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
0.424415	f453d9bd84de46b2ac4fbffff68a8361	SI-TI characteristics for test sequences of the three considered sets.	2012.14796	figure8.jpg
0.416233	f7a20fa365c04b2fa3fd74548aa9520b	Cascaded RF model prediction performance on test set.	2012.14796	figure9.jpg
0.450922	2673a04b56334476b16c98af4af7dcb2	Curves having a=-2, a=-1, a=0.2 and the rotation numbers 5,5,4, respectively.	2012.14797	fig1.jpg
0.484474	68154f3a6fce4c77a3399c61cd871acb	These curves have a=1.2, a=1.2, a=1.4, respectively.	2012.14797	fig2.jpg
0.468663	69ce833da3e44fff815837e7a562131d	These curves have a=1.5, a=1.5, a=1.75, respectively.	2012.14797	fig3.jpg
0.481723	c34d1a22038f40e09001f3772b707d02	These curves have a=2, a=2.5, a=3 and the rotation numbers 3,4,5, respectively.	2012.14797	fig4.jpg
0.38293	5404a0c5c93545968f674ae09a76e172	Number of attempts required to solve each task.	2012.14800	attempt.jpg
0.406134	521fe1916be94fe6a111297707f5018a	Usability	2012.14800	ease.jpg
0.419483	95fc438394854a669468fe65477eb0db	achieves higher F1 score than for all tasks.	2012.14800	fscore.jpg
0.406911	9e0ea09099ee46f38b6325143275a63e	Movie-watching frequency.	2012.14800	movieFrequency.jpg
0.40318	da93e67596634436af1b3e0028a0a133	Overall knowledge of movies and actors.	2012.14800	movieknowledge.jpg
0.473415	0f414656505c44dc8f80769024792a69	achieves higher precision than for all tasks.	2012.14800	precision.jpg
0.420534	a7660c84b21144afb3b62c472ee40311	Preference	2012.14800	preference.jpg
0.499912	09adaa4c4b1745ca919bcd95fb0d7aef	achieves higher recall than in two tasks.	2012.14800	recall.jpg
0.445789	7d7c5ee83a974eda88f810e6119740ca	Complete schema of the IMDb database with 8 main relations: , , , , , , , and ; and 7 connecting relations that associate the main relations: , , , , , , and .	2012.14800	schema_large.jpg
0.402566	e0786bc3af3d4f74be617004397b345c	Overall accuracy of results.	2012.14800	sqlsatisfaction.jpg
0.446376	970f19f713c34923a34c0c28eff79a67	(Adapted from <cit.>) 's offline module constructs an abduction-ready database that stores derived relations and semantic property statistics. The query intent discovery module serves the user: it takes the user-provided examples, consults the abduction-ready database, discovers the most likely query, executes it, and returns the results to the user.	2012.14800	squidArchitecture.jpg
0.386887	a7d5489a5f8a4c53b9dc8513e762520c	The graphical user interface of used in our user study. The task description is at the top. The left panel allows the users to provide examples with an auto-completion feature. infers the user's intended query from the examples, executes it, and shows the results in the right panel. The result set contains more actors, but we only show the first five (alphabetically) here.	2012.14800	squid_ui.jpg
0.410169	ab113f6c953845d9889ad0e0f4ca6a8e	Overall satisfaction with results.	2012.14800	squidsatisfaction.jpg
0.364037	3529aa26a1f744bcacc64a0c6f4cacf2	Time required to solve each task.	2012.14800	time.jpg
0.547601	1e4c1ea3761d47b496abc2aec2ad8850	Adding two types of half-twisted bands to a spanning surface S of the knot K.	2012.14801	HalfTwistedBands.jpg
0.482837	dfb9dd03c6c7464db2abcad17334f344	The family of pretzel knots P(a_1, …, a_n). Each 4-stranded box labeled with a nonzero integer a represents a pair of strands with a right-handed half-twists if a>0, and -a left-handed half-twists is a<0.	2012.14801	PretzelKnot.jpg
0.443153	6544467e91b8434dabf30154b8133355	The shaded spanning surface for the pretzel knot K=P(a_1, …, a_n) is orientable of genus (n-1)/2 if K is an odd pretzel knot (Subfigure (a) where K=P(-3,-3,3)), or nonorientable with first Betti number n-1 if K is an even pretzel knot (Subfigure (b) where K=P(-3,-3,3,2)).	2012.14801	SpanningSurfacesForPretzelKnots.jpg
0.438616	a1a1606151a141d2bd248d1a3576bdf3	Declarative Definition of the Clues for the makePayment and initiateGoingOut scripts.	2012.14803	clues.jpg
0.440611	36a8f96f83fc48f0acf4bdaa91ba809e	Eating_Out script with possible event flows	2012.14803	eatingOut.jpg
0.396467	5f136743ff5f467a9781d61b6303e002	Declarative evidence for the Eating_Out script.	2012.14803	evidence.jpg
0.402643	bcb57e17901b44eca2a88930662237b4	Simplified Facebook post analyzed according to the model	2012.14803	facebookPost.jpg
0.484214	08998cef8b864363a17a3ff84e38ee7b	Definition of EatingOut and GroceryShopping scripts	2012.14803	scriptDefinition.jpg
0.456911	c9fb99cdc4cd4780b676dc65ba8da2c1	Definitions of the AttendEatingOut sub-script and the parametric MakeAPayment<T>, GoToPlace<T> sub-scripts.	2012.14803	subscriptDef.jpg
0.396697	9739f060b2ed4afabe1d52c7885e1fdd	Double differential α emission cross section from the reaction ^120Sn(^6He,α)X at a laboratory energy of 22.2 MeV for a sum of two different angles, θ_lab=36^∘ and θ_lab=40^∘. The experimental data are taken from Ref. <cit.>	2012.14805	6He-120Sn-22-spectra.jpg
0.506837	1c8011b32804418aa52c5393e53779d9	Inclusive angular distribution for α particle emission in the reaction ^120Sn(^6He,α)X at the energies indicated in the frames. The angular distributions obtained using the experimental two-neutron separation energy in the α+dineutron model are also shown (orange lines). The experimental data were taken from Ref. <cit.>.	2012.14805	6He-120Sn-ang-dist.jpg
0.40273	fd638bc153ab4cd2963434a83946f3df	α particle angular distribution from the breakup of ^6He on ^64Zn at the laboratory energies indicated in the panels. Experimental data are taken from Refs. <cit.> (circles) and <cit.> (triangles).	2012.14805	6He-64Zn-ang-dist.jpg
0.459423	eb2d44b90f7044e1969abff22e50a27a	Comparison of inclusive α cross sections for ^6He incident on ^209Bi and ^64Zn to the data taken from <cit.> (squares), <cit.> (circles) and <cit.> (triangles). Compound nucleus cross sections (dotted lines) are computed using the EMPIRE code <cit.>. The binding energy of the projectile is taken as ε_b=- 1.6 MeV within the two-body model. See the discussion in the text.	2012.14805	6He-Ei.jpg
0.411782	80f4cbad8f3042e29bd584b1b6c5c21c	Inclusive α angular distribution in the ^6Li+^118Sn reaction. The small contribution of the compound nucleus was assumed to be isotropic. The experimental data are taken from Ref. <cit.>.	2012.14805	6Li-118Sn-ang-dist.jpg
0.450667	7037319d35e54202a9ac70fbfd0cb554	Inclusive α angular distribution in the ^6Li+^58Ni reaction. The contribution of compound nucleus emission was assumed to be isotropic. The experimental data are taken from Ref. <cit.>.	2012.14805	6Li-58Ni-ang-dist.jpg
0.490199	26da2944c51d44babb559bb0190025c0	Comparison of α angular distributions from the collisions of ^6Li and ^7Li with a target of ^58Ni at 16 MeV. The experimental data are taken from Ref. <cit.>.	2012.14805	6Li-7Li-58Ni-ang-dist.jpg
0.435781	4a3708c46a504125990bc91a1ba34814	Angular distribution of α particles from the reaction ^90Zr(^6Li,α)X. The experimental data are taken from Ref. <cit.>.	2012.14805	6Li-90Zr-ang-dist.jpg
