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.
685
datasets available to search
ShareScore release 0.7.1
Dataset results
685 results for “error”
Determining non-significant bits on a C++ implementation of the LeNet-5 convolutional neural network to be used for storing error correcting codes to protect weights and biases. Robustness assessment of the network after integrating the proposed codes.
<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" (<a href="https://ieeexplore.ieee.org/document/726791">https://ieeexplore.ieee.org/document/726791</a>) 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>S0</strong>/<strong>S1</strong>: multiple adjacent stuck-at-0 and stuck-at-1 faults to determine the least significant bits of weights and biases that could be used to store the proposed error correcting codes.</li><li><strong>BF</strong>: single, double, and triple bit-flip faults to assess the robustness of the considered CNN</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>All the images (10000) from the MNIST dataset have been used as workload.</p><p>The weights and biases of the LeNet-5 architecture have been protected using six different error correcting codes that have been deployed in the least significant bits of these elements.</p><p>The parity check matrices (H = P I) that define these ECCs are:</p><ul><li><strong>SEC(32, 26)</strong> (Hamming) under a <i>classic policy </i>(see methodology below):</li></ul><p><i> 11010010001000011101101000 100000</i></p><p><i> 10101001000100011011010100 010000</i></p><p><i> 01100100100010010110110010 001000</i></p><p><i> 00011100010001001110001101 000100</i></p><p><i> 00000011110000100001111011 000010</i></p><p><i> 00000000001111100000000111 000001</i></p><ul><li><strong>SEC(23, 18)</strong> (Hamming) under a <i>conservative policy</i> (see methodology below):</li></ul><p><i> 111100001111000000 10000</i></p><p><i> 110011101000111000 01000</i></p><p><i> 101011010100100110 00100</i></p><p><i> 010110110010010101 00010</i></p><p><i> 001101110001001011 00001</i></p><ul><li><strong>SEC(13, 9)</strong> (Hamming) under an <i>aggressive policy </i>(see methodology below):</li></ul><p><i> 110111000 1000</i></p><p><i> 101100110 0100</i></p><p><i> 011010101 0010</i></p><p><i> 111001011 0001</i></p><ul><li><strong>DEC(32, 21)</strong> (low redundancy and reduced overhead DEC) under a <i>classic policy </i>(see methodology below):</li></ul><p><i> 111000011001010010000 10000000000</i></p><p><i> 110110000011101000000 01000000000</i></p><p><i> 101011000110000010001 00100000000</i></p><p><i> 100101101000110001000 00010000000</i></p><p><i> 011010101100100000100 00001000000</i></p><p><i> 010101010100001001010 00000100000</i></p><p><i> 001100110010010100100 00000010000</i></p><p><i> 000011110001000110010 00000001000</i></p><p><i> 000000001111001101001 00000000100</i></p><p><i> 000000000000111100111 00000000010</i></p><p><i> 000000000000000011111 00000000001</i></p><ul><li><strong>DEC(28, 18)</strong> (low redundancy and reduced overhead DEC) under a <i>conservative policy </i>(see methodology below):</li></ul><p><i> 111111000000000000 1000000000</i></p><p><i> 110100111100000000 0100000000</i></p><p><i> 110000100011110000 0010000000</i></p><p><i> 001110010011001100 0001000000</i></p><p><i> 101100001010101010 0000100000</i></p><p><i> 010001001101010110 0000010000</i></p><p><i> 001011000101101001 0000001000</i></p><p><i> 101000011000110101 0000000100</i></p><p><i> 010001110000011011 0000000010</i></p><p><i> 000010100110000111 0000000001</i></p><ul><li><strong>DEC(17, 9)</strong> (low redundancy and reduced overhead DEC) under an <i>aggressive policy </i>(see methodology below):</li></ul><p><i> 111110000 10000000</i></p><p><i> 111001100 01000000</i></p><p><i> 110101010 00100000</i></p><p><i> 101010110 00010000</i></p><p><i> 101101001 00001000</i></p><p><i> 100110101 00000100</i></p><p><i> 100011011 00000010</i></p><p><i> 110000111 00000001</i></p><p>This dataset contains the raw data obtained from:</p><ul><li>running exhaustive fault injection campaigns for increasingly multiple stuck-at faults in the least significant bits of all weights and biases (simultaneously) and for all the images in the workload.</li><li>running statistical fault injection campaigns for single, double, and triple bit-flip faults, randomly targeting the considered locations and images in the workload.</li></ul><h3>Files information</h3><ul><li><i>no_ecc </i>folder: Results obtained for the original (not protected) version of the CNN.<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>sampling_SBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for single bit-flip faults.</li><li><i>sampling_DBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for double bit-flip faults.</li><li><i>sampling_TBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for triple bit-flip faults.</li><li><i>locating_sensitive_bits </i>folder: Prediction obtained for all the images considered in the workload in presence of stuck-at-0/stuck-at-1 faults that simultaneously target the N least significant bits of all weights and biases. There is one file for each parameter of type of fault and range of targeted bits. Files for bits in the range [11, 0] are not included as they obtain eactly the same results as the Golden Run (faults do not alter the behaviour of the network).