Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
1,049
datasets available to search
ShareScore release 0.9.0
Dataset results
1,049 results for “robustness”
Image 2 in Robust Trapdoor Tarantula Haploclastus validus Pocock, 1899: notes on taxonomy, distribution and natural history (Araneae: Theraphosidae: Thrigmopoeinae)
Image 2. Haploclastus validus male palpal bulb. Not to scale
Image 1 in Robust Trapdoor Tarantula Haploclastus validus Pocock, 1899: notes on taxonomy, distribution and natural history (Araneae: Theraphosidae: Thrigmopoeinae)
Image 1. Haploclastus validus male from Matheran (Raighad District, Maharashtra)
Data from: Quartet-based computations of internode certainty provide robust measures of phylogenetic incongruence
Incongruence, or topological conflict, is prevalent in genome-scale data sets. Internode Certainty (IC) and related measures were recently introduced to explicitly quantify the level of incongruence of a given internal branch among a set of phylogenetic trees and complement regular branch support measures (e.g., bootstrap, posterior probability) that instead assess the statistical confidence of inference. Since most phylogenomic studies contain data partitions (e.g., genes) with missing taxa and IC scores stem from the frequencies of bipartitions (or splits) on a set of trees, IC score calculation typically requires adjusting the frequencies of bipartitions from these partial gene trees. However, when the proportion of missing taxa is high, the scores yielded by current approaches that adjust bipartition frequencies in partial gene trees differ substantially from each other and tend to be overestimates. To overcome these issues, we developed three new IC measures based on the frequencies of quartets, which naturally apply to both complete and partial trees. Comparison of our new quartet-based measures to previous bipartition-based measures on simulated data shows that: 1) on complete data sets, both quartet-based and bipartition-based measures yield very similar IC scores; 2) IC scores of quartet-based measures on a given data set with and without missing taxa are more similar than the scores of bipartition-based measures; and 3) quartet-based measures are more robust to the absence of phylogenetic signal and errors in phylogenetic inference than bipartition-based measures. Additionally, the analysis of an empirical mammalian phylogenomic data set using our quartet-based measures reveals the presence of substantial levels of incongruence for numerous internal branches. An efficient open-source implementation of these quartet-based measures is freely available in the program QuartetScores (https://github.com/lutteropp/QuartetScores).
Data from: Parameterizing the robust design in the BUGS language: lifetime carry‐over effects of environmental conditions during growth on a long‐lived bird
1. Since the initial development of the robust design, this capture‐recapture model structure has been modified to estimate temporary emigration, and expanded to include auxiliary information such as band recovery and live resight data using maximum likelihood approaches. These developments have allowed investigators to separately assess individual and group effects on true survival, site fidelity, and temporary emigration. Additionally, recent advances in the BUGS language have allowed researchers to develop increasingly complex, user‐specified models in Bayesian frameworks. 2. The robust design has rarely been implemented in the BUGS language, and previous attempts to parameterize the robust design in BUGS exhibited strong bias in estimates of temporary emigration rates. Given the limitations of current parameterizations of the robust design in Bayesian frameworks, and our research objectives, we have developed a parameterization of the robust design in the BUGS language that produces unbiased estimates of all model parameters. 3. We use this novel model structure to examine lifetime carry‐over effects of environmental conditions during early life on annual breeding probabilities of Pacific black brent (Branta bernicla nigricans) breeding on the Yukon‐Kuskokwim River Delta in western Alaska. We found that individuals that were more structurally developed as goslings bred at increased rates as adults (β = 0.14, f = 0.94), with no effect on adult survival (β = 0.01, f = 0.62). Additionally, we provide evidence for long‐term declines in apparent survival of breeding adult females at the population level (β = ‐0.01, f = 0.90). 4. This novel model structure can be easily expanded (Gibson et al., in review), and has important implications for population modelling at broad scales, where we apply it to a declining population of Pacific black brent. Given long‐term declines in gosling growth on the Yukon‐Kuskokwim Delta, we predict future declines in population trajectories as a result of lifetime carry‐over effects of environmental conditions during growth on adult fecundity, and long‐term declines in adult survival.
