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
|
|---|---|---|---|---|
0.388954 | 6a8ddd1391ce4b3d87cbdbb5871e6cef | FWHM of main peak for both sample A (filled red bullets) and sample B (open blue squares), along with FWHM of side peak (open blue squares) and overall width (open green triangles) of sample B. The FWHM begins to broaden rapidly below ∼ 15 K and this region is highlighted in light blue. | 2012.12340 | Wang_Fig4.jpg |
0.402391 | 3c31751c3cb54e559d4e11ebe47887c2 | The normalized recovery curve M(t) observed for sample A at select temperatures. The solid lines in panel a) show the respective stretch fit results, and the solid lines in panel b) show the ILT fit results. The stretch fit fails to capture all the components when 1/T_1 becomes highly distributed while the ILT fits the data nicely at all temperatures. | 2012.12340 | Wang_Fig5.jpg |
0.411107 | df93131de0554cdab372d5f9ebe7fe52 | a) Spin lattice relaxation rate 1/T_1^stretch estimated from stretch fit at the main peak of sample A (filled red bullets) and the main (open red circles) and side (open blue squares) peak of sample B. ILT COG of sample A main peak (black X's) and sample B main (red +'s) and side (open blue diamonds) peaks seen in Fig.<ref> are also plotted. Note that below 10K in sample B, the rapidly broadening main peak is overshadowed by the side peak, and thus 1/T_1 is measured at the peak of the overall NMR line shape. The inset shows 1/T_1^stretch of the sample A main peak zoomed to better display the sharp peak at 9± 1 K. b) The stretch fit parameter β for samples A and B. For sample A main peak, we found β = 1 within experimental uncertainties down to 20 K, implying that there is no distribution in 1/T_1. | 2012.12340 | Wang_Fig6.jpg |
0.500344 | 13bc0d0198724ab2b7af8f85aad6f8f2 | Cascaded ILT density distributions P(1/T_1) for a) Sample A main peak and b) Sample B side peak 1/T_1 relaxation measurements. The distributions are normalized to have equal areas. The temperatures corresponding to each curve are labeled on the left side of the plot. The open diamonds indicate the calculated COG of the ILT distribution. The black dotted lines below 13 K are double Gaussian fits used to deduce the relative fraction of the extra-slow component. The P(1/T_1) results deduced for M(t) measured at the main peak of sample B (not shown) is similar to the side peak results in panel b), except that the relative intensity of the slower peak of P(1/T_1) is much smaller. | 2012.12340 | Wang_Fig7.jpg |
0.490698 | e1c028b3a5424d21aefa08a166199079 | Contour plot of the density distribution function P(1/T_1) deduced from M(t) by ILT for a) main peak of sample A and b) side peak of sample B. Note that the faster component in panel b) shows nearly identical behavior as the result in panel a. This means that the fast components observed for the side peak of sample B correspond to the intrinsic behaviour of the material arising from the superposed signals from the main peak. | 2012.12340 | Wang_Fig8.jpg |
0.386648 | 6c255386391748999bd7622fefc12e82 | Fraction of extra-slow component f_s in P(1/T_1) below 13 K for sample A. | 2012.12340 | Wang_Fig9.jpg |
0.403382 | 116221008b8c4491a4be6209b298cc07 | Large N phase diagram of 𝒩=4 SYM theory on S_R^3 for λ→ 0 as a function of inverse temperature β and grading θ. The theory is in a confined phase in the white region, and is partially confined in the colored regions. In between θ = π/3 and θ = 5π/3 center symmetry is only partially broken, with the unbroken subgroup of ℤ_N indicated by the color scheme in the legend. The theory confines at all β/R when θ = π/3,π,5π/6. | 2012.12341 | rainbowPlot2.jpg |