</li></ul></li><li><i>sec/classic</i>, <i>sec/conservative</i>, and <i>sec/aggressive</i> folders: They contain the results obtained for the CNN protected by SEC(32, 26), SEC(23, 18), and SEC(13, 9), respectively.<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. It must be noted that this file could be different that the golden_run.csv file for the original version of the CNN, as deploying the ECC in the weights and biases may have affected the behaviour of the network.</li><li><i>sampling_SBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for single bit-flip faults. They should all be tolerated by the definition of the ECC.</li><li><i>sampling_DBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for double bit-flip faults. They could be more harmful than for the unprotected version of the CNN, as the ECC may erroneously flip correct bits.</li></ul></li><li><i>dec/classic</i>, <i>dec/conservative</i>, and <i>dec/aggressive </i>folders: They contain the results obtained for the CNN protected by DEC(32, 21), DEC(28, 18), and DEC(17, 9), respectively.<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. It must be noted that this file could be different that the golden_run.csv file for the original version of the CNN, as deploying the ECC in the weights and biases may have affected the behaviour of the network.</li><li><i>sampling_DBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for double bit-flip faults. They should all be tolerated by the definition of the ECC.</li><li><i>sampling_TBF_10000.csv</i>: Prediction obtained for running 10000 statistical fault injection experiments for triple bit-flip faults. They could be more harmful than for the unprotected version of the CNN, as the ECC may erroneously flip correct bits.</li></ul></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>To locate non-significant bits in weights and biases, fault injection experiments were executed targeting all elements of all parameters of the CNN using the following procedure:</p><ul><li>The initial mask targeted only the least significant bit</li><li>Until the mask targets all bits of the elements (32 bits as they are single-precision floating point values):<ul><li>Affect the bits (setting them to 0 or 1 in case of stuck-at-0 or stuck-at-1 faults) identified by the mask for all elements of all parameters.</li><li>Classify all the images of the workload in the presence of this fault. The obtained output was stored in a given .csv file.</li><li>Remove the fault from the CNN by restoring the affected bits to its previous value.</li><li>Add the next adjacent bit to the mask, so it targets an additional least significant bit.</li></ul></li></ul><p>The analysis of the obtained results may help in determining which bits can be used to store an ECC:</p><ul><li>which bits never affect the behaviour of the CNN, as the predicted classification is exactly the same than in the absence of faults.</li><li>which bits midly affect the behaviour of the CNN, as although the predicted classifications differ from those in the absence of faults, the accuracy of the network is barely affected.</li><li>which bits greatly affect the behaviour of the CNN, as the accuracy of the network is significantly affected.</li></ul><p>Accordingly, three different policies have been identified for deploying an ECC using these bits:</p><ul><li><strong>Classic policy</strong>: The ECC protects as much bits as possible.</li><li><strong>Conservative policy</strong>: The ECC protects all those bits that may affect the prediction of the network.</li><li><strong>Aggressive policy</strong>: The ECC protects only those bits that significantly affect the accuracy of the network.</li></ul><p>After designing and deploying a single ECC and a double ECC for each of the identified policies, fault injection experiments were executed to verify their behaviour in the presence of faults.</p><p>Single and double ECCs were tested against single and double bit-flip, respectively (all faults should be tolerated,) and double and triple bit-flips, respectively (a correct bit could be erroneously flipped.)</p><p>Due to the heavy computational load of the decoders, statistical injection was used to run the required fault injection campaigns with a sample size (number of experiments) of 10000.</p><p>Each experiment consisted in:</p><ul><li>Randomly selecting the image to process, and the parameter, element, and bits (mask) to be targeted by the fault.</li><li>Affecting the bits (inverting them) identified by the mask.</li><li>Classifying the selected image 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 (1-9999).</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>
Error bounds for kernel-based approximations of the Koopman operator
<p>This repository contains python scripts and data to re-create the result shown in</p><p>`Error bounds for kernel-based approximations of the Koopman operator, arxiv:2301.08637`</p><p>See README for detailed instructions on how to re-create these data.</p>