0.47153	6a1eccb38edf439c8bf092908967bfbf	Differential energy spectrum of α particles from the breakup of ^6Li on ^90Zr at three different angles indicated in the frames. The dashed lines represent the elastic breakup. The experimental data are taken from Ref. <cit.>.	2012.14805	6Li-90Zr-spectra.jpg
0.458491	993d9aa7f4a34108ab19203210fdda08	Inclusive cross sections for α and deuteron emission. The experimental data in the panels are taken from Refs. <cit.> for ^6Li+^208Pb, and <cit.> for ^6Li+^90Zr. Compound nuclear reactions are computed using EMPIRE-III. See text for details.	2012.14805	6Li-Ei.jpg
0.414501	cfb6d6dcf99b46fd8641e60fe780daf5	Double differential cross section of α emission from ^7Li colliding with ^56Fe at a laboratory energy of 68 MeV for different angles. The experimental data are taken from Ref. <cit.>.	2012.14805	7Li-56Fe-spectra.jpg
0.376834	d9fa2dea62a647f0b5895f25dbf10fe1	α angular distribution in the collision of ^7Li on a target of ^58Ni compared to the experimental data taken from Ref. <cit.>. The large contribution from the compound nucleus was calculated with the EMPIRE code and was assumed to be isotropic.	2012.14805	7Li-58Ni-ang-dist.jpg
0.41018	9af2a63191c946618efedd6b425c428f	The computational time of each problem.	2012.14808	Consumed_time_uopt.jpg
0.450941	e0148f87fce44febb85443487aa010d5	Panel a) Dispersion of DDGPE linearized Bogoliubov-like excitations obtained via BdG equations for the system of σ^+ polarization component in sufficiently non-linear regime and σ^- component with lower density. Panel b) The dispersion from all space obtained from the numerical simulations. The sum of signals from σ^+ and σ^- is given.	2012.14814	new3.jpg
0.413909	b815d925bf1a48af99ca446cda596b13	Panel a) Time-averaged over 25 ns interference pattern \|ψ(𝐑_0)+ψ(𝐫)\|^2 of some fixed point 𝐑_0 with all other points r. The point 𝐑_0 is visible as bright spot. Fringes show the presence of a flow with k ≈ 0.3 μm^-1, which matches with instability regions of linearized excitations. Panel b) Dispersion from the region marked by dashed square in panel a). The dominating direction is well visible.	2012.14814	new5.jpg
0.457706	e200e6cb073c4a37a288f28259419840	Coarse graining a dislocation microstructure (left) with averaging voxels of size b indicated by the dashed lines results in the GND density shown on the right. Such coarse graining would completely lose the information of the dislocation dipole on the top left voxel and the difference of locations for single dislocations inside the other voxels. Any energy measure based on the GND density is afflicted by this information loss.	2012.14815	figure-1-2d.jpg
0.458655	7199bf85240a41a5bc74b92b71b26a53	Sketch of the averaging procedure: the energy density around a dislocation is sampled with a high resolution (≈ 1b) as indicated by the small, colored voxels (1.). Only the core region (2.) of the dislocation is excluded (note the missing voxels in the vicinity of the dislocation line). Averaging the energy data inside the coarse graining volumes (the larger volumes drawn in blue, (3.))	2012.14815	figure-1.jpg
0.427145	f177c084e7394afd8f0158f6125131f9	Summary of the fitting parameters for 2D and 3D dislocation microstructures. For 2D results, each data point is the average of 30 samples. For 3D results, the symbol size represents the sample size since both 30nm and 60nm sample are analyzed.	2012.14815	figure-10.jpg
0.450686	018028931c5441c68e0a8f2f27f93a75	Fitting parameters based on dislocation microstructures obtained in tension (top row) and bending (bottom row). The error bar is the standard deviation of 20 realizations. In the right column, the contribution to the energy from each term, are shown for different voxel sizes.	2012.14815	figure-11.jpg
0.430379	6247d02f95894ef38c30c02ea6ba65a3	Fitting parameters based on dislocation microstructures of random loops. Fitting parameters as a function of the voxel size are shown. The error bar is the standard deviation of 10 realizations. Curve colors correspond to the field quantity in the bottom right figure where the energy contribution of each term to the total strain energy density is shown.	2012.14815	figure-12.jpg
0.404855	f514389c9c654fd98311d34ed3aef5c4	Fitting parameters based on relaxed dislocation microstructures. Fitting parameters as a function of the voxel size are shown. The error bar is the standard deviation of 5 realizations. Curve colors correspond to the field quantity in the bottom right figure where the energy contribution of each term to the total strain energy density is shown.	2012.14815	figure-13.jpg
0.389629	220fd4b884a34830909bc1f1d6491d01	Schematic of the data-mining in 2D. The dislocation structure consists of different signs of dislocations (represented by different colors) is firstly coarse grained by the reference voxel of size Δ x=b, Δ y=b as shown in the top row. With the coarse grained dislocation density, the strain energy density (the third column) can be calculated. If we average the strain energy density data in a desired coarse voxel size, for example Δ x=10b, Δ y=10b indicated by the green dashed line, we get ψ^micro for the large voxel size, shown in the last column of the first row. For coarse graining voxel size Δ x=10b, Δ y=10b as shown in the middle row, we can calculate the corresponding dislocation density and the strain energy density ψ^meso. Δψ, shown in the last row, is the difference between ψ^micro and ψ^meso.	2012.14815	figure-2.jpg
0.394	ed631286ebd44bd89ad42f81023c0ec0	Top: Schematic of the two simulated problems. The red dots represent the positions of Frank-Read sources. M and θ stand for the bending moment and bending angle respectively. Middle: Typical mechanical responses (10 realizations) in DDD simulations. Different realizations have the same dislocation source density, but different source positions. Left: σ versus ε in tension; Right: bending moment versus bending angle in bending; Bottom left: Typical dislocation microstructure in tension; Bottom right: Typical dislocation microstructure in bending.	2012.14815	figure-4.jpg
0.415843	8c66d61301a9447a98feb9de4b2b7bef	Top row: coarse grained dislocation density. Middle left: ψ^micro in tension; Bottom left: ψ^meso in tension; Middle right: ψ^micro in bending; Bottom right: ψ^meso in bending.	2012.14815	figure-5.jpg
0.41315	a7d300ec75c54783949b4a98b3a83db7	Strain energy density as a function of dislocation density in the voxel of Δ x = Δ y=100b. Analysis of the energy based on a dislocation configuration obtained through (left) tension and (right) bending. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+CGb^2ρlog(ρ/ρ_0) is denoted by asterisks in black.	2012.14815	figure-6.jpg
0.436461	0c2b5d012b3241328890bdd81cadaf28	Dependence of the fitting parameter C on the voxel size, (left) tension and (right) bending. The red asterisks are data without consideration of the finite boundary while the black dots are data that include the finite boundary effect.	2012.14815	figure-7.jpg