0.445758 | 322c308c06de4eb38da1a3b40e8374f9 | Large N phase diagram of 𝒩=4 SYM theory on S_R^3 as a function of inverse temperature β and grading θ at zero and infinite 't Hooft coupling λ. The grading θ interpolates between the thermal and (-1)^F graded partition functionsf. At λ = 0 the theory has a deconfinement phase transition along the black curve, and has at least partially broken center symmetry everywhere in the colored regions. The analogous phase transition curve due to known bulk solutions at large coupling λ→∞ is shown in blue, with center symmetry broken everywhere in the hatched region. Note that the deconfined region at large-coupling is a subset of the deconfined region at zero coupling. | 2012.12341 | weak_strong_plot2.jpg |
0.410396 | 26410a0460da4b3dab7407a38d11f830 | Lennard-Jones potential obtained by deflection angle and exact values plotted for data in the well region for L=2500. | 2012.12342 | LJ407pt.jpg |
0.999988 | 444e147e5b764866a617bce9437c9844 | Relevant parameterizations for proof of <Ref>, Region 𝒮_1 | 2012.12347 | rank_1_fig.jpg |
0.999988 | 35211687d3bf4446b268e972f0279664 | Relevant parameterizations for proof of <Ref>, region 𝒮_2 | 2012.12347 | rank_3_fig.jpg |
0.43296 | 1db8b7af380e4c3d881df1889934f3f1 | ActionBert model architecture. On a high-level, the ActionBert model takes a pair of UIs, represented by their text, vision, positional and segment embedding as input. Three new UI-specific tasks, i.e. link component prediction, consecutive UI prediction and masked text prediction, are defined to pre-train ActionBert on large-scale UI sequences with user actions. | 2012.12350 | actionbert-archi.jpg |
0.428985 | 4e350865ce3f4cf386603df963a5df46 | Examples where low-level appearance does not reflect the semantics of UI components without context. Shown below are their semantics on the source screen. | 2012.12350 | icon-semantics.jpg |
0.379577 | 9110df1fe8ac4705b1040f10c2c7c62e | Referring expression UI component retrieval task. | 2012.12350 | ref-exp-retrieval.jpg |
0.411451 | f9119fc804bb4e39bcc871617d9c584c | An example of similar UI component retrieval. | 2012.12350 | sim-comp-retrieval-web.jpg |
0.406468 | 3ac068a9bffe4be59ca1b9d801f2a5eb | An example of user actions on UIs. The user clicks on the “tick” button in screen 1 and jumps to the boarding-pass page, screen 2. The semantics of the button, online check-in, can be inferred from components on the homepage (e.g., images of plane, airline name) and the components on the next UI (e.g., QR code, passenger information). | 2012.12350 | toy-example.jpg |
0.4514 | 9b7989adb4ef4d52a795143c2b2815de | BERT and transformer encoder blocks. | 2012.12350 | transformer-block.jpg |
0.379352 | 8a7bf54e08da44aabb1996a5da2d936d | An example of leaf nodes in a view hierarchy. The view hierarchy provides useful semantic information (marked red) for machines to understand the UI. | 2012.12350 | vh.jpg |
0.410843 | 98a5feef9cc94f99aca1b83ce5a3aee8 | Two examples of applying masked language modelling on the counting dataset with LXMERT. | 2012.12352 | MLM_example.jpg |
0.411843 | d43ef282a75b49a0968527fdabd88585 | Percentage of numerals in the counting data. Outer circle: standard split, inner circle: hard split. | 2012.12352 | percentages_hard_vs_standard_white_space.jpg |
0.447587 | bf84827ddb8648238a6d2d89e5704c8d | How many question from Visual7W dataset. | 2012.12352 | visual7w_figure_compact.jpg |
0.526734 | 7280f122794e4b3b8344c3eb1b792519 | (a) Time averaged phase difference between the particle's oscillation and the injected signal when the detuning δ is scanned over the locking range. Black line shows theoretical fits using Eq. (<ref>) and time averaged Eq. (<ref>) with 2Ω_m/2π=270 Hz. (b) Standard deviation of the phase difference σ(Δϕ(t)) measured for the same frequency range. The value π/√(3) rad represents the standard deviation under the limit where the phase difference distributes uniformly over the range [-π,π]. Rapid drop of this quantity to near zero inside the locking range is an evidence of a fairly fixed phase relationship between the particle and the external modulation. Black line shows theoretical fits using standard deviations of Eq. (<ref>) and Eq. (<ref>) with 2Ω_m/2π=270 Hz. (c) Increase in the RMS displacement when the particle is driven by the injected optical force. The locking range (with the width of 2Ω_m/2π) is shown by the shaded region. | 2012.12354 | injlock.jpg |