Supporting Dataset (Tables S1-S7, Figure S1) for "Heavy-mineral grain counting: Counting techniques, error estimation, and the number of grains to be counted"
<p>This Dataset comprises Tables S1-S7 for the article <strong>"Heavy-mineral grain counting: Counting techniques, error estimation, and the number of grains to be counted"</strong> authored by Jan Schönig and submitted to J<em>ournal of Geophysical Research: Earth Surface</em>.</p> <p>Figure S1: Error of different counting methods in comparison with theory and without considering the finite population correction</p> <p>Table S1: Heavy-mineral dataset from semi-automated Raman analysis</p> <p>Table S2: Summary of heavy-mineral composition for individual samples in percent</p> <p>Table S3: Computed ribbon compositions for individual samples and ribbon sizes given in numbers of counts</p> <p>Table S4: Relative errors at 95 % quantile for consecutive ribbon counting simulations</p> <p>Table S5: Relative errors at 95 % quantile for maximum distance ribbon counting simulations</p> <p>Table S6: Relative errors at 95 % quantile for cluster counting simulations</p> <p>Table S7: Numerical solution for determining the number of counts required to discriminate the content of two mineral species in an aliquot</p>
Error Related Potential at the start of the gait with a lower limb exoskeleton
<h2>Description</h2> <p>This dataset contains EEG signals from experiments designed to evoke Error Related Potentials (ErrP) at the onset of gait using a Brain-Computer Interface (BCI) to control a lower limb exoskeleton. The ErrP is elicited using three different stimuli: Tactile, Visual, and VisuoTactile.</p> <p>During the experiment, participants remain stationary and engage in two mental tasks: Relax (R) and Motor Imagery (I) of walking to activate the exoskeleton. These tasks can be executed correctly (RC, IC) or incorrectly (RE, IE). For example, during RC (Relax Correct), the subject maintains an idle state, whereas during RE (Relax Error), the exoskeleton activates unexpectedly. Conversely, in IC, the subject imagines the sensation of starting to walk in their muscles, and the exoskeleton activates, but during IE, the exoskeleton does not move despite the motor imagery. When the exoskeleton activates before starting to walk, the stimulus remains active for 2 seconds to alert the subject about the impending movement. Therefore, ErrP is elicited by the stimuli in RE and can be compared with the absence of ErrP in IC, where the stimulus activates but should not evoke an error.</p> <p>Each subject participates in three sessions, one for each stimulus, consisting of 12 trials. In each trial, 10 mental tasks are performed, 5 Relax and 5 Imagination, interleaved. Since the subject is never in control of the system, tasks are correctly performed 70% of the time (RC, IC), while the remaining 30% are incorrect (RE, IE). In an exception, subject R01_VisuoTactile performed 7 trials of 20 mental tasks each, 10 of each type. However, due to the extended duration of the trials and resulting fatigue, they were later split for subsequent sessions.</p> <p> </p> <h2>Data information</h2> <p>A trial consists of a Matlab structure that stores all information related to the trial experiment. </p> <ul> <li><em>data_EEG</em>: Original EEG signals recorded with a sampling rate of 250Hz, where each row is a channel (1-28 EEG, 29-32 EOG, 33-35 inertial electrodes).</li> <li><em>data_preprocessed_EEG</em>: Matrix that contains the preprocessed signals for each channel. Rows 1-35 are the original signals and then, the preprocessed signals in blocks of 35. Find the indexes of each filter in <em>session.conf.info.preprocessingSteps.ListPreprocessingSteps</em>.</li> <li><em>trigger_EEG</em>: Information related to signal quality and missing data while recording. </li> <li><em>data_EXO</em>: Exoskeleton recorded data with a sampling rate of 250Hz.</li> <li><em>data_preprocessed_EXO: </em>The same data recorded by the exoskeleton in <em>data_EXO</em>, since it does not require the application of any filter.</li> <li><em>trigger_EXO</em>: Empty vector. </li> <li><em>data_Actuators</em>: Arduino response when activates (1) and deactivates (-1) the feedback. </li> <li><em>data_preprocessed_Actuators: </em>The same Arduino resposes recorded in <em>data_Actuators</em>, because it does not require any filter application. </li> <li><em>trigger_Actuators</em>: Empty vector. </li> <li><em>task_EEG</em>: Vector that associates a task to each signal sample.</li> <li><em>task_index_EEG</em>: Zero vector with negative peaks at the samples indicating the start of a task. Each peak decrements by one unit with each task. </li> <li><em>task_order_EEG</em>: Vector that increments a unit with each task change. </li> <li><em>event_EEG</em>: Vector of commands to activate (1) and deactivate (-1) the feedback in Arduino. </li> <li><em>conf</em>: Configuration employed for data acquisition and preprocessing. <ul> <li><em>acquisition</em>: User and signals acquisition information. <ul> <li><em>user_code</em>: User code name.</li> <li><em>feedback</em>: Trial in openloop (User do not have control of the system).