0.415765	70fccfa1275b4db0ba7d059667deb6ca	Top left: dislocation density field of random dislocation loops when the voxel size is b; Top right: coarse-grained continuum dislocation density field when the voxel size is chosen as 20b; Bottom left: analysis of the energy density when the coarse voxel size is chosen as 20b; Bottom right: analysis of the energy density when the coarse voxel size is 40b. The domain size is 60nm. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+Gb^2ρlog(ρ/ρ_0) is denoted by asterisks in orange. The corresponding voxel size in each data plot is on the top right of each figure.	2012.14815	figure-8.jpg
0.371887	005750c147184e57a50440590e3485ca	Top left: dislocation density field of relaxed dislocation microstructure when the voxel size is b. The field approaches a discrete description of the dislocation microstructure; Top right: coarse-grained continuum dislocation density field when the voxel size is chose as 20b; Bottom left: analysis of the energy density when the coarse voxel size is chosen as 20b; Bottom right: analysis of the energy density when the coarse voxel size is 40b. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+Gb^2ρlog(ρ/ρ_0) is denoted by asterisks in orange. The corresponding voxel size in each data plot is on the top right of each figure.	2012.14815	figure-9.jpg
0.446369	ba1cff79d47646c28fa9f3ad61864baa	Schematic of boundary value problem of coarse grained dislocation configurations.	2012.14815	figure-A1.jpg
0.430509	dd685667fb114ffabac9ac9043285844	Fitting parameters as a function of the reference voxel size.	2012.14815	figure-A2.jpg
0.514749	e17ac13939b44a0c98c380a1f60032df	Architecture of LC-Net	2012.14816	LCNet_HR.jpg
0.489987	d84cd25be3b745df9bc97f7475d0d02d	Average probability of error with respect to the learning database size. Notice that the abscissa scale is logarithmic.	2012.14816	Results_curves_acc_recent.jpg
0.465947	4da127ef145e4420ad19fb270e1cf89a	Schematic generic evolution of the test error depending on the dataset size.	2012.14816	law_dataset_size.jpg
0.469509	3567a70227d44793b34e7ae5f0690e97	Schematic generic evolution of the test error depending on the model size.	2012.14816	law_model_size.jpg
0.456698	ae5e4bf9a9ad47f0ac2e74b216086ca1	The summary of the feasibility of the proposed method. The pixel-wise differences are obtained via subtracting the pixel values in the artificial images from the corresponding pixel values in the corresponding reconstructed images. Two, three, and four pixels from each side of the images are excluded from the pixel-wise-difference calculations when the size of the resampling matrix is three-by-three, five-by-five, and seven-by-seven, respectively. As a comparative base, the values in parentheses are the corresponding results when the elements in Equations (10), (11), and (12) are replaced by the values: 1/9, 1/25, and 1/49, respectively.	2012.14817	Table_1.jpg
0.476528	29aad7f9415545c6a9a9e0ef50b93e47	GI I diagrams for W^+ W^- fusion.	2012.14818	fig13_new.jpg
0.47218	65b26d020b5c4a57af836b0a9d158cf4	(A) Hall resistivity measured at several different temperatures as a function of magnetic field with 𝐣∥[100] and 𝐁∥[001]. The inset presents the Hall resistivity in weak magnetic field region. (B) Magnetic field dependent Hall conductivity measured at T= 2 K. The red solid line represents the fit with Equation (1). Quality of the fit was excellent (with R^2>99.99). Inset shows the geometry of measurement.	2012.14819	Fig1new.jpg
0.523852	0d76cb9d9f364d0799fca2131ae7cc15	(A) MR(B) isotherms measured at T=2 K for different angles between current and magnetic field. Inset shows schematically the geometry of measurement. (B) LMR(B) isotherms for several temperatures. (C) Analysis of the transverse magnetoconductivity data obtained at T= 2 K. Blue points - experimental data, green solid line - σ_xx^D(B) calculated with Eq. (<ref>). Grey solid curve represents the WAL contribution derived as described in the text, and red solid line curve shows the fit of Eq. (<ref>). Inset: temperature dependence of the critical field approximated by power law B_c∝T^1.35. (D) Magnetic field dependence of the longitudinal magnetoconductivity for several temperatures. Red solid lines correspond to the fits of Eq. (<ref>). Inset: temperature-dependent chiral coefficient (solid line is a guide for the eye).	2012.14819	Fig2new.jpg
0.437477	2f9231d9e3b240a9b194108ff929f446	Angular dependence of the planar Hall resistivity measured in different external magnetic fields at 2 K (A), and 300 K (B), with experimental geometry sketched in inset to panel (A). Red solid lines represent the fits obtained using Eq. (<ref>). (C) Magnetic field variations of the PHE magnitude observed at several different temperatures. Lines are guides for the eye.	2012.14819	Fig3new.jpg
0.476286	3a1ce03608874775a044fd3c4cbf4628	(A) Angular dependence of the electrical resistivity measured at several temperatures in a magnetic field of 14 T. The inset shows the geometry of measurement. (B) Anisotropic magnetoresistance isotherms calculated from the data presented in (A). (C) Conductivity enhancement derived from the data shown in (A). Red lines represent the results of fitting Eq. (<ref>) to the experimental data. The inset displays the temperature dependence of the narrowing parameter L defined in the text. The solid line is a guide for the eye.	2012.14819	Fig4new.jpg
0.456577	7806999e69154ab586e36f560bb30a26	(A) Band structure and (B) total DOS calculated (including spin-orbit interaction) for the ideal crystal structure of ScPtBi. (C) Band structure and (D) total DOS calculated (including spin-orbit interaction) for the model with Pt occupying the 16e Wyckoff position. The red ellipse in C marks a gap between inverted bands where Dirac points may occur. (E,F) Total DOS for models: (E) with different antisites, and (F) with different vacancies.	2012.14819	Fig5new.jpg
0.44438	395e79ddd75b4b3aaa38cd13a5fd7f68	Energy-dispersive X-ray spectrum (EDS), SEM image (left inset) and photograph (right inset) of ScPtBi single crystal. Chemical composition determined from the spectrum is close to the equiatomic: Sc_33.33Pt_34.03Bi_32.64.	2012.14819	FigS1_EDS.jpg
0.433978	afa11c8cfa90427789f3336036530e96	Temperature dependent specific heat of ScPtBi. Red solid line corresponds to the fit with C=γ T+β T^3 equation. Blue bar corresponds to the specific heat jump expected for a conventional superconductor.	2012.14819	FigS3_HC.jpg
0.483915	4c617d4ad2144cb78cb8148b1d12cf6e	Electronic band structure of ScPtBi for crystal structure models with different antisite defects. Note both hole-like and electron-like bands crossing the Fermi energy E_ F.	2012.14819	FigS8rev.jpg
0.477867	b340829a47f7476e81e9e6f07670798c	Hall resistance measured at several temperatures with electric current 𝐣∥ [100]. Inset shows zoom in weak magnetic field range. Hall resistivity data shown in Fig. 1A of the main text were obtained through anti-symmetrization of data shown here, as: ρ_xy(B)=t(R_xy(B)-R_xy(-B))/2, where t was the thickness of the sample.	2012.14819	ScPtBi_raw_data_Hall.jpg