0.395028 | 793e73d8fe6f4c73b910bcc2c777404b | Experimental setup for injection locking. AOM – acousto-optic modulator, EOM – electro-optic modulator, AFBE – analog feedback electronics, BPs – three bandpass filters, 2Ω_0 – three frequency doublers, ΔΦ_c – three phase shifters, G_c gain-tunable summing amplifier, DLIA – digital lock-in amplifier, PLL – phase-locked loop, LO – local oscillator, BPDs – balanced photodiodes and their associated optics for split-detecting the probing beam in three orthogonal directions, DBS – dichroic beam splitter, PBS – polarizing beam splitter. The trapping beam is feedback modulated via the EOM, while the probe beam is separately controlled by an AOM. Forward-scattered probe beam is collected in a network of BPDs from which the 3–axis parametric feedback cooling and the 1–axis injection locking signals are separately derived. The injection laser is intensity modulated with an additional AOM driven by a LO that generates the harmonic modulating signal. | 2012.12354 | injlock_expsetup.jpg |
0.580812 | 3c338bb221b14f01b4bd7c9ef8ca1c82 | Characteristic phase pull-in effect in the injection locking of a levitated nanoparticle. Red shift in the particle's oscillation frequency can be seen in the shaded area when the detuning of the injected signal is swept from far below resonance to inside the locking range. Enhancement of the oscillation amplitude along with a significant reduction in its linewidth can be clearly seen for the locked oscillation. Red: experimental data, black: Lorentzian fit to data. | 2012.12354 | phase_pulling.jpg |
0.508135 | c08c0c14e9ab4f54a32874ae87ff6d5d | Average Running Time of Gurobi and <Ref> with Different N in Experiment 1. The plot is shown on a logarithmic scale. | 2012.12356 | alg1_gurobi_color.jpg |
0.402527 | f196a16f63cc474abda68628ee9a15fb | Decision Boundary for Our Framework | 2012.12356 | framework_20.jpg |
0.407626 | a029c554825e4c23b65c105b11aab4e0 | Decision Boundary for SVM | 2012.12356 | scatter_one_2.jpg |
0.427224 | ae5260b96cf14313abf251e8f408afc0 | Decision Boundary for GSVMF | 2012.12356 | scatter_two_2.jpg |
0.390847 | 903eae9fd65544019d6b614e4ba6fced | Decision Boundary for Vanilla Classifier | 2012.12356 | vanilla_20.jpg |
0.413009 | cf1e7759b5e843ae9598c8cf8dd3695e | Our direct pose estimation pipeline. Given a single RGB input, we jointly estimate scene coordinates and depth. Using the known camera focal length, depth is converted to camera frame coordinates by unprojecting using π^-1. The second stage of our network computes a correspondence weight for each point pair. These weights are used for point normalization and final pose regression using weighted rigid alignment to estimate camera pose. | 2012.12360 | finalarch.jpg |
0.452245 | 500fe4507d63430983bf89cbde202814 | Cumulative histograms of error for all scenes from 7Scenes for several methods, truncated to 1 meter and 25 degrees error. | 2012.12360 | histogrammain.jpg |
0.444062 | faa8b9facfa24bcdab687b70154b08d7 | Individual cumulative histograms of error for each scene from 7Scenes for several methods, truncated to 1 meter and 25 degrees error. | 2012.12360 | histogramsall.jpg |
0.471327 | 8068b3b8f89649639c2db38b4cb37a5b | The component ratio Γ_1^1 of the first excited state for the (d)→ (-) flow. The red dashed line marks the predicted value Γ_1^1=-77/3≈ -25.6. | 2012.12361 | l20dtominusG11_v2.jpg |