</li> <li><em>feedbackErrP</em>: Feedback type employed during the trial.</li> <li><em>readfile</em>: Path to read files after its acquisition.</li> <li><em>saveSession_Script</em>: Script used to save the recorded data.</li> <li><em>writeResults</em>: Path to save the recorded data.</li> <li><em>device</em>: List of connected devices during the trial and their related information, such as name, sampling rate, connection order, etc. </li> <li><em>task</em>: Information about tasks occurring during the trial. <ul> <li><em>task_list</em>: Decodes tasks numbers. The first number is the global task/mental activity, the second one is the physiological state of the user, and the third one indicates the task version (preparation or basic task).</li> <li><em>sequence_tasks</em>: List of tasks in order of execution.</li> <li><em>sequence_times</em>: List with the duration of each task in the sequence.</li> </ul> </li> <li><em>deviceOutput</em>: List of devices that receive commands to execute orders, such as the exoskeleton for walking and stopping and the VibroLed for turning feeedback on and off.</li> <li><em>eye_index</em>: Indexes of EOG electrodes.</li> <li><em>EEG_index</em>: Indexes of EEG electrodes.</li> <li><em>inertial_index</em>: Indexes of inertial electrodes.</li> <li><em>file_name</em>: Trial name.</li> <li><em>num_epochs</em>: Number of epochs within a trial. An epoch is the half of sampling rate (250Hz), this means that an epoch has a duration of 0.5s and 125 samples. </li> </ul> </li> <li><em>preadjustment</em>: Empty list. </li> <li><em>preprocessing</em>: Information of the preprocessing filters, parameters and order of application.</li> <li><em>processing</em>: Not necessary for this analysis. </li> <li><em>static</em>: Information used internally by the architecture for its correct operation.</li> <li><em>info</em>: Important information about filters, their order and indexes in <em>data_processed_EEG</em>.</li> </ul> </li> <li><em>times</em>: Struct with information of the devices synchronization and preprocessing times.</li> <li><em>times_processing</em>: Processing duration times. </li> </ul>
Data for publication: "Quantifying the relationship between observed variables that contain censored values using Bayesian error-in-variables regression"
<p>This archive contains the two datasets used in the publication: Vermeiren, Charles, Munoz: Quantifying the relationship between observed variables that contain censored values using Bayesian error-in-variables regression <br>Preprint: <a href="https://hal.science/hal-04764660" rel="nofollow">https://hal.science/hal-04764660</a><br><br>The first dataset is used to develop and test the model using cross-validation, the 2nd dataset is used as an independent, external dataset to test the model. For details, see the publication.</p> <p>The model code, combined with the data and outputs, are also available on GitHub: https://github.com/Peter-Vermeiren/EIVmodels </p>
Database of PV output forecast errors
<p>This database is extracted from 180 studies on PV output forecasting and is used for the research paper "What drives the accuracy of PV output forecasts?". The data of 21 key variables including the publishing year of the papers, the error values, data processing techniques used by the models, the length of the test sets, the forecast resolution, the country and region of the studies, the methodology of the forecast models, the forecast horizon, and the error metrics are included. Besides, other information such as the weather condition of the forecasts, the number of power plants, the installed capacity... is also included. </p>
Data - Coherent combining of low-power optical signals based on optically amplified error feedback
<p>This dataset contains measurement data and processing code for the results published in "Coherent combining of low-power optical signals based on optically amplified error feedback". Code for the Micro-controllers used in the work is also attached.</p> <p>This work was funded by the Swedish Research Council (grant VR-2015-00535).</p>
Video-based learning of coping strategies for common errors improves laparoscopy training - a randomized study
<p>Instructional video material from the doctoral thesis "Introducing a coping role model to increase the learning efficiency in learning laparoscopic knot tying - a randomised controlled trial" by Ms Annabelle Gerhäuser at Heidelberg University, Medical School.</p> <p>Developing the skills needed to perform minimally invasive procedures requires intensive training of trainee surgeons. Laparoscopic knot tying is a particularly challenging technique. Subjects of the study learned this technique with the help of instructional videos developed specifically for this purpose. These "coping videos" show typical mistakes made by beginners and the corresponding solution strategies. They have been validated before use in the above-mentioned study.</p> <p>Coping Video 1 - Needle Load</p> <p>Coping Video 2 - Handling & Tissue</p> <p>Coping Video 3 - Tail length</p> <p>Coping Video 4 - Difficulties in knot tying</p> <p>Coping Video 5 - Lift & drift</p> <p> </p>