0.43861	8ff25da609604970a5e5128a51a0701d	The analyzed data.	2012.14820	data.jpg
0.442649	1cf29d1d8d0a48d395ebd5926f5f0b99	The unit root transformations od the analyzed data.	2012.14820	data_trans.jpg
0.411929	90f0e22399e247d691f6f6f5181edcdc	Kinematic distributions for different decay modes of charged kaons at the NA62 experiment <cit.>. Here the normalized differential decay width for a given mode is plotted against the squared missing energy s. The two rare kaon decay modes K^+→π^+ + νν̅/νν have been multiplied by a factor 10^10 for visibility, all other decay modes are treated as background processes. The two shaded areas correspond to the two signal regions of the NA62 experiment.	2012.14825	distributions.jpg
0.477919	556ce11875ac47c7978aef9548967ce0	(Color online) Phase diagrams for the aluminum-nickel Al_(100-x)Ni_x system (left column) and the iron-nickel Fe_(100-x)Ni_x system (right column) <cit.>. The investigated region of the phase diagram in this study is highlighted in rose color.	2012.14826	fig01.jpg
0.56404	20cda57ed2fc428db3d72ddefc5a0619	Concentration dependences of shear viscosity of aluminum-nickel (left column) and iron-nickel (right column) melts at different temperatures: ( ) – results of molecular dynamics simulation; (∘ ∘ ∘) – results of the experiment on viscometry; markers (□ □ □), ( ), ( ), (▪ ▪ ▪), (∙ ∙ ∙) – experimental data taken from Refs. <cit.>, <cit.>, <cit.>, <cit.>, <cit.>, respectively.	2012.14826	fig02.jpg
0.605049	6e93dd9eb6d443c8b68fccadf932c1c4	(Color online) Temperature dependencies of the reduced viscosity for Al_(100-x)Ni_x (left column) and Fe_(100-x)Ni_x melts (right column) at various concentrations of nickel atoms: markers represent the results of molecular dynamics simulation; solid line – universal viscosity curve calculated by the formula (<ref>). Insert: Temperature dependence of the logarithm of viscosity, showing the deviation from Arrhenius's law below the temperature T_A.	2012.14826	fig03.jpg
0.40695	7e08fd21cfc24129a9f148c8a96f3ee9	(Color online) Reduced viscosity η^* as a function of configurational entropy for binary nickel-containing metal melts: markers – the results of molecular dynamics simulation; the dotted line is the Rosenfeld's scale ratio for viscosity with parameters B=1.02 and β=0.65.	2012.14826	fig04.jpg
0.418258	ccd4063c4f3d443cabefde6d722625de	An example dialog from ShARC benchmark dataset <cit.>. At each turn, the machine can give a decision regarding the initial question put up by the user. If the decision is Inquire, the machine will ask a clarification question to help with decision making. The corresponding rule document and the question are marked in the same color in the figure.	2012.14827	example.jpg
0.420559	70971304035c47d7bb26ea4bb4ed6573	Examples selected from the dev set of ShARC where Discern fails but our model succeeds.	2012.14827	examples.jpg
0.465718	da102ff4260d497d97ff5bcb9587c175	The overall structure for our proposed model. With segmented EDUs and tagged relations, the inputs including user initial question, user scenario and dialog history are sent for embedding and graph modeling to make the final decision. If the decision is Inquire, the question generation stage will be activated and use the under-specified span of rule document to generate a follow-up question.	2012.14827	pipeline.jpg
0.536929	7baf04473e054db7a85842f5fedcd405	The figure displays free energy F of the thermalon configuration as a function of the Rényi temperature T_R^(+) for several values of the non-extensive parameter α. We have used L=1, and λ =1.1382. From left to right: α= -0.25 (blue), α =-0.10 (green), α=0.00 (red), α=0.10 (magenta) and α=0.50 (orange). For each value of the non-extensive parameter α, the upper branch beyond the cusp is unphysical where it corresponds to Π^+ = - Π^- solutions while the lower branch is the physical solutions of Π^+=Π^-.	2012.14828	FT.jpg
0.504544	441ce2e5e60a44db919278d081f16529	VLBI image of NGC 315 at 86 GHz.	2012.14831	86_1.jpg
0.385117	186e015198c840d193f2e2741fe75c2e	Core positions as a function of frequency relative to the mm-VLBI cores (43 GHz and 86 GHz).	2012.14831	CoreShift.jpg
0.395806	14665c929f7945ae8787b899e7774faf	Residuals in the innermost jet region for a single power-law fit performed considering the VLBI data only and the entire data set (VLBI+VLA).	2012.14831	NGC315_SingleRes_vla.jpg
0.392072	0ab3b0be72a743b5abee2e2dc0dca39c	Jet structure of NGC 315 from Mega-parsec to sub-parsec scales. The top image was obtained based on VLA data at 1.4 GHz, while the rest of the maps show the VLBI structure at different frequencies (5 GHz, 22 GHz, 43 GHz, 86 GHz) based on some of analyzed the data presented in Sect. 2, Table 1. The complete series of images considered in the article is presented in the Appendix.	2012.14831	multinew.jpg
0.481661	4c0d92a6ea5d4d9d8d2ca64ab91ad65d	Block diagram overview of the multi-modal image registration process utilized.	2012.14833	alignment_process.jpg
0.394197	bb2873c3f7214762becda7042debfde6	Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Car-Nighttime dataset.	2012.14833	car_night_selected.jpg
0.387944	97af855fd4244ba0a30cd9723cbb1aea	Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Car-Daytime dataset.	2012.14833	car_selected.jpg
0.414932	336ffcbd274b41babd31869b99733502	Indicative images from the raw visual and thermal camera frames of the released dataset. Top row: Car-Daytime dataset, second row: Car-Nighttime dataset, third row: Aerial Scout-Mine dataset, bottom row: Aerial Scout-Tunnel dataset.	2012.14833	dataset_raw_images.jpg
0.425407	5b54717fee7f4f4db034a2f7be670bbb	Indicative histogram plots (256 bins) for visual and thermal camera data from the released datasets.	2012.14833	histograms_examples.jpg
0.444436	030d2945f680417da380fe09e7634247	Indicative 32× 32 image patches from visual and thermal data with the centroid of each visual patch being the coordinates of a FAST detection and the centroid of the thermal patch acquired through the identified alignment transformation. One image pair for each of the datasets is considered for this result and only six (randomly selected) detected corners are presented for each image pair. Naturally, datasets that are rich both visually and thermally allow for rather more “matching” multi-modal frames.	2012.14833	local_features_datasets.jpg
0.421678	e5d59bca84c447d787118b62a275851c	Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Aerial Scout-Mine dataset.	2012.14833	trjv_selected.jpg
0.418685	d10f9ae67336480c8f7ce59d9c8c2c30	Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Aerial Scout-Tunnel dataset.	2012.14833	truckee_selected.jpg
0.420594	6cbb205158b04275b1e6f09781f279cf	Optimal and maximum EH time	2012.14834	EHopt_vs_max_v2.jpg
0.383272	41564076ba1f4314b9441839163a2cb2	Optimal and maximum power	2012.14834	Popt_vs_fix_v2.jpg
0.419032	186ac6c791854ad3ad35d755fbdfecfc	Different ToA allocations	2012.14834	SFalloc_v2.jpg