Classifying Handedness in Chiral Nanomaterials Using Label Noise-Robust Deep Learning
<p>Images of individual Te chiral nanoparticles and labels of their handedness for classifier training.</p>
Data from: Mechanical feedback and robustness of apical constrictions in Drosophila embryo ventral furrow formation
<p>Formation of the ventral furrow in the Drosophila embryo relies on the apical constriction of cells in the ventral region to produce bending forces that drive tissue invagination. Recently [J Phys Condens Matter. 2016;28(41):414021], we observed that apical constrictions during the initial phase of ventral furrow formation produce elongated patterns of cellular constriction chains prior to invagination, and argued that these are indicative of tensile stress feedback. Here, we quantitatively analyze the constriction patterns preceding ventral furrow formation and find that they are consistent with the predictions of our active-granular-fluid model of a monolayer of mechanically coupled stress-sensitive constricting particles. Our model shows that tensile feedback causes constriction chains to develop along underlying precursor tensile stress chains that gradually strengthen with subsequent cellular constrictions. As seen in both our model and available optogenetic experiments, this mechanism allows constriction chains to penetrate or circumvent zones of reduced cell contractility, thus increasing the robustness of ventral furrow formation to spatial variation of cell contractility by rescuing cellular constrictions in the disrupted regions.</p>
DATA - Effects of Hearing Aid Amplification on Robust Neural Coding of Speech
<p>This data is presented in the following dissertation:<br /> Effects of Hearing Aid Amplification on Robust Neural Coding of Speech<br /> http://docs.lib.purdue.edu/open_access_dissertations/190/</p> <p>The code for analyzing this data is here:<br /> http://dx.doi.org/10.5281/zenodo.49296</p> <p>The data is organized as follows:</p> <ul> <li>The main file is Research.zip. This contains the directory structure within the "Research" folder.</li> <li>The contents of PhaseModulation.zip should be moved to "Research\Matlab"</li> <li>Everything else should be moved to "Research\Matlab\Vowel_STMP\ExpData"</li> </ul> <p> </p>
GERBIL evaluation data of Robust and Collective Entity Disambiguation through Semantic Embeddings
<p>A table containing the SIGIR 2016 experiments performed with GERBIL in context of the SIGIR 2016 work "Robust and Collective Entity Disambiguation through Semantic Embeddings" by Stefan Zwicklbauer, Christin Seifert and Michael Granitzer</p> <p>It also contains the original URL to the GERBIL website</p> <p>Corresponding GitHub Repository:</p> <p>https://github.com/quhfus/</p> <p> </p>
EEG Motor Imagery Dataset from the PhD Thesis "Commande robuste d'un effecteur par une interface cerveau machine EEG asynchrone"
<p>This Dataset contains EEG recordings from 8 subjects, performing 2 task of motor imagination (right hand, feet or rest). Data have been recorded at 512Hz with 16 wet electrodes (Fpz, F7, F3, Fz, F4, F8, T7, C3, Cz, C4, T8, P7, P3, Pz, P4, P8) with a g.tec g.USBamp EEG amplifier.</p> <p>File are provided in MNE raw file format. A stimulation channel encoding the timing of the motor imagination. The start of a trial is encoded as 1, then the actual start of the motor imagination is encoded with 2 for imagination of a right hand movement, 3 for imagination of both feet movement and 4 with a rest trial.</p> <p>The duration of each trial is 3 second. There is 20 trial of each class.</p>
Experimental data for "Spot-On: robust model-based analysis of single-particle tracking experiments"
<p><strong>Overview of experimental spaSPT data</strong></p> <p>To comprehensively test Spot-On over many different conditions, we conducted 1064 spaSPT experiments. The raw data is freely available and the purpose of this ReadMe file is to describe the organization, acquisition parameters and format of the data. The data is for 4 different cell lines imaged over 15 different conditions yielding a total of 60 different conditions. The four cell lines were:</p> <ul> <li> <p>U2OS C32 Halo-CTCF</p> </li> <li> <p>U2OS H2B-Halo-SNAP</p> </li> <li> <p>U2OS Halo-3xNLS</p> </li> <li> <p>mESC (JM8.N4) C3 Halo-Sox2</p> </li> </ul> <p>The cell lines were constructed in different ways. U2OS C32 