Data for: Assortative mating in an ecological context: Effects of mate choice errors and relative species abundance on the frequency and asymmetry of hybridization
<p><span>The frequency and asymmetry of mixed-species mating set the initial stage for the ecological and evolutionary implications of hybridization. How such patterns of mixed-species mating, in turn, are influenced by the combination of mate choice errors and relative species abundance remain largely unknown. We develop a mathematical model that generates predictions for how relative species abundances and mate choice errors affect hybridization patterns. When mate choice errors are small (<5%) the highest frequency of hybridization occurs when one of the hybridizing species is at low abundance, but when mate choice errors are high (>5%) the highest hybridization frequency occurs when species occur in equal proportions. Furthermore, females of the less abundant species are overrepresented in mixed-species matings. We compare our theoretical predictions with empirical data on naturally hybridizing Ficedula flycatchers and find that hybridization is highest when the two species occur in equal abundance, implying rather high mate choice errors. We discuss ecological and evolutionary implications of our findings and encourage future work on hybrid zone dynamics that take demographic aspects, such as relative species abundance, into account.</span></p>
Dataset for paper "Mitigating the effect of errors in source parameters on seismic (waveform) inversion"
<p>Dataset corresponding to the journal article "Mitigating the effect of errors in source parameters on seismic (waveform) inversion" by Blom, Hardalupas and Rawlinson, accepted for publication in Geophysical Journal International. In this paper, we demonstrate the effect or errors in source parameters on seismic tomography, with a particular focus on (full) waveform tomography. We study effect both on forward modelling (i.e. comparing waveforms and measurements resulting from a perturbed vs. unperturbed source) and on seismic inversion (i.e. using a source which contains an (erroneous) perturbation to invert for Earth structure. These data were obtained using Salvus, a state-of-the-art (though proprietary) 3-D solver that can be used for wave propagation simulations (Afanasiev et al., GJI 2018).</p> <p>This dataset contains:</p> <ul> <li>The entire Salvus project. This project was prepared using Salvus version 0.11.x and 0.12.2 and should be fully compatible with the latter.</li> <li>A number of Jupyter notebooks used to create all the figures, set up the project and do the data processing.</li> <li>A number of Python scripts that are used in above notebooks.</li> <li>two conda environment .yml files: one with the complete environment as used to produce this dataset, and one with the environment as supplied by Mondaic (the Salvus developers), on top of which I installed basemap and cartopy.</li> <li>An overview of the inversion configurations used for each inversion experiment and the name of hte corresponding figures: inversion_runs_overview.ods / .csv .</li> <li>Datasets corresponding to the different figures. <ul> <li>One dataset for Figure 1, showing the effect of a source perturbation in a real-world setting, as previously used by Blom et al., Solid Earth 2020</li> <li>One dataset for Figure 2, showing how different methodologies and assumptions can lead to significantly different source parameters, notably including systematic shifts. This dataset was kindly supplied by Tim Craig (Craig, 2019).</li> <li>A number of datasets (stored as pickled Pandas dataframes) derived from the Salvus project. We have computed: <ul> <li>travel-time arrival predictions from every source to all stations (df_stations...pkl)</li> <li>misfits for different metrics for both P-wave centered and S-wave centered windows for all components on all stations, comparing every time waveforms from a reference source against waveforms from a perturbed source (df_misfits_cc.28s.pkl)</li> <li>addition of synthetic waveforms for different (perturbed) moment tenors. All waveforms are stored in HDF5 (.h5) files of the ASDF (adaptable seismic data format) type</li> </ul> </li> </ul> </li> </ul> <p>How to use this dataset:</p> <ul> <li>To set up the conda environment: <ol> <li>make sure you have anaconda/miniconda</li> <li>make sure you have access to Salvus functionality. This is not absolutely necessary, but most of the functionality within this dataset relies on salvus. You can do the analyses and create the figures without, but you'll have to hack around in the scripts to build workarounds.</li> <li>Set up Salvus / create a conda environment. This is best done following the instructions on the Mondaic website. Check the changelog for breaking changes, in that case download an older salvus version.</li> <li>Additionally in your conda env, install basemap and cartopy: <pre><code class="language-bash">conda-env create -n salvus_0_12 -f environment.yml conda install -c conda-forge basemap conda install -c conda-forge cartopy</code></pre> </li> <li> <p>Install LASIF (https://github.com/dirkphilip/LASIF_2.0) and test. The project uses some lasif functionality.