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.
73
datasets available to search
ShareScore release 0.7.1
Dataset results
73 results for “Network robustness”
Robustness assessment of a C++ implementation of a quantized (int8) version 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" (<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, and 8-bit integers for weights and activations instead of floating-point.</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><i><strong>relu2</strong></i>: Rectified Linear Unit (6@14x14).</li><li>max2: Subsampling buy max pooling (6@7x7).</li><li><i><strong>fc1</strong></i>: Fully connected (294, 147)</li><li><i><strong>fc2</strong></i>: Fully connected (147, 10)</li></ul><p>The fault hypotheses for this work include the occurrence of:</p><ul><li><strong>BF</strong>: single, double-adjacent and triple-adjacent 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 (int8)</li><li><strong>zw</strong>: zero point of the weights (int8)</li><li><strong>b</strong>: biases (int32)</li><li><strong>z</strong>: zero point (int8)</li><li><strong>m</strong>: m (int32)</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><p>In addition, the raw data have been lightly processed to obtain global data related to the particular bits and parameters affected by the faults, and the obtained failure modes.</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/bit_flip</i> folder: Prediction obtained for all the images considered in the workload in presence of double adjacent bit-flip 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/bit_flip</i> folder: Prediction obtained for all the images considered in the workload in presence of triple adjacent bit-flip 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 golden_run.csv file.</p><p>After that, one fault injection experiment was executed for each bit of each element of each parameter of the CNN.</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.zw, 3 - conv1.m, 4 - conv1.b, 5 - conv1.z, 6 - conv2.w, 7 - conv2.zw, 8 - conv2.m, 9 - conv2.b, 10 - conv2.z, 11 - fc1.w, 12 - fc1.zw, 13 - fc1.m, 14 - fc.b, 15 - fc1.z, 16 - fc2.w, 17 - fc2.zw, 18 - fc2.m, 19 - fc2.b, 20 - fc2.z)</li><li><strong>ELEMID</strong>: Integer number identiying the element of the parameter affected by the fault (-1 - No fault, [0-2] - {conv1.b, conv1.m, conv1.zw}, [0-74] - conv1.w, 0 - conv1.z, [0-5] - {conv2.b, conv2.m, conv2.zw}, [0-149] - conv2.w, 0 - {conv1.z, conv2.z, fc1.z, fc2.z}, [0-146] - {fc1.b, fc1.m, fc1.zw}, [0-43217] - fc1.w, [0-9] - {fc2.b, fc2.m, fc2.zw}, [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>OUTPUT</strong>: 10 integer numbers provided by the CNN as output after processing the image. The highest value identifies the selected category for classification.</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>
Robust Method for Property Prediction via Artificial Neural Networks: Incorporating Key Structural Features for Carbon Dioxide – Ionic Liquid Mixtures
<p>This Dataset comprises two sub-sets of information:</p> <ul> <li>Database and Results of the work present in the paper "Robust Method for Property Prediction via Artificial Neural Networks: Incorporating Key Structural Features for Carbon Dioxide – Ionic Liquid Mixtures" published in The Journal of Physical Chemistry B (https://doi.org/10.1021/acs.jpcb.4c04432).</li> <li>Sample of the code used, in order to reproduce any of the results presented above. This can be found in the previous version of this Dataset (v1.0 https://zenodo.org/records/11216901)</li> </ul> <p> </p> <p>Regarding the sample code, an example for all ANN Models used in this work is provided. This includes the three models used:</p> <ol> <li>One based only on Critical Properties of Ionic Liquids (CRT Model)</li> <li>One based only on Structural Properties of Ionic Liquids (STR Model)</li> <li>One combination of the previous models, taking into account both Critical and Structural Properties (COMB Model)</li> </ol> <p>In this manner, it is possible to observe the differences between the performance of the different models, either through statiscal analysis or using graphical representation. This allows for the benchmarking to be done in a more concise way.</p>
Generated Data for the Manuscript "Nonideality-Aware Training for Accurate and Robust Low-Power Memristive Neural Networks"
<p>The file contains data generated and referred to in the text and the figures of the manuscript.</p>