Halo-CTCF was made by homozygous endogenous N-terminal tagging of CTCF in human osteosarcoma U2OS cells using CRISPR/Cas9-mediated genome-editing as described (C32 refers to clone number 32)<sup>1</sup>. We note the CTCF is an essential gene and that N-terminal tagging did not appear to affect CTCF function or expression level according to a series of control experiments<sup>1</sup>. Moreover, C32 Halo-CTCF has been authenticated using Short Tandem Repeat (STR) profiling (performed by Dr. Alison N. Killilea at the UC Berkeley Cell Culture Facility) against the following loci: THO1, D5S818, D13S317, D7S820, D16S539, CSF1PO, AMEL, vWA and TPOX. The C32 Halo-CTCF cell line showed a 100% match with U2OS.</p> <p>U2OS H2B-Halo-SNAP was made through random integration of a H2B-HaloTag-SNAP-Tag transgene expressed using the EF1a promoter with an IRES-NeoR gene for drug selection. After transfection, cells were selected using G418 until a pure cell population was obtained. This cell line has also been described previously<sup>1</sup>. The wild-type U2OS cell line used to make this cell line was also authenticated using STR profiling against the same loci as C32 and also showed a 100% match with U2OS.</p> <p>U2OS Halo-3xNLS was made through random integration of a FLAG-Halo-3xNLS (3x SV40 NLS: PKKKRKV) transgene expressed using the EF1a promoter. NeoR for drug selection was separately expressed using an SV40 promoter. After transfection, cells were selected using G418 until a pure cell population was obtained. This cell line has also been described previously<sup>1</sup>. The wild-type U2OS cell line used to make this cell line was also authenticated using STR profiling against the same loci as C32 and also showed a 100% match with U2OS.</p> <p>mESC C3 Halo-Sox2 was made through homozygous N-terminal tagging of Sox2 in JM8.N4<sup>2</sup> mouse embryonic stem cells using CRISPR/Cas9-mediated genome editing as previously described (C3 refers to clone number 3)<sup>3</sup>. The functionality of the C3 Halo-Sox2 knock-in was validated through control experiments and pluripotency through teratoma assays as described previously<sup>3</sup>.</p> <p>Each file contains single-molecule trajectories from a single cell imaged over 30,000 frames. Localization and tracking was performed using a custom-written Matlab implementation of the MTT-algorithm<sup>4</sup> and the following settings: Localization error: 10<sup>-6.25</sup>; deflation loops: 0; Blinking (frames): 1; max competitors: 3; max <em>D</em> (m<sup>2</sup>/s): 20.</p> <p>The same 15 conditions were used for each of the 4 cell lines.</p> <p><strong>ExpA PA-JF549</strong></p> <p>The purpose of this experiment was to test the effect of “motion-blurring” on the Spot-On estimated <em>D</em><sub>FREE</sub> and <em>F</em><sub>BOUND</sub>. 5 different experimental conditions were considered. Full details are given in the Methods section. Briefly, cells were grown overnight on plasma-cleaned 25 mm circular coverslips either directly (U2OS) and MatriGel coated as described<sup>1</sup>. Cell were labeled with 5-50 nM PA-JF549<sup>5</sup> for around 15-30 min, washed twice and medium exchanged to phenol-red free medium. 30,000 frames were collected at a camera exposure time (Andor iXon Ultra 897; frame-transfer mode; vertical shift speed: 0.9 μs; -70C) of 9.5 ms which together with a ~447 μs camera integration time gave a frame rate of ~100 Hz. PA-JF549 dyes were photo-activated during the ~447 μs camera integration time using 405 nm pulses and the 405 nm pulse intensity optimized to achieve a mean density of 1 molecule per frame per nucleus. The JF549 dye was excited using a 561 nm laser and the total number of excitation photons kept constant but either delivered during a 1 ms pulse, a 2 ms pulse, a 4 ms pulse, a 7 ms pulse or with constant illumination.</p> <p>For each cell line and condition, 4 replicates were performed. We count a replicate as an independent experiment performed on a different day. For each replicate around 5 cells were imaged. Occasionally, fewer than 5 cells are available. To avoid tracking errors, we removed cells with too high a localization density from the analysis. All of this information is available in the file name. For example, “U2OS_C32_Halo-CTCF_PA-JF549_1ms-561nm_100Hz_rep2_cell03” refers to the third cell imaged in the second replicate of U2OS C32 Halo-CTCF using a 1 ms excitation pulse of 561 nm laser at a frame rate of 100 Hz. Similarly, “U2OS_C32_Halo-CTCF_PA-JF549_cont-561nm_100Hz_rep4_cell01” refers to the first cell imaged in the fourth replicate of U2OS C32 Halo-CTCF using constant 561 nm laser at a frame rate of 100 Hz.