</p> </li> <li> <p> </p> </li> <li> <p> </p> </li> </ol> </li> <li>To recreate the figures: This is extremely straightforward. Every figure has a corresponding Jupyter Notebook. Suffices to run the notebook in its entirety. <ul> <li>Figure 1: separate notebook, Fig1_event_98.py</li> <li>Figure 2: separate notebook, Fig2_TimCraig_Andes_analysis.py</li> <li>Figures 3-7: Figures_perturbation_study.py</li> <li>Figures 8-10: Figures_toy_inversions.py</li> </ul> </li> <li>To recreate the dataframes in DATA: This can be done using the example notebook Create_perturbed_thrust_data_by_MT_addition.py and Misfits_moment_tensor_components.M66_M12.py . The same can easily be extended to the position shift and other perturbations you might want to investigate.</li> <li>To recreate the complete Salvus project: This can be done using: <ul> <li>the notebook Prepare_project_Phil_28s_absb_M66.py (setting up project and running simulations)</li> <li>the notebooks Moment_tensor_perturbations.py and Moment_tensor_perturbation_for_NS_thrust.py</li> <li>For the inversions: using the notebook Inversion_SS_dip.M66.28s.py as an example. See the overview table inversion_runs_overview.ods (or .csv) as to naming conventions.</li> </ul> </li> </ul> <p> </p> <p>References:</p> <ul> <li>Michael Afanasiev, Christian Boehm, Martin van Driel, Lion Krischer, Max Rietmann, Dave A May, Matthew G Knepley, Andreas Fichtner, Modular and flexible spectral-element waveform modelling in two and three dimensions, <em>Geophysical Journal International</em>, Volume 216, Issue 3, March 2019, Pages 1675–1692, <a href="https://doi.org/10.1093/gji/ggy469">https://doi.org/10.1093/gji/ggy469</a></li> <li>Nienke Blom, Alexey Gokhberg, and Andreas Fichtner, Seismic waveform tomography of the central and eastern Mediterranean upper mantle, <em>Solid Earth</em>, Volume 11, Issue 2, 2020, Pages 669–690, 2020, <a href="https://doi.org/10.5194/se-11-669-2020">https://doi.org/10.5194/se-11-669-2020</a></li> <li>Tim J. Craig, Accurate depth determination for moderate-magnitude earthquakes using global teleseismic data. <em>Journal of Geophysical Research: Solid Earth</em>, 124, 2019, Pages 1759– 1780. <a href="https://doi.org/10.1029/2018JB016902">https://doi.org/10.1029/2018JB016902</a></li> </ul> <p> </p>
Dataset for "The first mitotic division of human embryos is highly error prone"
<p>Data for The first mitotic division of human embryos is highly error prone. [Dataset]<br> Research output available from https://omero.warwick.ac.uk/webclient/?show=project-8301<br> Details for how to access the dataset are available on the public data page: https://warwick.ac.uk/fac/sci/med/research/biomedical/facilities/camdu/publicdata/</p> <p>The dataset contains movies of human embryos consented to research progressing through the first two embryonic mitoses. The chromosomes have been visualised using SiR-DNA dye and imaged using a deltavision widefield microscope.The dataset consists only of time lapse imaging movies.</p>
Widespread Error Detection in Large Scale Continuous Integration Systems
<p>Dataset of 5000 json documents describing verification process of <a href="https://github.com/facebook/react">React</a> project collected between 2023 and 2024. Errors recorded in this dataset were used in the presentation at CCIW workshop.</p> <p><strong>Abstract: </strong>Continuous Integration systems are widely used in the software industry to validate and integrate code changes into central repositories. Their effectiveness can be impacted by non-deterministic tests which can fail in the absence of any regression. Integration tests which depend on external services are particularly prone to this problem. We present a system which allows us to reduce the impact of non-deterministic failures by detecting widespread errors. The key assumption, which works well in practice, is that developers tend not to make identical mistakes simultaneously. If we observe a widespread error, it strongly suggests there is a problem with upstream services and not with the code change being evaluated. The detection algorithm consists of three main phases. First, the error text gets extracted from logs using predefined heuristics or automated methods. Then, this text gets fuzzy matched against a database of recently observed errors. Finally, statistics get checked to determine if they meet the criteria for a widespread error. When an error meets the criteria it either gets demoted to a warning or it gets enriched with information about an ongoing incident.</p>
figure 5 in The sawFIsh (Rhinopristiformes, Pristidae) rostrum displayed in the "Basilica Santuario del Carmine Maggiore" in Naples, Italy: A long story of legends and taxonomic errors
figure 5 Alignment of coi (A) and nd2 (B) sequences obtained for the rostrum of the bscm (indicated by a pink box) and those available in GenBank for sawfish species (see main text for details and Accession numbers). nd 2 haplotypes for P. pectinata from Faria et al. (2013) are indicated as HapC, HapD, HapE. Unrooted neighbor-joining tree (C) constructed using coi + nd2 sequences and displaying only>50% percentage bootstrap scores.
