 | b8cc8cd8c628458e917c9c1e95e19e41 | BER performance of the (a) joint, and (b) integrated RIS and relay transmission schemes for increasing N values. | 2012.12329 | 1_BER3.jpg |
0.41992 | 4b0066e3e0ec432f95a83952dd4825ed | Achievable rate of the joint RIS and relay scheme when (a) ideal and (b) non-ideal relaying are considered for a varying distance in the y-axis between the RIS and the DF relay. | 2012.12329 | 2_ideal.jpg |
0.436606 | 8836b2eb745640b7875a4db3a139b9fc | Achievable rates with the two proposed RIS and DF relay schemes for varying N values. | 2012.12329 | 3_concept2.jpg |
0.474728 | b25ec82427b74d3794b84cfc8d007b7c | Achievable rates of the joint and integrated RIS and relay transmission schemes for (a) N=256 at 2.4 GHz and (b) N=2048 at 28 GHz. | 2012.12329 | 4_two.jpg |
0.436534 | d612d8b4708e403ca029d6e3b3208b36 | (a) Achievable rate and (b) P_1 analysis for the joint RIS and relay transmission with the sequential algorithm for N=1024 at 28 GHz. | 2012.12329 | 5_Opt.jpg |
0.409983 | 4542bb6c2dfd4d5db902dbb435c26ad6 | The system model of the joint RIS and DF relay transmission scheme where the RIS and the relay are placed close to the S and D, respectively. | 2012.12329 | Fig1_2.jpg |
0.378827 | ff51f2466f0940cf9fcc152c883ec206 | The system model of the integrated RIS and DF relay transmission scheme. | 2012.12329 | Fig2_2.jpg |
0.447434 | 1b5f6925ecab44eb909cf1c4b9c4fd09 | Methodology of our approach and its validation. (a) FER lookup table generator: A hardware-in-the-loop (HiL) setup for generating the FER lookup table. The channel impulse response from a stochastic channel model with the condensed channel parameters vector Ψ (RMS delay spread σ_τ, Doppler bandwidth f_Dmax, LOS Doppler shift f_LOS, K-factor K, and received power P) is emulated. (b) System-level simulation: GSCM module for simulating the environment and computing the propagation paths; channel parameter extraction module for estimating the condensed channel parameter vector Ψ' and FER lookup table for obtaining the FER during run-time. (c) Validation: The FER obtained by emulating the channel impulse response from real world measurements is compared with the FER obtained by the system-level simulator. | 2012.12331 | process10.jpg |
0.486853 | 4d67ea8fc0644885b6f54e8ece7ac14e | Comparison with three different hash coding strategies on Cifar-10(I) for both 32 and 64 bits. Our Bi-half layer outperforms the others. | 2012.12334 | Coding_histogram2.jpg |
0.4227 | 14275e8644714e68b8d93338e6be0601 | Experimental results on Flickr25k dataset. | 2012.12334 | MAP_batch_median.jpg |
0.441047 | 0635b13941a44bb389879793d21550a5 | The estimation of median value for one hash bit vs. numbers of training epochs. Note that the estimated median value is sensitive to the batch size. | 2012.12334 | batch_translation.jpg |
0.480572 | 465c5f885a854e3098d1b7811ea13b77 | Retrieval experiments(MAP@ALL) with shorter encoding length on the Flickr25k data set. | 2012.12334 | bitnumber_Flicker.jpg |
0.432463 | bf557361a8404fb68e469ec3c6a339a3 | Hashing compresses N images to K bits per image. K continuous features are learned, and thresholded to K binary values. We see the continuous to binary transition as a lossy communication channel –a hash channel– between a single continuous value u to a single discrete binary value b. The green histograms (standard hashing) and blue histograms (our approach) show how a single feature is distributed over the N images. Instead of adding an additional loss term, we design a Bi-half layer to explicitly maximizes the bit capacity in the hash channel, leading to more informative hash codes, as measured by the entropy ℋ of the bits over the images which leads to improved hashing accuracy. | 2012.12334 | fig1.jpg |
0.443134 | 4b19d96621bc4259ba438a153ce746c3 | The proposed bi-half layer. M is the mini-batch size and K is the feature dimensions. A bi-half layer (middle part in white and blue) is used to quantize continuous features in 𝐔 into binary codes in 𝐁 via minimizing W_1(P_u, P_b) in Eq.(<ref>). The assignment strategy is the optimal probabilistic coupling π_0. For each bit, per column of 𝐔, we first rank its elements and then the top half elements is assigned to +1 and the remaining half elements to -1. In contrast, the commonly used sign function directly during training assigns the continuous features to their nearest binary codes which minimizes the Euclidean distance. The blue boxes indicate where our method differs from the sign function as the code in that position should flip. | 2012.12334 | new2.jpg |