</p> <p>The five ExpA_PAJF549 conditions are separated by cell line such that each cell line is provided in a separate directory. E.g. the directory “U2OS_H2B_ExpA_PAJF549” contains all data for the U2OS H2B-Halo-SNAP cell line.</p> <p><strong>ExpA PA-JF646</strong></p> <p>This experiment was exactly identical to the “ExpA_PA-JF549” experiment except cell were labeled with PA-JF646<sup>5</sup> and excited using a 633 nm laser. The file names and data organization was otherwise the same and the same five excitation conditions were considered.</p> <p><strong>ExpB PA-JF646</strong></p> <p>The purpose of this experiment was to test if the Spot-On estimated <em>D</em><sub>FREE</sub> and <em>F</em><sub>BOUND</sub> values would depend on the frame rate. In particular, all four proteins exhibit some levels of apparent anomalous diffusion, which could cause a dependence on the frame rate. Cells were labeled with PA-JF646 and grown and imaged as described above. Photo-activation took place during the ~447 μs camera integration time and JF646 dyes were excited using 1 ms stroboscopic 633 nm excitation pulses. To change the frame rate, the camera exposure time was set to 4.5 ms (~201 Hz), 5.5 ms (~167 Hz), 7 ms (~134 Hz), 13 ms (~74 Hz) and 19.5 ms (~50 Hz) when also counting the ~447 μs camera integration time. All of this information is available in the file name. For example, “U2OS_Halo-3xNLS_PA-JF646_1ms-633nm_74Hz_rep2_cell04” refers to the fourth cell imaged in the second replicate of U2OS Halo-3xNLS using a 1 ms excitation pulse of 633 nm laser at a frame rate of 74 Hz. Similarly, “mESC_C3_Halo-Sox2_PA-JF646_1ms-633nm_201Hz_rep1_cell03” refers to the third cell imaged in the first replicate of mESC Halo-Sox2 using a 1 ms excitation pulse of 633 nm laser at a frame rate of 201 Hz.</p> <p><strong>Data format</strong></p> <p>All data is available in two different formats: CSV-files and Matlab MAT-files. Both file formats are readable by the web-version of Spot-On. The Matlab version of Spot-On is only able to read the MAT-files. The CSV format consists of comma-separated values and contains headers. If opened with Microsoft Excel, it should appear as shown:</p> <p>Here the “frame” column contains the frame number in which the molecule was detected. The “t” column contains the timestamp. The “trajectory” column contains the trajectory number. For example, trajectory number 1 was only detected in frame 13 after which it disappeared. In contrast, trajectory number 4 was detected in frames 20, 21 22, 23 and 24. Finally, the “x” and “y” columns contain the x,y coordinates of the localization in units of micrometers (μm).</p> <p>The MAT-files contain a structure array named “trackedPar”. trackedPar contains three variables:</p> <ul> <li> <p>trackedPar.xy: “xy” is a matrix with 2 columns and a number of rows corresponding to the number of localizations in that trajectory. The first column is the x-coordinate and the second column is the y-coordinate. The units are micrometers (μm).</p> </li> <li> <p>trackedPar.Frame: “Frame” is a column vector where each element is the frame where the particle was localized.</p> </li> <li> <p>trackedPar.TimeStamp: “TimeStamp” is a column vector where each element is the timepoint where the particle was localized.</p> </li> </ul> <p>Each element in the structure array “trackedPar” correspond to a different trajectory.</p>
Simulated data for "Spot-On: robust model-based analysis of single-particle tracking experiments"