0.402696 | 0dc8fbb6553f4d0b97e98ea7da028ea3 | Ablation study of Transformer-Guided-Attn when dropping each role individually. | 2012.12366 | barplotperf.jpg |
0.492721 | 369d690c2c0745219fa07ff40b17d466 | Scaled-dot product with role mask or padding mask. | 2012.12366 | mask_scale.jpg |
0.396592 | c147ac040b8c460f9c75028b4a0689ae | Comparison of PSSG (green), FSG (red), GSSG (black) and standard SGD (blue) on misclassification in testing as a function of cumulative samples over HIV-1 dataset with ρ=0.1 and log-scale x-axis. | 2012.12367 | hiv_cumsamples_vs_test.jpg |
0.508393 | de6ffcca769d4add8ca55bd8425d2dbb | Comparison of PSSG (green), FSG (red), GSSG (black) and standard SGD (blue) on misclassification in testing as a function of cumulative samples over RCV1 dataset with ρ=0.1 and log-scale x-axis. | 2012.12367 | reuters_cumsamples-vs-test.jpg |
0.461165 | ea882801fae54344a0df64d03eff6d72 | CIFAR100: 1st iteration (up to 2.25k randomly selected samples per class). Even with 100 classes in total and only 450 train images per class, is able to select a diverse range of task-related images from the pool of 80MTI, which contains mostly unrelated samples. | 2012.12372 | cifar100_overview_ODST_iteration_0.jpg |
0.48524 | 15bf0e7181724993b5768141ad6d760c | CIFAR100: 2nd iteration (up to 4.5k randomly selected samples per class). With , class representations remain stable as the student model is not learning to classify unrelated images with high confidence. | 2012.12372 | cifar100_overview_ODST_iteration_1.jpg |
0.42205 | ebc2386ade944d369e1ce035d06c063e | CIFAR100: 3nd iteration (up to 6.75k randomly selected samples per class). Taking into account the difficulty of the task shows great stability even in the third iteration, however sample quality depends on the class and some classes like "otter" and "worm" show a larger false positive rate. For the class otter there are only two images above the thresholds. | 2012.12372 | cifar100_overview_ODST_iteration_2.jpg |
0.405287 | b50ace182a714b94aed72caac13172ae | CIFAR100: 1st iteration (up to 2.25k randomly selected samples per class). Disregarding few outlier classes, the baseline with OD thresholding is even better than and has a good sample quality even though clearly worse than . | 2012.12372 | cifar100_overview_ST-OT_iteration_0.jpg |
0.413064 | ff66bf868c5246628e5d6813fd8be590 | CIFAR100: 2nd iteration (up to 4.5k randomly selected samples per class). While thresholding slightly stabilizes sample selection, without OD-aware training, the model starts to include bad samples, for example text as "raccoon" or graphics as "clock". | 2012.12372 | cifar100_overview_ST-OT_iteration_1.jpg |
0.432719 | e9ecaee790414d4f80539035b2abceda | CIFAR100: 3nd iteration (up to 6.75k randomly selected samples per class). In the third iteration, we notice more and more mistakes for a large number of classes. This highlights the challenges of self-training. Once a teacher model has learned a wrong representation, it will always pass on wrong information to the student, which results in even worse sample selection. | 2012.12372 | cifar100_overview_ST-OT_iteration_2.jpg |
0.377403 | 1af822238c824e32882b1dc1f6fc5541 | CIFAR100: 1st iteration (up to 2.25k randomly selected samples per class). Due to the relatively small number of additional samples per class in the first iteration, even the baseline is able to select comparatively good sampls. | 2012.12372 | cifar100_overview_ST_iteration_0.jpg |
0.415161 | 17a3fac410d044d08e192dbe96756531 | CIFAR100: 2nd iteration (up to 4.5k randomly selected samples per class). With the 2nd iteration we can notice an increase in task unrelated samples across various classes. As most images are classified with very large confidence, the additional samples are not well suited for training another student. | 2012.12372 | cifar100_overview_ST_iteration_1.jpg |