0.424415

f453d9bd84de46b2ac4fbffff68a8361

SI-TI characteristics for test sequences of the three considered sets.

2012.14796

figure8.jpg

0.416233

f7a20fa365c04b2fa3fd74548aa9520b

Cascaded RF model prediction performance on test set.

2012.14796

figure9.jpg

0.450922

2673a04b56334476b16c98af4af7dcb2

Curves having a=-2, a=-1, a=0.2 and the rotation numbers 5,5,4, respectively.

2012.14797

fig1.jpg

0.484474

68154f3a6fce4c77a3399c61cd871acb

These curves have a=1.2, a=1.2, a=1.4, respectively.

2012.14797

fig2.jpg

0.468663

69ce833da3e44fff815837e7a562131d

These curves have a=1.5, a=1.5, a=1.75, respectively.

2012.14797

fig3.jpg

0.481723

c34d1a22038f40e09001f3772b707d02

These curves have a=2, a=2.5, a=3 and the rotation numbers 3,4,5, respectively.

2012.14797

fig4.jpg

0.38293

5404a0c5c93545968f674ae09a76e172

Number of attempts required to solve each task.

2012.14800

attempt.jpg

0.406134

521fe1916be94fe6a111297707f5018a

Usability

2012.14800

ease.jpg

0.419483

95fc438394854a669468fe65477eb0db

achieves higher F1 score than for all tasks.

2012.14800

fscore.jpg

0.406911

9e0ea09099ee46f38b6325143275a63e

Movie-watching frequency.

2012.14800

movieFrequency.jpg

0.40318

da93e67596634436af1b3e0028a0a133

Overall knowledge of movies and actors.

2012.14800

movieknowledge.jpg

0.473415

0f414656505c44dc8f80769024792a69

achieves higher precision than for all tasks.

2012.14800

precision.jpg

0.420534

a7660c84b21144afb3b62c472ee40311

Preference

2012.14800

preference.jpg

0.499912

09adaa4c4b1745ca919bcd95fb0d7aef

achieves higher recall than in two tasks.

2012.14800

recall.jpg

0.445789

7d7c5ee83a974eda88f810e6119740ca

Complete schema of the IMDb database with 8 main relations: , , , , , , , and ; and 7 connecting relations that associate the main relations: , , , , , , and .

2012.14800

schema_large.jpg

0.402566

e0786bc3af3d4f74be617004397b345c

Overall accuracy of results.

2012.14800

sqlsatisfaction.jpg

0.446376

970f19f713c34923a34c0c28eff79a67

(Adapted from <cit.>) 's offline module constructs an abduction-ready database that stores derived relations and semantic property statistics. The query intent discovery module serves the user: it takes the user-provided examples, consults the abduction-ready database, discovers the most likely query, executes it, and returns the results to the user.

2012.14800

squidArchitecture.jpg

0.386887

a7d5489a5f8a4c53b9dc8513e762520c

The graphical user interface of used in our user study. The task description is at the top. The left panel allows the users to provide examples with an auto-completion feature. infers the user's intended query from the examples, executes it, and shows the results in the right panel. The result set contains more actors, but we only show the first five (alphabetically) here.

2012.14800

squid_ui.jpg

0.410169

ab113f6c953845d9889ad0e0f4ca6a8e

Overall satisfaction with results.

2012.14800

squidsatisfaction.jpg

0.364037

3529aa26a1f744bcacc64a0c6f4cacf2

Time required to solve each task.

2012.14800

time.jpg

0.547601

1e4c1ea3761d47b496abc2aec2ad8850

Adding two types of half-twisted bands to a spanning surface S of the knot K.

2012.14801

HalfTwistedBands.jpg

0.482837

dfb9dd03c6c7464db2abcad17334f344

The family of pretzel knots P(a_1, …, a_n). Each 4-stranded box labeled with a nonzero integer a represents a pair of strands with a right-handed half-twists if a>0, and -a left-handed half-twists is a<0.

2012.14801

PretzelKnot.jpg

0.443153

6544467e91b8434dabf30154b8133355

The shaded spanning surface for the pretzel knot K=P(a_1, …, a_n) is orientable of genus (n-1)/2 if K is an odd pretzel knot (Subfigure (a) where K=P(-3,-3,3)), or nonorientable with first Betti number n-1 if K is an even pretzel knot (Subfigure (b) where K=P(-3,-3,3,2)).

2012.14801

SpanningSurfacesForPretzelKnots.jpg

0.438616

a1a1606151a141d2bd248d1a3576bdf3

Declarative Definition of the Clues for the makePayment and initiateGoingOut scripts.

2012.14803

clues.jpg

0.440611

36a8f96f83fc48f0acf4bdaa91ba809e

Eating_Out script with possible event flows

2012.14803

eatingOut.jpg

0.396467

5f136743ff5f467a9781d61b6303e002

Declarative evidence for the Eating_Out script.

2012.14803

evidence.jpg

0.402643

bcb57e17901b44eca2a88930662237b4

Simplified Facebook post analyzed according to the model

2012.14803

facebookPost.jpg

0.484214

08998cef8b864363a17a3ff84e38ee7b

Definition of EatingOut and GroceryShopping scripts

2012.14803

scriptDefinition.jpg

0.456911

c9fb99cdc4cd4780b676dc65ba8da2c1

Definitions of the AttendEatingOut sub-script and the parametric MakeAPayment<T>, GoToPlace<T> sub-scripts.

2012.14803

subscriptDef.jpg

0.396697

9739f060b2ed4afabe1d52c7885e1fdd

Double differential α emission cross section from the reaction ^120Sn(^6He,α)X at a laboratory energy of 22.2 MeV for a sum of two different angles, θ_lab=36^∘ and θ_lab=40^∘. The experimental data are taken from Ref. <cit.>

2012.14805

6He-120Sn-22-spectra.jpg

0.506837

1c8011b32804418aa52c5393e53779d9

Inclusive angular distribution for α particle emission in the reaction ^120Sn(^6He,α)X at the energies indicated in the frames. The angular distributions obtained using the experimental two-neutron separation energy in the α+dineutron model are also shown (orange lines). The experimental data were taken from Ref. <cit.>.

2012.14805

6He-120Sn-ang-dist.jpg

0.40273

fd638bc153ab4cd2963434a83946f3df

α particle angular distribution from the breakup of ^6He on ^64Zn at the laboratory energies indicated in the panels. Experimental data are taken from Refs. <cit.> (circles) and <cit.> (triangles).

2012.14805

6He-64Zn-ang-dist.jpg

0.459423

eb2d44b90f7044e1969abff22e50a27a

Comparison of inclusive α cross sections for ^6He incident on ^209Bi and ^64Zn to the data taken from <cit.> (squares), <cit.> (circles) and <cit.> (triangles). Compound nucleus cross sections (dotted lines) are computed using the EMPIRE code <cit.>. The binding energy of the projectile is taken as ε_b=- 1.6 MeV within the two-body model. See the discussion in the text.

2012.14805

6He-Ei.jpg

0.411782

80f4cbad8f3042e29bd584b1b6c5c21c

Inclusive α angular distribution in the ^6Li+^118Sn reaction. The small contribution of the compound nucleus was assumed to be isotropic. The experimental data are taken from Ref. <cit.>.

2012.14805

6Li-118Sn-ang-dist.jpg

0.450667

7037319d35e54202a9ac70fbfd0cb554

Inclusive α angular distribution in the ^6Li+^58Ni reaction. The contribution of compound nucleus emission was assumed to be isotropic. The experimental data are taken from Ref. <cit.>.