<p><strong>Generation of simulated data</strong></p> <p>To systematically evaluate the performance of Spot-On as well as other common analysis tools such as MSD<sub>i</sub> and vbSPT, we considered a comprehensive set of 3480 realistic SPT simulations spanning the range of plausible dynamics. The simulations were performed using simSPT, which is freely available at GitLab: https://gitlab.com/tjian-darzacq-lab/simSPT. The simulation methods are described in detail at GitLab. A full description of the parameters which allows exact reproduction of the simulations is available together with the data (see Data Availability section). Briefly, we parameterized simSPT to consider that particles diffuse inside a sphere (the nucleus) of 8 µm diameter illuminated using HiLo illumination (assuming a HiLo beam width of 4 µm), with an axial detection range of ~700 nm, centered at the middle of the HiLo beam. Molecules are assumed to have a half-life of 4 frames (when inside the HiLo beam) and of 40 frames when outside the HiLo beam. The localization error was set to 25 nm and the simulation was run until 100000 in-focus trajectories were recorded. More specifically, the effect of the exposure time (1 ms, 4 ms, 7 ms, 13 ms, 20 ms), the free diffusion constant (from 0.5 µm²/s to 14.5 µm²/s in 0.5 µm²/s increments) and the fraction bound (from 0 % to 95 % in 5 % increments) were investigated, yielding a dataset consisting of 3480 simulations. The advantage of simulations is that the ground truth is known. This allows a quantitative assessment of which method works the best.</p> <p><strong>Content of the archives:</strong></p> <ol> <li>170718_simSPT_simulations.zip the code and instructions to reproduce the simulations</li> <li>4um.tar.bz2 simulated data inside a 4 µm nucleus</li> <li>20um.tar.bz2 simulated data inside a 20 µm nucleus, in which virtually no confinement occurs.</li> <li>subsampled.tar.bz2 is a set of subsampled datasets, containing either 99999, 30000, 10000, 3000, 1000, 300, 100 or 30 trajectories. Each subsampling was done 50 times, yielding 50 files per subsmpling.</li> </ol> <p><strong>Formats:</strong></p> <p>The data is provided both in CSV and .mat formats. .mat files are provided in the following dataset: 10.5281/zenodo.835541</p>
Data for "Early and Widespread Emergence of Regional Warming is Robust to Observational and Model Uncertainty"
<p>These data can be used to reproduce all figures in "Early and Widespread Emergence of Regional Warming is Robust to Observational and Model Uncertainty". Figure code is hosted at https://github.com/jshaw35/RegionalToE_ShawAndLenssen/releases/tag/v1.0</p>
Social network structure is robust to parasite induced changes in contact behavior of domestic sheep
<p>Understanding how parasitism may affect social behavior and social networks is key to understanding the impact of infection on a population. Infection can disrupt social networks by altering the behavior of both infected individuals (e.g. by reducing activity) and the behavior of uninfected individuals (e.g. avoiding sick individuals), both of which can <span>have an impact on social group dynamics and parasite transmission</span>. Here we test experimentally how parasitism affects social contact behavior and social network structure using a common parasite infection of sheep. Three treatment groups, each with 4 replicate social groups were established (i) Parasitised; all lambs were infected with a parasitic nematode, (ii) Non-parasitised; all lambs remained uninfected (iii) Mixed; part of each group were infected, and part of the group remained uninfected. Contact behaviours of each individual were recorded using proximity loggers during four phases of infection (pre-parasite, pre-patent, patent-parasite, post-parasite). We found infected individuals in the parasitised and mixed groups reduced contact frequency following infection. Infected individuals in mixed groups however reduced contact frequency to a greater extent than infected animals in the fully parasitised group. D<span>espite the reduction in contacts between infected animals in the mixed group, the social network structure was unaffected, as non-infected individuals maintained pre-parasite levels of social interactions with their infected conspecifics. </span><span>These results demonstrate </span>how infection can impact the social behavior of all animals within a group, and how the expression of behavioral change may depend on the parasitic status of all group members and the response of uninfected conspecifics.</p>
Data for "Strong El Niño events lead to robust multi-year ENSO predictability"