0.501949 | 5c2a46f25a4a4f66aacbef03f50c7732 | An overview of the planning and acting architecture. | 2012.12335 | ArchitectureOverview.jpg |
0.447043 | 421db8f92ce34efe8fe0d03cc01755b4 | A level of the BoulderDash game. | 2012.12335 | BoulderDash_lv0.jpg |
0.427804 | 0d36ac0a39304d49a1ee29e8cdc9adca | SEAN-EP architecture running on multiple machines. Each host has all the components shown in Fig.<ref>. | 2012.12336 | lb.jpg |
0.461905 | 17423ee4799b434294377ca1b19d3726 | SEAN-EP enables fast data collection of HRI data in the context of navigation via 3 steps: (1) experimenters specify navigation tasks, (2) they integrate interactive simulations based on the tasks with online surveys, and (3) they collect data in parallel from multiple users. See the text for more details. | 2012.12336 | process.jpg |
0.418249 | 02e73da40b2b4a87a6c6fe068f8b5a07 | Qualtrics survey with embedded SEAN <cit.> simulations. Left: Warehouse scene. Right: Lab scene. | 2012.12336 | surveys.jpg |
0.441136 | 1b0c47511a02431284060fce82069c68 | SEAN-EP architecture running on a single machine. The host machine accepts inbound connections via the NGINX reverse proxy. | 2012.12336 | system-diagram.jpg |
0.410898 | 5f6f2cfe27dd40c4ba30f64f6b64dc6a | The prior probabilities of K (in grey) and K_+ (in black) for the three modelling approaches. | 2012.12337 | figure1.jpg |
0.452863 | 86b8fdc5dac848afa8a820595bec6b27 | The prior on the number of data clusters K_+ in dependence of γ or α characterised by prior mean, prior standard deviation (SD), prior 99%-quantile and the prior probability of K_+ = 1 for the three modelling approaches. | 2012.12337 | figure2.jpg |
0.443473 | 1a79c944ba924fbc82d42c19492d3438 | The prior on the number of data clusters K_+ in relation to the sample size N characterised by prior mean, prior standard deviation (SD), prior 99%-quantile and the prior probability of K_+ = 1 for the three modelling approaches. | 2012.12337 | figure3.jpg |
0.407822 | df797053a318446589e9eb25c53daaf0 | The prior mean and standard deviation of the relative entropy of the partitions for the three modelling approaches for K_+ ∈{2, 4, 6, 8}. | 2012.12337 | figure4.jpg |
0.393653 | 5fbe0f890f0d47bca6b12b689851bbf0 | The prior mean and standard deviation of the relative entropy of the partitions for the dynamic MFM approach with different priors on K for K_+ ∈{2, 4, 6, 8}. | 2012.12337 | figure5.jpg |
0.408982 | b0c357366c3e496388bf4692b0de98f3 | The prior mean and standard deviation of the weighted relative entropy of the partitions for the three modelling approaches. | 2012.12337 | figure6.jpg |
0.459628 | cf254338ebb9477a93f4c05c147a6be4 | The prior mean and standard deviation of the number of singletons in the partitions for the three modelling approaches for K_+ ∈{2, 4, 6, 8}. | 2012.12337 | figure7.jpg |
0.446946 | 67ecd46d966c4e2680ef6c872541ad5c | a) Ir^4+ Honeycomb planes in Ag_3LiIr_2O_6 crystal structure <cit.>. Each Ir^4+ ion is surrounded by 6 O^2- ions (red), forming an octahedron (highlighted in tan). Li^+ (green) ions are situated at the center of the Ir hexagons. b) Inter-layer Ag^+ ions (grey) in-between honeycomb planes. The grey lines indicate the unit cell of the lattice. | 2012.12340 | Wang_Fig1.jpg |
0.499176 | b6351c7c164d4eba9ed5181094a3ca2e | Knight shift of main peak for sample A (filled red bullets) along with main (open red circles) and side (open blue squares) peaks for sample B . The bulk spin susceptibility is plotted as the pink line and black line for sample A and B respectively. The dashed green horizontal line indicates the chemical shift ^7K_chem≃ 0.018% of sample A. The vertical axis of χ_spin is scaled in proportion to the Knight shift of sample A, and its origin is vertically shifted to account for the chemical shift of sample A. The inset shows Knight shift plotted against spin susceptibility χ_spin by choosing temperature as their implicit parameter. | 2012.12340 | Wang_Fig2.jpg |
0.568254 | 596d13c58b224fd0a2c51d6abb3c0785 | NMR line shapes of a) sample A and b) sample B at 30K for the fully relaxed signal (green triangles), along with the line shapes of the fast (red circles) and slow (blue squares) components. c) cascaded plot of the NMR lineshapes for sample A, observed at temperatures ranging from 4.2 K to 100 K. | 2012.12340 | Wang_Fig3.jpg |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.