0.388494 | a8f22d42d950440fab61bf4123d7796b | CIFAR100: 3nd iteration (up to 6.75k randomly selected samples per class). For most classes, standard self-training breaks down at this point. Notice that while the first model was selecting mostly good images and had an intact class representations, this one has learned systematically wrong representations, labeling food as "crab" with high confidence. | 2012.12372 | cifar100_overview_ST_iteration_2.jpg |
0.466408 | 34dd78ee68114e059f933fe7c43a6af9 | CIFAR100: Histogram plot showing the number of accepted samples per class for , and over three iterations. Each methods adds up to 2250 samples per class in the first, 4500 in the second and 6750 in the third iteration (note that the number of labeled examples is 450 per class). We see that all methods add the full number of samples for the majority of classes but has the most conservative but also the best sample selection. | 2012.12372 | cifar100_samples_hist.jpg |
0.395733 | 72777bb51a314b3ab293b1b67a9d7e87 | 1. iteration - max. 25k samples per class | 2012.12372 | cifar10_overview_ODST_iteration_0.jpg |
0.358236 | d9e2e4a8b0e74a9cbe8109a08655174d | 2. iteration - max. 50k samples per class | 2012.12372 | cifar10_overview_ODST_iteration_1.jpg |
0.382006 | dbf54142582a4dfb9ae90cd670c89cc6 | - 2nd iteration - 50k samples per class | 2012.12372 | cifar10_overview_ODST_iteration_1_10_marked.jpg |
0.388086 | 15f1d90b98c64bc1b220c3f8a4763da6 | 3. iteration - max. 75k samples per class | 2012.12372 | cifar10_overview_ODST_iteration_2.jpg |
0.431707 | c08da9d2fde045f385049ec0d0a1fb3c | 1. iteration - max. 25k samples per class | 2012.12372 | cifar10_overview_ST-OT_iteration_0.jpg |
0.410296 | 0876a1ba54a04c3fa554e3286635590b | 2. iteration - max. 50k samples per class | 2012.12372 | cifar10_overview_ST-OT_iteration_1.jpg |
0.398755 | 3943d8c382f046769af22f37405f93a1 | 3. iteration - max. 75k samples per class | 2012.12372 | cifar10_overview_ST-OT_iteration_2.jpg |
0.375568 | c8a37bca777d44a5a7dd55f36ea0287e | 1. iteration - max. 25k samples per class | 2012.12372 | cifar10_overview_ST_iteration_0.jpg |
0.458005 | 0fefcddc32c1409992000806654c3696 | 2. iteration - max. 50k samples per class | 2012.12372 | cifar10_overview_ST_iteration_1.jpg |
0.442901 | 2a98ea09885b40789e47f9a1e0c3fac8 | 3. iteration - max. 75k samples per class | 2012.12372 | cifar10_overview_ST_iteration_2.jpg |
0.399194 | 4cc67a2716e145f382ab09a448495020 | 500k-TI <cit.> - 50k samples per class | 2012.12372 | cifar10_overview_carmon_marked.jpg |
0.385401 | f0c14ed4b38b47ceb8cbf35225bdec1a | Random Forest AUROC as a function of lookahead window size, N. Error bars indicate the standard deviation of the cross-validated error across folds. | 2012.12373 | auroc_both_rfs.jpg |
0.517641 | f1f3fa8c765747b9be01239046e6f179 | The CDF in Figure <ref> split across infant failures (occurring at age ≤ 90 days) and mature failures (occurring at age > 90 days). | 2012.12373 | cdf_pe_count_split.jpg |
0.480162 | 607c34a493134f9c8d32507acae6c018 | The distribution P/E cycle counts of failed drives (solid) and the proportion of drives that fail (dashed) at a given P/E level, binned in increments of 250 cycles. | 2012.12373 | prob_fail_bype.jpg |
0.406234 | f4278c5d8c97485ca75b22d6b971ccb4 | Quartiles of the daily write intensity per month of drive age. The line plot shows the median write intensity for each month. The 1st and 3rd quartiles are shown as the boundaries of the shaded area. | 2012.12373 | writerates.jpg |
0.440468 | 4814d106c6aa4ad08cbc7dd43eeed93d | The Hoffman-Singleton graph. The 10 red vertices form a graphical design which integrate the first 49 eigenvectors of the graph Laplacian in a suitable basis. The induced subgraph of this design is 3-regular. | 2012.12376 | Hoffman.jpg |