<p>The full raw data, intermediate steps, and final results for all three experiments used in Lenssen et al. (202X), "Strong El Niño events lead to robust multi-year ENSO predictability," as submitted to GRL. Each zip file contains the raw data, intermediate steps, and final results for model analog forecast issued to predict that model or observational record. All data has been gridded to a common 2x2 grid using nco following the include `mygrid_2x2` file.</p><p>Archived codebase for this data available at <a href="https://zenodo.org/doi/10.5281/zenodo.10045615">https://zenodo.org/doi/10.5281/zenodo.10045615</a><br>Live codebase at <a href="https://github.com/nlenssen/LongLeadENSO/">https://github.com/nlenssen/LongLeadENSO/</a></p>
Data associated with "Robust Structured Illumination Microscopy with Bayesian Noise Control"
<p>Experimental and synthetic data saved in tiff and/or zarr formats and SIM reconstruction scripts written in Python (Wiener and FISTA-SIM0. These also include estimated SIM patterns which are needed for B-SIM</p><ul><li>Synthetic data consisting of variably spaced line pairs. Found in <a href="https://zenodo.org/uploads/10037823">2023_10_02_synthetic_line_pairs.zip</a></li><li>Experimental data. Fluorescence images of one of the variably spaced line pair patterns on an ArgoSIM calibration slide. Found in <a href="https://zenodo.org/uploads/10037823">2023_08_02_folder=002_argosim_slide.zip</a></li><li>Experimental data. MitoTracker Red labelled mitochondria in live HeLa cells. Found in <a href="2023_08_07_folder=011_mitos_live_hela.zip">2023_08_07_folder=011_mitos_live_hela.zip</a></li><li>Camera calibration maps, including gain, variance, and offset. Found in <a href="camera_calibration.zip">camera_calibration.zip</a></li></ul>
Robust evidence for bats as reservoir hosts is lacking in most African virus studies – a review and call to optimize sampling and conserve bats
<p><span>Africa experiences frequent emerging disease outbreaks among humans, with bats often proposed as zoonotic pathogen hosts. We comprehensively reviewed virus-bat findings from papers published between 1978 and 2020 to evaluate the evidence that African bats are reservoirs and/or bridging hosts for viruses that cause human disease. We present data from 162 papers (of 1322) with original findings on (1) numbers and species of bats sampled across bat families and the continent, (2) how bats were selected for study inclusion, (3) if bats were terminally sampled, (4) what types of ecological data, if any, were recorded, and (5) which viruses were detected and with what methodology. We propose a scheme for evaluating presumed virus-host relationships by evidence type and quality, using the contrasting available evidence for </span><em>Orthoebolavirus</em> (formerly <em>Ebolavirus</em>) versus <em>Orthomarburgvirus</em> (formerly <em>Marburgvirus</em>) as an example. We review the wording in abstracts and discussions of all 162 papers, identifying key framing terms, how these refer to findings, and how they might contribute to people's beliefs about bats. We discuss the impact of scientific research communication on public perception and emphasize the need for strategies that minimize human-bat conflict and support bat conservation. Finally, we make recommendations for best practices that will improve virological study metadata.</p>
Out-of-plane ferroelectricity and robust magnetoelectricity in quasi two-dimensional materials
<p>Thin film ferroelectrics have been pursued for capacitive and nonvolatile memory devices. They rely on polarizations that are oriented in an out-of-plane direction to facilitate integration and addressability with CMOS architectures. The internal depolarization field, however, formed by surface charges can suppress the out-of-plane polarization in ultrathin ferroelectric films that could otherwise exhibit lower coercive fields and operate with lower power. Here we unveil stabilization of a polar longitudinal optical (LO) mode in the <em>n</em>=2 Ruddlesden–Popper family that produces out-of-plane ferroelectricity, persists under open-circuit boundary conditions, and is distinct from hyperferroelectricity. Our first-principles calculations show the stabilization of the LO mode is ubiquitous in chalcogenides and halides and relies on anharmonic trilinear mode coupling. We further show that the out-of-plane ferroelectricity can be predicted with a crystallographic tolerance factor, and we use these insights to design a room-temperature multiferroic with strong magnetoelectric coupling suitable for magneto-electric spin-orbit transistors. </p>
Robustness assessment of a C++ implementation of the LeNet-5 convolutional neural network.