figure 1 in The sawFIsh (Rhinopristiformes, Pristidae) rostrum displayed in the "Basilica Santuario del Carmine Maggiore" in Naples, Italy: A long story of legends and taxonomic errors
figure 1 (A) Picture of the sawfish rostrum kept in the bscm and (B) illustration of the "miraculous event" in which it allegedly took part; note the fish tail protruding from the leak circled in white (from page 81; Bacco, 1605). (C) The original lacquer wax stamp present on the rostrum (central) compared to the logo of the Roman Curia of the Carmelite Order (left) and the logo of the Carmelite fathers of Naples (right). The logo of the Roman Curia of the Carmelite Order was downloaded from https://fraticarmelitani.wixsite.com/fraticarmelitani and, the Downloadedlogo of fromtheBrill.com 06/21/2024 06:27:30PM via Open Access . This is an open access article distributed under the terms Carmelite fathers of Naples was photographed by N. Maio . of the CC BY 4.0 license . https://creativecommons.org/licenses/by/4.0/
Рис. 1. Вероятность обнаружения меченых животных (среΑнее ± ошибка) при пяти- и Αесятиметровых интерваΛах межΑу прикормочными станциями в Αвух экспериментах. По второму эксперименту расчеты сΑеΛаны ΑΛя резуΛьтатов отΛова в течение первых трех и поΛных Αесяти Αней. Значение «p» отражает уровень статистической значимости разΛичий межΑу ΑоΛями животных с меткой при Αвух интерваΛах Fig. 1. Probability of finding marked animals (average±standard error) between feeding stations placed at intervals of five and ten meters in the two experiments. In the second experiment, calculations were made for the results of trapping during the first three days and during the whole period of ten days. The p value reflects the statistical significance of differences between the fractions of animals with a mark for two types of intervals in Verification of the bottle-based method for estimating abundance of small mammals using biomarkers
Рис. 1. Вероятность обнаружения меченых животных (среΑнее ± ошибка) при пяти- и Αесятиметровых интерваΛах межΑу прикормочными станциями в Αвух экспериментах. По второму эксперименту расчеты сΑеΛаны ΑΛя резуΛьтатов отΛова в течение первых трех и поΛных Αесяти Αней. Значение «p» отражает уровень статистической значимости разΛичий межΑу ΑоΛями животных с меткой при Αвух интерваΛах Fig. 1. Probability of finding marked animals (average±standard error) between feeding stations placed at intervals of five and ten meters in the two experiments. In the second experiment, calculations were made for the results of trapping during the first three days and during the whole period of ten days. The p value reflects the statistical significance of differences between the fractions of animals with a mark for two types of intervals
Figure 9. Mean Square Error for different bands-Classification of Human Emotion from Deap EEG Signal Using Hybrid Improved Neural Networks with Cuckoo Search
<p>Figure 9 has shown in different epochs using neural networks with mean square error<br> performance and the Table 3 shows that different bands mean square error values while training the<br> neural networks with particle swarm optimization.</p>
Figure 6. Mean square error of alpha band-Classification of Human Emotion from Deap EEG Signal Using Hybrid Improved Neural Networks with Cuckoo Search
<p>Particle swarm optimization algorithm first optimizes the neural networks weight and bias<br> and provides the minimum mean square error with nearer by zero. The following figure 6 has<br> shown that minimum mean square error when training the particular band features.</p>
Figure 7. Results obtained for the mean, the mean standard deviations and the mean standard errors of the independent speech variable in dyslalic subjects (both the control and the experimental group)
<p>Following the logopedic assessment of the investigated subjects, we moved on to the statistic processing of the gathered data, and we analysed the results obtained from the tests administered to the two groups after one year of speech therapy. The conducted analysis was directed both at the overall effectiveness of the speech therapy, and at the importance of strategies for the language development and stimulation, through the use of the computer-assisted Terapers system. It was noticed that the results obtained for the dyslalic subjects (children with pronunciation disorders) from the experimental group improved significantly due to the computer-based therapeutic program, compared to the subjects in the control group (who underwent classical therapy). The mean, the standard deviations and the standard error of the mean obtained for the independent speech variable in dyslalic children (control group and experimental group) are presented in Figure 7.</p>
New Ideas for Brain Modelling 4-Figure 4. LHS relates to neuron binding ensemble mass, with central column activated. RHS relates to hierarchy, with a direct mapping. The two red lines show where the ensemble is missing and so it needs to be learned. The blue lines show extra neurons from the hierarchy back to the ensemble, but can be removed as error. The other paired black squares represent where the patterns match and can oscillate together.