0.393032 | 9a132477fee447969a32f0476d31bb3a | A multi-color HST image of Haro 11, one of the nearest BCGs (and Lyman break analogues). Its irregular structure, of total diameter ∼2 kpc, is probably the result of a galaxy merger, which also induced the SFR ∼20 M_⊙ yr^-1. The galaxy is a weak Lyman continuum leaker (Section <ref>). It is a emitter in some of the star-forming knots and a absorber elsewhere. Credit: ESO/ESA/Hubble and NASA (https://www.eso.org/public/images/). | 2012.12378 | 1024px-Frenzied_Star_Birth_in_Haro_11_crop2.jpg |
0.445409 | c4fffecd7bf24112bab171865ac1dcee | Starburst region 30 Doradus (also known as Tarantula nebula) in the Large Magellanic Cloud. The multi-colour UV and optical image was obtained using the Hubble Space Telescope and shows the central concentration of young stars (called R136) in the NGC 2070 star cluster, which is situated at the core of the nebula. R136 contains several O stars and Wolf-Rayet stars and produces most of the energy that makes the 30 Doradus visible. Credit: NASA/ESA. | 2012.12378 | 30Dor.jpg |
0.439294 | a28c7e6174924b5f8361c12607f57902 | UV absorption lines of the galaxy SDSS-J030321.41-075923.2. The major lines originating both in the galaxy and in the Milky Way have been labeled. Source: HST archive. | 2012.12378 | GP0303_LIS_FUVB.jpg |
0.618811 | 7df0d587fb8c44be80f8e27d3931c61f | Green Pea optical SDSS spectrum (left); optical SDSS image – unresolved (middle); and near-UV HST image – obtained for Cosmic Spectrograph spectral acquisition (right). All three panels are for the galaxy SDSS J092532.36+140313.0 at z∼0.3. The diameter 0.5” corresponds to ∼2 kpc. Credit: SDSS and HST archives. | 2012.12378 | GP_spec_imgSDSS_imgHST_0925_violet2.jpg |
0.425979 | c4a671d5675a4b8e9559dece788379b6 | spectrum of Mrk 259, a nearby irregular starburst galaxy from the LARS sample. The P-Cygni profile with its asymmetric red peak and the absorption trough blueward from the line center (1215.67 Å, dashed vertical line) result from radiative transfer. Credit: HST archive. | 2012.12378 | Lya_spectrum.jpg |
0.455121 | fd0efbf936304d2fb17429fd97624216 | The Simple-QE architecture, with and without adding numeric features corresponding to the original complex sentence (Original Features) and the system output (System Features). R denotes a regressor layer. | 2012.12382 | fig.jpg |
0.463096 | 1f1ba566417b455281728a1bdad5408e | SGD with large step size. | 2012.12383 | LargeA.jpg |
0.455382 | d3b14a8fb2e84d68abe495f1129269f8 | Local controllers. | 2012.12383 | agent_convergence.jpg |
0.43986 | ee592ae7410a4498b06490be3ade5586 | Decaying excitation noise. | 2012.12383 | compare_decrease_fix_noise.jpg |
0.454772 | b96592dc4bbb451ca9c517e5bd92c6f8 | ST-Q with different N. | 2012.12383 | compare_differentN.jpg |
0.465189 | 23af4a37da644b30934a825367b4d2d3 | Convergence of K̂-K. | 2012.12383 | compare_error_fix_gst.jpg |
0.430016 | 9866377f65124f1dad5d286739a09c48 | Global controllers. | 2012.12383 | compare_gd_new.jpg |
0.456998 | 420efd54ddff4d3a85acfbdddf0dc215 | Illustration of the result in (<ref>). | 2012.12383 | lemma2figure.jpg |
0.453712 | 7622162b0a344c6f81fdeb106f69acf6 | SGD with small step size. | 2012.12383 | smallA.jpg |
0.481769 | 70976fe9d2224892aacfd22553edab9e | ST-Q with N=50. | 2012.12383 | smallN.jpg |
0.454042 | 4572cb266b484d6581c4e61aafce64b7 | Illustration of the time scales in Algorithms (<ref>) and (<ref>): t denotes the time steps, and t_q denotes the time instance for the q-th policy iteration. Policy evaluation is carried out N times within each policy iteration. | 2012.12383 | timescale.jpg |