2012.14805

6Li-58Ni-ang-dist.jpg

0.490199

26da2944c51d44babb559bb0190025c0

Comparison of α angular distributions from the collisions of ^6Li and ^7Li with a target of ^58Ni at 16 MeV. The experimental data are taken from Ref. <cit.>.

2012.14805

6Li-7Li-58Ni-ang-dist.jpg

0.435781

4a3708c46a504125990bc91a1ba34814

Angular distribution of α particles from the reaction ^90Zr(^6Li,α)X. The experimental data are taken from Ref. <cit.>.

2012.14805

6Li-90Zr-ang-dist.jpg

0.47153

6a1eccb38edf439c8bf092908967bfbf

Differential energy spectrum of α particles from the breakup of ^6Li on ^90Zr at three different angles indicated in the frames. The dashed lines represent the elastic breakup. The experimental data are taken from Ref. <cit.>.

2012.14805

6Li-90Zr-spectra.jpg

0.458491

993d9aa7f4a34108ab19203210fdda08

Inclusive cross sections for α and deuteron emission. The experimental data in the panels are taken from Refs. <cit.> for ^6Li+^208Pb, and <cit.> for ^6Li+^90Zr. Compound nuclear reactions are computed using EMPIRE-III. See text for details.

2012.14805

6Li-Ei.jpg

0.414501

cfb6d6dcf99b46fd8641e60fe780daf5

Double differential cross section of α emission from ^7Li colliding with ^56Fe at a laboratory energy of 68 MeV for different angles. The experimental data are taken from Ref. <cit.>.

2012.14805

7Li-56Fe-spectra.jpg

0.376834

d9fa2dea62a647f0b5895f25dbf10fe1

α angular distribution in the collision of ^7Li on a target of ^58Ni compared to the experimental data taken from Ref. <cit.>. The large contribution from the compound nucleus was calculated with the EMPIRE code and was assumed to be isotropic.

2012.14805

7Li-58Ni-ang-dist.jpg

0.41018

9af2a63191c946618efedd6b425c428f

The computational time of each problem.

2012.14808

Consumed_time_uopt.jpg

0.450941

e0148f87fce44febb85443487aa010d5

Panel a) Dispersion of DDGPE linearized Bogoliubov-like excitations obtained via BdG equations for the system of σ^+ polarization component in sufficiently non-linear regime and σ^- component with lower density. Panel b) The dispersion from all space obtained from the numerical simulations. The sum of signals from σ^+ and σ^- is given.

2012.14814

new3.jpg

0.413909

b815d925bf1a48af99ca446cda596b13

Panel a) Time-averaged over 25 ns interference pattern |ψ(𝐑_0)+ψ(𝐫)|^2 of some fixed point 𝐑_0 with all other points r. The point 𝐑_0 is visible as bright spot. Fringes show the presence of a flow with k ≈ 0.3 μm^-1, which matches with instability regions of linearized excitations. Panel b) Dispersion from the region marked by dashed square in panel a). The dominating direction is well visible.

2012.14814

new5.jpg

0.457706

e200e6cb073c4a37a288f28259419840

Coarse graining a dislocation microstructure (left) with averaging voxels of size b indicated by the dashed lines results in the GND density shown on the right. Such coarse graining would completely lose the information of the dislocation dipole on the top left voxel and the difference of locations for single dislocations inside the other voxels. Any energy measure based on the GND density is afflicted by this information loss.

2012.14815

figure-1-2d.jpg

0.458655

7199bf85240a41a5bc74b92b71b26a53

Sketch of the averaging procedure: the energy density around a dislocation is sampled with a high resolution (≈ 1b) as indicated by the small, colored voxels (1.). Only the core region (2.) of the dislocation is excluded (note the missing voxels in the vicinity of the dislocation line). Averaging the energy data inside the coarse graining volumes (the larger volumes drawn in blue, (3.))

2012.14815

figure-1.jpg

0.427145

f177c084e7394afd8f0158f6125131f9

Summary of the fitting parameters for 2D and 3D dislocation microstructures. For 2D results, each data point is the average of 30 samples. For 3D results, the symbol size represents the sample size since both 30nm and 60nm sample are analyzed.

2012.14815

figure-10.jpg

0.450686

018028931c5441c68e0a8f2f27f93a75

Fitting parameters based on dislocation microstructures obtained in tension (top row) and bending (bottom row). The error bar is the standard deviation of 20 realizations. In the right column, the contribution to the energy from each term, are shown for different voxel sizes.

2012.14815

figure-11.jpg

0.430379

6247d02f95894ef38c30c02ea6ba65a3

Fitting parameters based on dislocation microstructures of random loops. Fitting parameters as a function of the voxel size are shown. The error bar is the standard deviation of 10 realizations. Curve colors correspond to the field quantity in the bottom right figure where the energy contribution of each term to the total strain energy density is shown.

2012.14815

figure-12.jpg

0.404855

f514389c9c654fd98311d34ed3aef5c4

Fitting parameters based on relaxed dislocation microstructures. Fitting parameters as a function of the voxel size are shown. The error bar is the standard deviation of 5 realizations. Curve colors correspond to the field quantity in the bottom right figure where the energy contribution of each term to the total strain energy density is shown.

2012.14815

figure-13.jpg

0.389629

220fd4b884a34830909bc1f1d6491d01

Schematic of the data-mining in 2D. The dislocation structure consists of different signs of dislocations (represented by different colors) is firstly coarse grained by the reference voxel of size Δ x=b, Δ y=b as shown in the top row. With the coarse grained dislocation density, the strain energy density (the third column) can be calculated. If we average the strain energy density data in a desired coarse voxel size, for example Δ x=10b, Δ y=10b indicated by the green dashed line, we get ψ^micro for the large voxel size, shown in the last column of the first row. For coarse graining voxel size Δ x=10b, Δ y=10b as shown in the middle row, we can calculate the corresponding dislocation density and the strain energy density ψ^meso. Δψ, shown in the last row, is the difference between ψ^micro and ψ^meso.

2012.14815

figure-2.jpg

0.394

ed631286ebd44bd89ad42f81023c0ec0

Top: Schematic of the two simulated problems. The red dots represent the positions of Frank-Read sources. M and θ stand for the bending moment and bending angle respectively. Middle: Typical mechanical responses (10 realizations) in DDD simulations. Different realizations have the same dislocation source density, but different source positions. Left: σ versus ε in tension; Right: bending moment versus bending angle in bending; Bottom left: Typical dislocation microstructure in tension; Bottom right: Typical dislocation microstructure in bending.

2012.14815

figure-4.jpg

0.415843

8c66d61301a9447a98feb9de4b2b7bef

Top row: coarse grained dislocation density. Middle left: ψ^micro in tension; Bottom left: ψ^meso in tension; Middle right: ψ^micro in bending; Bottom right: ψ^meso in bending.

2012.14815

figure-5.jpg

0.41315

a7d300ec75c54783949b4a98b3a83db7

Strain energy density as a function of dislocation density in the voxel of Δ x = Δ y=100b. Analysis of the energy based on a dislocation configuration obtained through (left) tension and (right) bending. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+CGb^2ρlog(ρ/ρ_0) is denoted by asterisks in black.

2012.14815

figure-6.jpg

0.436461

0c2b5d012b3241328890bdd81cadaf28

Dependence of the fitting parameter C on the voxel size, (left) tension and (right) bending. The red asterisks are data without consideration of the finite boundary while the black dots are data that include the finite boundary effect.