Coping with Collapse: Functional Robustness of Coral-Reef Fish Network to Simulated Cascade Extinction
<p>Data set, codes and results related to the article "Coping with Collapse: Functional Robustness of Coral-Reef Fish Network to Simulated Cascade Extinction", accepted in the periodic Global Change Biology. Stored are the full results of site occupancy models fitted to fish data, with coral and turf algae cover as predictor variables (results published in Luza et al. 2022, Scientific Reports), and the results of the present article. The RData also contains site coordinates, and the fish traits used in trait-based analyzes.</p>
Emergence of cooperative bistability and robustness of gene regulatory networks
<p>Simulation and analysis source codes and obtained data set for "Emergence of cooperative bistability and robustness of gene regulatory network" (<a href="https://doi.org/10.1371/journal. pcbi.1007969">PLoS Comput Biol 16 (2020) e1007969</a> and <a href="https://arxiv.org/abs/1907.12030">arXiv:1907.12030</a>) by Nagata and Kikuchi. </p> <p>Source codes and figures are compiled in Jupyter notebook. Detailed discription of data sets is found in "readme.txt" file.</p> <p> </p>
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>
A Robust Generative Adversarial Network Approach for Climate Downscaling and Weather Generation
<h1>Dataset Description for "A Robust Generative Adversarial Network Approach for Climate Downscaling and Weather Generation"</h1> <p>This dataset accompanies the research paper titled <strong>"A Robust Generative Adversarial Network Approach for Climate Downscaling and Weather Generation"</strong>, currently under review for the AGU Journal JAMES. The study introduces a novel Regional Climate Model (RCM) emulator focusing on high-resolution climate downscaling for the New Zealand region. For additional insights and access to the codebase utilized in this research, please refer to our <a href="https://github.com/nram812/A-Robust-Generative-Adversarial-Network-Approach-for-Climate-Downscaling" target="_new">GitHub repository</a>.</p> <h2>Aims</h2> <p>Our study's overarching goal was to assess the effectiveness of Generative Adversarial Networks (GANs) in a climate downscaling context and is structured around two aims. The first aim of our study is to examine whether GANs can overcome several important limitations of regression-based climate downscaling algorithms (i.e. underestimating the magnitude of extreme events). The second and most important aim of our study is to assess the robustness GAN performance to different training hyperparameters. Our robustness assessment thoroughly scrutinizes GANs for their application in climate downscaling contexts, ensuring that they can learn and capture regional climate processes</p> <h2>Geographic Focus</h2> <p>Our research focuses only on the New Zealand Region (165°E-184°W, 33°S-51°S).</p> <p> </p> <h2>Data Overview</h2> <h3>Training and Evaluation Data</h3> <p>The training data used in this study (for our RCM emulator) only spans the historical period of simulation. It comprises daily accumulated precipitation as the primary target variable, alongside large-scale predictor variables. </p> <ul> <li> <p><strong>Resolution:</strong> The target variable is presented at a 12km resolution, reflecting the highest resolution face of RCM for the New Zealand region. Predictor variables are coarsened to a 1.5-degree resolution from original CCAM outputs using conservative interpolation. </p> </li> <li> <p><strong>Period Coverage:</strong></p> <ul> <li>Training Data: 1960-2014</li> <li>Validation Data: 1986-2005</li> </ul> </li> <li> <p><strong>Models:</strong></p> <ul> <li>Training on: ACCESS-CM2</li> <li>Validated on: EC-Earth3, NorESM2-MM</li> </ul> </li> </ul> <h3>File Structure</h3> <ul> <li> <p><strong>Training Data:</strong></p> <ul> <li>Target/Ground Truth (Y): <code>predictor_ACCESS-CM2_hist.nc</code></li> <li>Predictor (X): <code>pr_ACCESS-CM2_hist.nc</code></li> </ul> </li> <li> <p><strong>Evaluation Data:</strong></p> <ul> <li><strong>NorESM2-MM:</strong> <ul> <li>Target (Y): <code>NorESM2-MM_historical_precip_compressed.nc</code></li> <li>Predictor (X): <code>NorESM2-MM_histupdated_compressed.nc</code></li> </ul> </li> <li><strong>EC-Earth3:</strong> <ul> <li>Target: <code>EC-Earth3_historical_precip_compressed.nc</code></li> <li>Predictor: <code>EC-Earth3_histupdated_compressed.nc</code></li> </ul> </li> </ul> </li> </ul> <h2>Methodological Insights</h2> <ul> <li> <p><strong>Regional Climate Model</strong>, Our Regional Climate Model training data is from the Conformal Cubic Atmospheric Model (CCAM) which is a global non-hydrostatic atmospheric model renowned for its variable-resolution cubic grid. . For more information about CCAM, please see the following <a href="https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2023JD038530">paper</a>.