0.526856 | 5c3dbd8855ea4a2d88c9e3a8cff69ee0 | An example of the communication graph (dashed lines) and the interconnection graph (solid lines). All the agents are interconnected with physical connection, while the direct communication channel does not exist between neither Agent 4 and Agent 2 nor Agent 4 and Agent 1. | 2012.12383 | twograph.jpg |
0.584428 | 269d1eb78bb640dbac2c4fd5bd2d807d | The white and black dots represent samples with different model outputs. +s represent the virtual samples generated by the mixup operation. | 2012.12384 | mixup_w_ranges.jpg |
0.582674 | 38e47994743f4ca38b0787bac08aa919 | The white and black dots represent samples with different labels. | 2012.12384 | orig_points_w_ranges.jpg |
0.499593 | 92070e9014954ae998881b7dbdbf64e0 | If Γ (orange) is any circumconic of ABC, ℰ_D the pedal-circle (purple) w.r. to D and γ_D, (doted orange hyperbola) is the negative-pedal of ℰ_D, then (Γ, γ_D) form a poristic pair for n=3. All triangles (e.g. green and blue) inscribed in Γ and circumscribed to γ_D share the same pedal-circle ℰ_D. The arc D_1D_2 of Γ, situated inside the conic γ_D is infertile: it cannot contain vertices of such triangles, since there is no tangent line to γ_D passing through these points. | 2012.12385 | 2041_Euler_podal_bis.jpg |
0.425972 | 9c2a793d016d4fd7a3ed1398b56ec18a | The negative-pedal triangle of ABC w.r. to pedal-point D is A' B' C', whose sides are the perpendiculars in A, B, and C to DA,DB and DC respectively. Its circumcircle, 𝒞_D (orange circle) is the negative-pedal circle of ABC w.r. to D. Γ_D, the i-conic of A' B' C' focused in D (orange elipse) is the negative-pedal of circumcircle 𝒞. (𝒞_D,Γ_D) form a poristic pair for n=3. By Poncelet's porism, two triangles inscribed in 𝒞 (blue and green) share the same negative-pedal circle 𝒞_D if and only if the sides of their negative-pedal triangles tangents the same conic Γ_D. The arc L_1 L_2 of 𝒞_D located inside the in-ellipse, cannot contain vertices of A' B' C' (there is no tangent from on L_1 L_2 to Γ_D). Circle 𝒞 is pedal-circle of A'B'C' w.r. to pedal-point D. | 2012.12385 | 2043_Euler_podal.jpg |
0.465527 | 4149cfdf837d447383773420c591ebf6 | A_p, B_p, C_p are, respectively, the poles of BC, AC, and AB w.r. to ℐ (doted black).ℰ'_D (doted purple), the circumcircle of A_p B_p C_p is the inverse of the pedal-circle ℰ_D (purple). The negative-pedal A' B' C' (blue) and A_p B_p C_p are homothetic and so are their circumcircles 𝒞_D and ℰ'_D. Therefore 𝒞_D is fixed iff ℰ'_D is fixed iff ℰ_D is fixed. | 2012.12385 | 2047_Euler_podal_completo.jpg |
0.422162 | 4692912d1db14759aef763967b3cd0eb | The projections of pedal-point D on the sides of ABC are on pedal-circle ℰ_D (purple) iff lines AB, AC, BC tangent γ_D the negative-pedal of circumcircle 𝒞 (solid green hyperbola). The vertices A,B,C lie outside the hyperbola, hence the arc D_1D_2 situated inside the i-conic γ_D is infertile. ABC preserves pedal-circle iff the sides of pedal A' B' C' (solid blue) tangent the negative-pedal of the circumcircle 𝒞, the in-ellipse Γ_D (blue). ℰ_D is fixed iff 𝒞_𝒟, the circumcircle of negative-pedal triangle A'B'C' is fixed. | 2012.12385 | 2049_Euler_podal_iconicas.jpg |
0.450424 | 91d369a0474d4635ad8dae13ab6159d3 | ℰ_D the pedal-circle of ABC (purple) coincides with the inverse of 𝒞_p, the circumcircle of the polar triangle A_p B_p C_p (blue).Similarly, the pedal of the polar triangle (solid violet) is the inverse of 𝒞, the circumcircle of the original. The perpendiculars from D to the sides of ABC intercepts them on points located on pedal-circle ℰ_D and pass through the vertices of A_p B_p C_p. | 2012.12385 | 2050_lema_poristica_2_infertil.jpg |