2012.14815

figure-7.jpg

0.415765

70fccfa1275b4db0ba7d059667deb6ca

Top left: dislocation density field of random dislocation loops when the voxel size is b; Top right: coarse-grained continuum dislocation density field when the voxel size is chosen as 20b; Bottom left: analysis of the energy density when the coarse voxel size is chosen as 20b; Bottom right: analysis of the energy density when the coarse voxel size is 40b. The domain size is 60nm. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+Gb^2ρlog(ρ/ρ_0) is denoted by asterisks in orange. The corresponding voxel size in each data plot is on the top right of each figure.

2012.14815

figure-8.jpg

0.371887

005750c147184e57a50440590e3485ca

Top left: dislocation density field of relaxed dislocation microstructure when the voxel size is b. The field approaches a discrete description of the dislocation microstructure; Top right: coarse-grained continuum dislocation density field when the voxel size is chose as 20b; Bottom left: analysis of the energy density when the coarse voxel size is chosen as 20b; Bottom right: analysis of the energy density when the coarse voxel size is 40b. The energy function Gb^2ρlog(ρ/ρ_0) is plotted as solid lines. ψ^meso_ V+Gb^2ρlog(ρ/ρ_0) is denoted by asterisks in orange. The corresponding voxel size in each data plot is on the top right of each figure.

2012.14815

figure-9.jpg

0.446369

ba1cff79d47646c28fa9f3ad61864baa

Schematic of boundary value problem of coarse grained dislocation configurations.

2012.14815

figure-A1.jpg

0.430509

dd685667fb114ffabac9ac9043285844

Fitting parameters as a function of the reference voxel size.

2012.14815

figure-A2.jpg

0.514749

e17ac13939b44a0c98c380a1f60032df

Architecture of LC-Net

2012.14816

LCNet_HR.jpg

0.489987

d84cd25be3b745df9bc97f7475d0d02d

Average probability of error with respect to the learning database size. Notice that the abscissa scale is logarithmic.

2012.14816

Results_curves_acc_recent.jpg

0.465947

4da127ef145e4420ad19fb270e1cf89a

Schematic generic evolution of the test error depending on the dataset size.

2012.14816

law_dataset_size.jpg

0.469509

3567a70227d44793b34e7ae5f0690e97

Schematic generic evolution of the test error depending on the model size.

2012.14816

law_model_size.jpg

0.456698

ae5e4bf9a9ad47f0ac2e74b216086ca1

The summary of the feasibility of the proposed method. The pixel-wise differences are obtained via subtracting the pixel values in the artificial images from the corresponding pixel values in the corresponding reconstructed images. Two, three, and four pixels from each side of the images are excluded from the pixel-wise-difference calculations when the size of the resampling matrix is three-by-three, five-by-five, and seven-by-seven, respectively. As a comparative base, the values in parentheses are the corresponding results when the elements in Equations (10), (11), and (12) are replaced by the values: 1/9, 1/25, and 1/49, respectively.

2012.14817

Table_1.jpg

0.476528

29aad7f9415545c6a9a9e0ef50b93e47

GI I diagrams for W^+ W^- fusion.

2012.14818

fig13_new.jpg

0.47218

65b26d020b5c4a57af836b0a9d158cf4

(A) Hall resistivity measured at several different temperatures as a function of magnetic field with 𝐣∥[100] and 𝐁∥[001]. The inset presents the Hall resistivity in weak magnetic field region. (B) Magnetic field dependent Hall conductivity measured at T= 2 K. The red solid line represents the fit with Equation (1). Quality of the fit was excellent (with R^2>99.99). Inset shows the geometry of measurement.

2012.14819

Fig1new.jpg

0.523852

0d76cb9d9f364d0799fca2131ae7cc15

(A) MR(B) isotherms measured at T=2 K for different angles between current and magnetic field. Inset shows schematically the geometry of measurement. (B) LMR(B) isotherms for several temperatures. (C) Analysis of the transverse magnetoconductivity data obtained at T= 2 K. Blue points - experimental data, green solid line - σ_xx^D(B) calculated with Eq. (<ref>). Grey solid curve represents the WAL contribution derived as described in the text, and red solid line curve shows the fit of Eq. (<ref>). Inset: temperature dependence of the critical field approximated by power law B_c∝T^1.35. (D) Magnetic field dependence of the longitudinal magnetoconductivity for several temperatures. Red solid lines correspond to the fits of Eq. (<ref>). Inset: temperature-dependent chiral coefficient (solid line is a guide for the eye).

2012.14819

Fig2new.jpg

0.437477

2f9231d9e3b240a9b194108ff929f446

Angular dependence of the planar Hall resistivity measured in different external magnetic fields at 2 K (A), and 300 K (B), with experimental geometry sketched in inset to panel (A). Red solid lines represent the fits obtained using Eq. (<ref>). (C) Magnetic field variations of the PHE magnitude observed at several different temperatures. Lines are guides for the eye.

2012.14819

Fig3new.jpg

0.476286

3a1ce03608874775a044fd3c4cbf4628

(A) Angular dependence of the electrical resistivity measured at several temperatures in a magnetic field of 14 T. The inset shows the geometry of measurement. (B) Anisotropic magnetoresistance isotherms calculated from the data presented in (A). (C) Conductivity enhancement derived from the data shown in (A). Red lines represent the results of fitting Eq. (<ref>) to the experimental data. The inset displays the temperature dependence of the narrowing parameter L defined in the text. The solid line is a guide for the eye.

2012.14819

Fig4new.jpg

0.456577

7806999e69154ab586e36f560bb30a26

(A) Band structure and (B) total DOS calculated (including spin-orbit interaction) for the ideal crystal structure of ScPtBi. (C) Band structure and (D) total DOS calculated (including spin-orbit interaction) for the model with Pt occupying the 16e Wyckoff position. The red ellipse in C marks a gap between inverted bands where Dirac points may occur. (E,F) Total DOS for models: (E) with different antisites, and (F) with different vacancies.

2012.14819

Fig5new.jpg

0.44438

395e79ddd75b4b3aaa38cd13a5fd7f68

Energy-dispersive X-ray spectrum (EDS), SEM image (left inset) and photograph (right inset) of ScPtBi single crystal. Chemical composition determined from the spectrum is close to the equiatomic: Sc_33.33Pt_34.03Bi_32.64.

2012.14819

FigS1_EDS.jpg

0.433978

afa11c8cfa90427789f3336036530e96

Temperature dependent specific heat of ScPtBi. Red solid line corresponds to the fit with C=γ T+β T^3 equation. Blue bar corresponds to the specific heat jump expected for a conventional superconductor.

2012.14819

FigS3_HC.jpg

0.483915

4c617d4ad2144cb78cb8148b1d12cf6e

Electronic band structure of ScPtBi for crystal structure models with different antisite defects. Note both hole-like and electron-like bands crossing the Fermi energy E_ F.

2012.14819

FigS8rev.jpg

0.477867

b340829a47f7476e81e9e6f07670798c

Hall resistance measured at several temperatures with electric current 𝐣∥ [100]. Inset shows zoom in weak magnetic field range. Hall resistivity data shown in Fig. 1A of the main text were obtained through anti-symmetrization of data shown here, as: ρ_xy(B)=t(R_xy(B)-R_xy(-B))/2, where t was the thickness of the sample.