<p>This paper continues the research that considers a new cognitive model based strongly on the human brain, last updated in Greer (2016). In particular, it considers figure 4 of that paper (Figure below) and how it might be useful in practice. The paper also describes some new methods in the areas of image processing and behaviour simulation. The image processing introduces a most classical form of pattern cross-referencing, while the behaviour equations used feedback for a memory-type of cross-referencing. The work is all based on earlier research by the author and the new additions are intended to fit in with the overall design. For image processing, a grid-like structure is used with ‘full linking’, if you like. Each cell in the classifier grid stores a list of all other cells it gets associated with and this is used as the learned image that new input is compared with. For the behaviour metric, a new prediction equation is suggested, as part of a simulation, that uses feedback and history to dynamically determine its current state and course of action. While the new methods are from widely different topics, both can be compared with the binary-analog type of interface that is the main focus of the paper. Sensory input may be static and binary, but cross- references result in variable comparisons that make the input more dynamic. It is suggested that the simplest of linking between a tree and ensemble can explain neural binding and variable signal strengths.</p>
Data Set for Article "Verification-Aided Debugging: An Interactive Web-Service for Exploring Error Witnesses", Proc. CAV'16
<p>This is the description of the supplementary archive of example interactive reports for the approach described in the article "Verification-Aided Debugging: An Interactive Web-Service for Exploring Error Witnesses", Proc. CAV'16.</p> <p>This archive contains a static snapshot of our system that allows the reader to<br> a) experience the features of our web-service without relying on its online availability and<br> b) reproduce the bug reports displayed in this static snapshot by validating the provided witnesses against the source code and the corresponding specifications using CPAchecker.</p> <p>The witness database is available at:<br> static/index.html<br> The supplied verification tasks can be found at:<br> static/programs/<br> The supplied error witnesses are grouped by their corresponding verification tasks and can be found at:<br> static/witnesses/<br> The software verifier CPAchecker is placed at:<br> CPAchecker/</p> <p>To browse the witness database and explore the supplied error reports, we recommend using the Firefox web browser,<br> because not all features of our bug reports are guaranteed to be available in other browsers.</p> <p>Like the supplementary archive originally provided to the reviewers, this witness database contains only a small selection of the witnesses harvested from the "Competition on Software Verification 2016", because we do not want to burden the reader with an enormous amount of data that likely is not relevant for understanding the concepts. Also, error witnesses produced by some competition candidates that were not even syntactically correct were removed, because they do not add any value to the evaluation. However, the full data is still available online via our web service, for example, the list of witnesses for a verification task can be requested by computing the SHA-1 hash of the verification task's source code and submitting the following query:<br> http://vcloud.sosy-lab.org/webclient/master/witness?inputFile=<program-hash><br> The resulting JSON data contains all hashes of witnesses stored for the given program.<br> A witness stored in the database can be requested via its SHA-1 hash by submitting the following query:<br> https://vcloud.sosy-lab.org/webclient/files/<hash><br> All verification tasks are available at the SV-COMP repository:<br> https://github.com/dbeyer/sv-benchmarks<br> If you use verification tasks from the repository and are interested in validating witnesses produced for SV-COMP '16,<br> please use the 'svcomp16' tag, because the tasks and their hashes might have changed since then.</p> <p>You can use CPAchecker to validate a witness for a verification task and generate an error report.<br> First, navigate to the CPAchecker directory:</p> <p> cd CPAchecker/</p> <p>Now, perform the validation by providing the verification task (consisting of specification and program source code) and a witness:</p> <p> scripts/cpa.sh -generateReport -witness-validation \<br> -spec <specification> \<br> <source-code> \<br> -spec <witness></p> <p>For example:</p> <p> scripts/cpa.sh -generateReport -witness-validation \<br> -spec ../static/programs/loop-acceleration/ALL.prp \<br> ../static/programs/loop-acceleration/array_false-unreach-call3.i \<br> -spec ../static/witnesses/loop-acceleration/array_false-unreach-call3.i/a4572a0c1b505b1d1170b7347e48a2a93cb3f4c1</p> <p>The report will be generated in the subdirectory<br> output/report/</p> <p> </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.