0.366888 | 4b0e9cdb401c4935aa2292a0d716a6f5 | 𝒞 is the circumcircle of ABC (violet circle) and 𝒞_p (green circle), the circumcircle of the polar triangle A_p B_p C_p. 𝒞_p is fixed, does not depend on point A on 𝒞. The dual of 𝒞 w.r. to inversion circle centred in D is Γ_D (violet hyperbola) and the dual of 𝒞_p is γ_D (green ellipse). Then γ_D is the i-conic of ABC and Γ_D is the i-conic of A_p B_p C_p. There are three poristic pairs: i) (𝒞,γ_D); the fertile arc D_1D_2 of 𝒞 lies outside the i-conic; ii) (𝒞_p,Γ_D); the fertile arc L_1L_2 of 𝒞_p lies outside the i-conic Γ_D; iii) the polar-poristic (𝒞,𝒞_p); (shown ABC and its polar A_p B_p C_p). The poristic pairs (𝒞,γ_D) and (𝒞_p,Γ_D) are dual images of each other. | 2012.12385 | 2052_lema_poristica_2_infertil.jpg |
0.487067 | f01b250d1b1a438aa9dbd802d6dbe0d1 | A general CDPR with a genetic auxiliary counterbalance system. | 2012.12387 | J8_Figure_1.jpg |
0.554216 | 8fcb62fa937a4f978285555d48ec32bc | Possible robot configurations and coordinates assignment. (a) Mounting a single pulley on the top side of the platform. (b) Mounting two pulleys on the top side of the platform. (c) Mounting two pulleys on the bottom side of the platform. (d) Mounting two pulleys on the top side of the platform and one fixed pulley on the static frame. | 2012.12387 | J8_Figure_2.jpg |
0.457686 | 9f2b7ce97e8b422eab8d57d218ccb850 | Area of reachable workspace versus w_p. | 2012.12387 | J8_Figure_3.jpg |
0.408689 | 6ce7bb9e7a3f45e69bad06ccb5626a75 | Area of reachable workspace versus T_5. | 2012.12387 | J8_Figure_4.jpg |
0.451676 | 0ffcabae6a064020b5e080db92d2c378 | Desired workspace versus reachable workspace. | 2012.12387 | J8_Figure_5.jpg |
0.431091 | 4912c57859c94e1ba7643f9d987c8a52 | Area of reachable workspace of different robot configurations versus T_5. (a) Responses of the configuration <ref> (b). (b) Responses of the configuration <ref> (c). (c) Responses of the configuration <ref> (d). | 2012.12387 | J8_Figure_6.jpg |
0.495668 | 76034f98973a48ff8a7fa1c85b955796 | Desired workspace versus reachable workspace of different robot configurations. (a) Responses of the configuration <ref> (b). (b) Responses of the configuration <ref> (c). (c) Responses of the configuration <ref> (d). | 2012.12387 | J8_Figure_7.jpg |
0.578731 | 7084f12b2afe4b9fb67caa6592a3d917 | Desired workspace versus reachable workspace with the active control of T_5. (a) T_5 = 0-5000 N. (b) T_5 = 0-10000 N. (c) T_5 = 0-20000 N. (d) T_5 = 0-26000 N. | 2012.12387 | J8_Figure_8.jpg |
0.488275 | b47f5a6a8b5c43d49262cdc00e9e4153 | Real world COVID-19 surveillance data for Scotland, from 8th March–30th November. This data is publicly available online at <cit.>. (a) Reported cases; (b) number of tests; (c) case intensity (the number of cases scaled by the number of tests with the baseline of 1,000 tests per day); (d) current number of hospitalised patients; (e) current number of patients in critical care, and (d) daily deaths. | 2012.12390 | Figure_1.jpg |
0.452677 | edfadf3945a1485098326dbbbf0282ee | SEIR-HCD model flow-chart, illustrating the model's structure. This is the assumed clinical pathway of COVID-19 infection. | 2012.12390 | Figure_2.jpg |
0.468452 | 8dc35aba764f48e4a20a97a91b5a179d | Simulation results, produced by fitting the SEIR-HCD to all four Scottish surveillance data streams, from 8th March–30th November. Median prediction is shown by the black line, alongside the corresponding 95% credible intervals (CI) (light grey) and 50% CI (dark grey). The real-world data is shown as red points. For convenience of view, the y-axes use a logarithmic scale. (a) Reported cases; (b) hospitalisation; (c) critical care; and (d) daily deaths. | 2012.12390 | Figure_3.jpg |
0.427516 | 740bc4f97864435d99a18421ad432bde | Estimated effective reproductive number R, for COVID-19 in Scotland, for the period from 8th March–30th November. Median prediction is shown by the black line, alongside the corresponding 95% CI (light grey) and 50% CI (dark grey). Horizontal dashed blue line corresponds to R_0=1. | 2012.12390 | Figure_4.jpg |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.