<p>The architecture of the LeNet-5 convolutional neural network (CNN) was defined by LeCun in its paper "Gradient-based learning applied to document recognition" (https://ieeexplore.ieee.org/document/726791) to classify images of hand written digits (MNIST dataset).</p><p>This architecture has been customized to use Rectified Linear Unit (ReLU) as activation functions instead of Sigmoid.</p><p>It consists of the following layers:</p><ul><li><strong>conv1</strong>: Convolution 2D, 1 input channel (28x28), 3 output channels (28x28), kernel size 5, stride 1, padding 2.</li><li><strong>relu1</strong>: Rectified Linear Unit (3@28x28).</li><li><strong>max1</strong>: Subsampling buy max pooling (3@14x14).</li><li><strong>conv2</strong>: Convolution 2D, 3 input channels (14x14), 6 output channels (14x14), kernel size 5, stride 1, padding 2.</li><li><strong>relu2</strong>: Rectified Linear Unit (6@14x14).</li><li><strong>max2</strong>: Subsampling buy max pooling (6@7x7).</li><li><strong>fc1</strong>: Fully connected (294, 147)</li><li><strong>fc2</strong>: Fully connected (147, 10)</li></ul><p>The fault hypotheses for this work include the occurrence of:</p><ul><li><strong>BF</strong>: single bit-flip faults</li><li><strong>S0</strong>: single, double-adjacent and triple-adjacent stuck-at-0 faults</li><li><strong>S1</strong>: single, double-adjacent and triple-adjacent stuck-at-1 faults</li></ul><p>In the memory cells containing all the parameters of the CNN: </p><ul><li><strong>w</strong>: weights (float32)</li><li><strong>b</strong>: biases (float32)</li></ul><p>Images 200 to 249 from the MNIST dataset have been used as workload.</p><p>This dataset contains the raw data obtained from running exhaustive fault injection campaigns for all considered fault models, targeting all considered locations and for all the images in the workload.</p><h3>Files information</h3><ul><li><i>golden_run.csv</i>: Prediction obtained for all the images considered in the workload in the absence of faults (Golden Run). This is intended to act as oracle to determine the impact of injected faults.</li><li><i>single_faults/bit_flip</i> folder: Prediction obtained for all the images considered in the workload in presence of single bit-flip faults. There is one file for each parameter of each layer.</li><li><i>single_faults/stuck_at_0</i> folder: Prediction obtained for all the images considered in the workload in presence of single stuck-at-0 faults. There is one file for each parameter of each layer.</li><li><i>single_faults/stuck_at_1</i> folder: Prediction obtained for all the images considered in the workload in presence of single stuck-at-1 faults. There is one file for each parameter of each layer.</li><li><i>double_adjacent_faults/stuck_at_0</i> folder: Prediction obtained for all the images considered in the workload in presence of double adjacent stuck-at-0 faults. There is one file for each parameter of each layer.</li><li><i>double_adjacent_faults/stuck_at_1</i> folder: Prediction obtained for all the images considered in the workload in presence of double adjacent stuck-at-1 faults. There is one file for each parameter of each layer.</li><li><i>triple_adjacent_faults/stuck_at_0</i> folder: Prediction obtained for all the images considered in the workload in presence of triple adjacent stuck-at-0 faults. There is one file for each parameter of each layer.</li><li><i>triple_adjacent_faults/stuck_at_1 </i>folder: Prediction obtained for all the images considered in the workload in presence of triple adjacent stuck-at-1 faults. There is one file for each parameter of each layer.</li></ul><h3>Methodology information</h3><p>First, the CNN was used to classify all the images of the workload in the absence of faults to get a reference to determine the impact of faults. This is <i>golden_run.csv</i> file.</p><p>After that, one fault injection experiment was executed for each of the 16 most significant bits bit of each element of each parameter of the CNN, as previous fault injection experiments showed that the occurrence of the considered faults in the 16 least significant bits does not impact the behaviour of the network.</p><p>Each experiment consisted in:</p><ul><li>Affecting the bits (inverting it in case of bit-flip faults, setting it to 0 or 1 in case of stuck-at-0 or atuck-at-1 faults) identified by the mask.</li><li>Classifying all the images of the workload in the presence of this fault. The obtained output was stored in a given .csv file.</li><li>Removing the fault from the CNN by restoring the affected bits to its previous value.</li></ul><h3>List of variables (Name : Description (Possible values))</h3><ul><li><strong>IMGID</strong>: Integer number identifying the considered image (200-249).</li><li><strong>TENSORID</strong>: Integer number identiying the parameter affected by the fault (0 - No fault, 1 - conv1.w, 2 - conv1.b, 3 - conv2.w, 4 - conv2.b, 5 - fc1.w, 6 - fc1.b, 7 - fc2.w, 8 - fc2.b).