2012.14819

ScPtBi_raw_data_Hall.jpg

0.43861

8ff25da609604970a5e5128a51a0701d

The analyzed data.

2012.14820

data.jpg

0.442649

1cf29d1d8d0a48d395ebd5926f5f0b99

The unit root transformations od the analyzed data.

2012.14820

data_trans.jpg

0.411929

90f0e22399e247d691f6f6f5181edcdc

Kinematic distributions for different decay modes of charged kaons at the NA62 experiment <cit.>. Here the normalized differential decay width for a given mode is plotted against the squared missing energy s. The two rare kaon decay modes K^+→π^+ + νν̅/νν have been multiplied by a factor 10^10 for visibility, all other decay modes are treated as background processes. The two shaded areas correspond to the two signal regions of the NA62 experiment.

2012.14825

distributions.jpg

0.477919

556ce11875ac47c7978aef9548967ce0

(Color online) Phase diagrams for the aluminum-nickel Al_(100-x)Ni_x system (left column) and the iron-nickel Fe_(100-x)Ni_x system (right column) <cit.>. The investigated region of the phase diagram in this study is highlighted in rose color.

2012.14826

fig01.jpg

0.56404

20cda57ed2fc428db3d72ddefc5a0619

Concentration dependences of shear viscosity of aluminum-nickel (left column) and iron-nickel (right column) melts at different temperatures: ( ) – results of molecular dynamics simulation; (∘ ∘ ∘) – results of the experiment on viscometry; markers (□ □ □), ( ), ( ), (▪ ▪ ▪), (∙ ∙ ∙) – experimental data taken from Refs. <cit.>, <cit.>, <cit.>, <cit.>, <cit.>, respectively.

2012.14826

fig02.jpg

0.605049

6e93dd9eb6d443c8b68fccadf932c1c4

(Color online) Temperature dependencies of the reduced viscosity for Al_(100-x)Ni_x (left column) and Fe_(100-x)Ni_x melts (right column) at various concentrations of nickel atoms: markers represent the results of molecular dynamics simulation; solid line – universal viscosity curve calculated by the formula (<ref>). Insert: Temperature dependence of the logarithm of viscosity, showing the deviation from Arrhenius's law below the temperature T_A.

2012.14826

fig03.jpg

0.40695

7e08fd21cfc24129a9f148c8a96f3ee9

(Color online) Reduced viscosity η^* as a function of configurational entropy for binary nickel-containing metal melts: markers – the results of molecular dynamics simulation; the dotted line is the Rosenfeld's scale ratio for viscosity with parameters B=1.02 and β=0.65.

2012.14826

fig04.jpg

0.418258

ccd4063c4f3d443cabefde6d722625de

An example dialog from ShARC benchmark dataset <cit.>. At each turn, the machine can give a decision regarding the initial question put up by the user. If the decision is Inquire, the machine will ask a clarification question to help with decision making. The corresponding rule document and the question are marked in the same color in the figure.

2012.14827

example.jpg

0.420559

70971304035c47d7bb26ea4bb4ed6573

Examples selected from the dev set of ShARC where Discern fails but our model succeeds.

2012.14827

examples.jpg

0.465718

da102ff4260d497d97ff5bcb9587c175

The overall structure for our proposed model. With segmented EDUs and tagged relations, the inputs including user initial question, user scenario and dialog history are sent for embedding and graph modeling to make the final decision. If the decision is Inquire, the question generation stage will be activated and use the under-specified span of rule document to generate a follow-up question.

2012.14827

pipeline.jpg

0.536929

7baf04473e054db7a85842f5fedcd405

The figure displays free energy F of the thermalon configuration as a function of the Rényi temperature T_R^(+) for several values of the non-extensive parameter α. We have used L=1, and λ =1.1382. From left to right: α= -0.25 (blue), α =-0.10 (green), α=0.00 (red), α=0.10 (magenta) and α=0.50 (orange). For each value of the non-extensive parameter α, the upper branch beyond the cusp is unphysical where it corresponds to Π^+ = - Π^- solutions while the lower branch is the physical solutions of Π^+=Π^-.

2012.14828

FT.jpg

0.504544

441ce2e5e60a44db919278d081f16529

VLBI image of NGC 315 at 86 GHz.

2012.14831

86_1.jpg

0.385117

186e015198c840d193f2e2741fe75c2e

Core positions as a function of frequency relative to the mm-VLBI cores (43 GHz and 86 GHz).

2012.14831

CoreShift.jpg

0.395806

14665c929f7945ae8787b899e7774faf

Residuals in the innermost jet region for a single power-law fit performed considering the VLBI data only and the entire data set (VLBI+VLA).

2012.14831

NGC315_SingleRes_vla.jpg

0.392072

0ab3b0be72a743b5abee2e2dc0dca39c

Jet structure of NGC 315 from Mega-parsec to sub-parsec scales. The top image was obtained based on VLA data at 1.4 GHz, while the rest of the maps show the VLBI structure at different frequencies (5 GHz, 22 GHz, 43 GHz, 86 GHz) based on some of analyzed the data presented in Sect. 2, Table 1. The complete series of images considered in the article is presented in the Appendix.

2012.14831

multinew.jpg

0.481661

4c0d92a6ea5d4d9d8d2ca64ab91ad65d

Block diagram overview of the multi-modal image registration process utilized.

2012.14833

alignment_process.jpg

0.394197

bb2873c3f7214762becda7042debfde6

Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Car-Nighttime dataset.

2012.14833

car_night_selected.jpg

0.387944

97af855fd4244ba0a30cd9723cbb1aea

Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Car-Daytime dataset.

2012.14833

car_selected.jpg

0.414932

336ffcbd274b41babd31869b99733502

Indicative images from the raw visual and thermal camera frames of the released dataset. Top row: Car-Daytime dataset, second row: Car-Nighttime dataset, third row: Aerial Scout-Mine dataset, bottom row: Aerial Scout-Tunnel dataset.

2012.14833

dataset_raw_images.jpg

0.425407

5b54717fee7f4f4db034a2f7be670bbb

Indicative histogram plots (256 bins) for visual and thermal camera data from the released datasets.

2012.14833

histograms_examples.jpg

0.444436

030d2945f680417da380fe09e7634247

Indicative 32× 32 image patches from visual and thermal data with the centroid of each visual patch being the coordinates of a FAST detection and the centroid of the thermal patch acquired through the identified alignment transformation. One image pair for each of the datasets is considered for this result and only six (randomly selected) detected corners are presented for each image pair. Naturally, datasets that are rich both visually and thermally allow for rather more “matching” multi-modal frames.

2012.14833

local_features_datasets.jpg

0.421678

e5d59bca84c447d787118b62a275851c

Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Aerial Scout-Mine dataset.

2012.14833

trjv_selected.jpg

0.418685

d10f9ae67336480c8f7ce59d9c8c2c30

Indicative results of the considered visual-thermal camera frames alignment process assuming no access to calibration or depth information for the case of the Aerial Scout-Tunnel dataset.

2012.14833

truckee_selected.jpg

0.420594

6cbb205158b04275b1e6f09781f279cf

Optimal and maximum EH time

2012.14834

EHopt_vs_max_v2.jpg

0.383272

41564076ba1f4314b9441839163a2cb2

Optimal and maximum power

2012.14834

Popt_vs_fix_v2.jpg

0.419032

186ac6c791854ad3ad35d755fbdfecfc

Different ToA allocations

2012.14834

SFalloc_v2.jpg