</li><li><strong>ELEMID</strong>: Integer number identiying the element of the parameter affected by the fault (-1 - No fault, [0-2] - conv1.b, [0-74] - conv1.w, [0-5] - conv2.b, [0-149] - conv2.w, [0-146] - fc1.b, [0-43217] - fc1.w, [0-9] - fc2.b, [0-1469] - fc2.w).</li><li><strong>MASK</strong>: 8-digit hexadecimal number identifying those bits affected by the fault ([00000000 - No fault, FFFFFFFF - all 32 bits faulty]).</li><li><strong>FAULT</strong>: String identiying the type of fault (NF - No fault, BF - bit-flip, S0 - Stuck-at-0, S1 - Stuck-at-1).</li><li><strong>SOFTMAX</strong>: 10 decimal numbers obtained after applying the softmax function to the provided output. They represent the probability of the image of belonging to the corresponding category for classification.</li><li><strong>PRED</strong>: Integer number representing the category predicted for the processed image.</li><li><strong>LABEL</strong>: integer number representing the actual category for the processed image.</li></ul>
Evaluation of tracking performance and robustness for a hybrid locomotion controller
<p>Legged locomotion is a complex control problem that requires both accuracy and robustness to cope with real-world challenges. Legged systems have traditionally been controlled using trajectory optimization with inverse dynamics. Such hierarchical model-based methods are appealing due to intuitive cost function tuning, accurate planning, generalization, and most importantly, the insightful understanding gained from more than one decade of extensive research. However, model mismatch and violation of assumptions are common sources of faulty operation. Simulation-based reinforcement learning, on the other hand, results in locomotion policies with unprecedented robustness and recovery skills.<br>Yet, all learning algorithms struggle with sparse rewards emerging from environments where valid footholds are rare, such as gaps or stepping stones. In this work, we propose a hybrid control architecture that combines the advantages of both worlds to simultaneously achieve greater robustness, foot-placement accuracy, and terrain generalization. Our approach utilizes a model-based planner to roll out a reference motion during training. A deep neural network policy is trained in simulation, aiming to track the optimized footholds. We evaluate the accuracy of our locomotion pipeline on sparse terrains, where pure data-driven methods are prone to fail. Furthermore, we demonstrate superior robustness in the presence of slippery or deformable ground when compared to model-based counterparts. Finally, we show that our proposed tracking controller generalizes across different trajectory optimization methods not seen during training. In conclusion, our work unites the predictive capabilities and optimality guarantees of online planning with the inherent robustness attributed to offline learning.</p>
Data from: Urbanisation and agricultural intensification modulate plant-pollinator network structure and robustness
<p>Land use change is a major pressure on pollinator abundance, diversity, and plant-pollinator interactions. Far less is known about how land use alters the structure of plant-pollinator networks and their robustness to plant-pollinator coextinctions.</p> <p>We analyzed the structure of plant-pollinator networks sampled in 12 landscapes along an urbanisation and agricultural intensity gradient, from early spring to late summer 2021, and used a stochastic coextinction model to correlate plant-pollinator coextinction risk with network structure (species and network-level metrics) and landscape context.</p> <p>Networks in intensively managed (i.e. agricultural and urban) landscapes had a lower risk of initiating a coextinction cascade, while networks in less-intensively managed landscapes may be less robust. Network structure modulated the frequency and severity of coextinctions and species loss, while the strength of species interactions increased robustness.</p> <p>Urban networks were more species-rich and symmetrical due to the high diversity of ornamental plants, while intensively managed agricultural landscapes had smaller, more tightly connected, and nested networks.</p> <p>Network structure modulated the frequency of extinctions, which was decreased by greater linkage density, interaction asymmetry, and interaction dependence in the networks, while once an extinction occurred, nestedness and linkage density propagated the degree of the coextinction cascade and species loss. At the species level, species strength was inversely correlated with extinction risk, implying that generalist species with a high number of interactions with specialists had the lowest extinction risk.</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.