</p> </li> <li> <p><strong>Predictor and Target Variables:</strong> Daily-averaged large-scale prognostic variables, including zonal wind, meridional wind, temperature, and specific humidity, are employed as predictors at the 500mb and 850mb pressure levels. These are normalized (see the GitHub repository for the mean and standard deviation fields). Precipitation is taken as is from CCAM and accumulated for each given day. Static predictors are also used in our model, which is stored in a GitHub repository.</p> </li> <li> <p><strong>Training Framework:</strong> Our dataset benefits from the "perfect framework" training strategy, which uses CCAM-coarsened predictor variables. For more information about the perfect and imperfect training frameworks, see the following <a title="review" href="https://journals.ametsoc.org/view/journals/aies/3/2/AIES-D-23-0066.1.xml">review</a></p> </li> </ul>
Group and individual social network metrics are robust to changes in resource distribution in experimental populations of forked fungus beetles
<p>Social interactions drive many important ecological and evolutionary processes. It is therefore essential to understand the intrinsic and extrinsic factors that underlie social patterns. A central tenet of the field of behavioral ecology is the expectation that the distribution of resources shapes patterns of social interactions.</p> <p>We combined experimental manipulations with social network analyses to ask how patterns of resource distribution influence complex social interactions.</p> <p>We experimentally manipulated the distribution of an essential food and reproductive resource in semi-natural populations of forked fungus beetles (Bolitotherus cornutus). We aggregated resources into discrete clumps in half of the populations and evenly dispersed resources in the other half. We then observed social interactions between individually marked beetles. Half-way through the experiment, we reversed the resource distribution in each population, allowing us to control any demographic or behavioral differences between our experimental populations. At the end of the experiment, we compared individual and group social network characteristics between the two resource distribution treatments.</p> <p>We found a statistically significant but quantitatively small effect of resource distribution on individual social network position and detected no effect on group social network structure. Individual connectivity (individual strength) and individual cliquishness (local clustering coefficient) increased in environments with clumped resources, but this difference explained very little of the variance in individual social network position. Individual centrality (individual betweenness) and measures of overall social structure (network density, average shortest path length, and global clustering coefficient) did not differ between environments with dramatically different distributions of resources.</p> <p>Our results illustrate that the resource environment, despite being fundamental to our understanding of social systems, does not always play a central role in shaping social interactions. Instead, our results suggests that sex differences and temporally fluctuating environmental conditions may be more important in determining patterns of social interactions.</p>
Robust quantum dots charge autotuning using neural network uncertainty - Output data
<p>Outputs of the model training and the offline autotuning experiments presented in the paper: "<em>Robust quantum dots charge autotuning using neural network uncertainty</em>".</p> <p>For convenience, the results are splitted in several zipped files:</p> <ul> <li><strong>run_outputs_light.zip</strong>: contains only settings and results text files (sufficient for compiling result tables).</li> <li><strong>run_outputs_full_scan.zip</strong>: contains complete scan of the diagrams (for qualitative analyse)</li> <li><strong>run_outputs_part<N>.zip</strong>: contains all autotuning simulation output, grouped by seed (images and video output types might vary between seeds)</li> </ul> <p>Each folder in the zipped files represent a run that includes:</p> <ul> <li>log file</li> <li>plots / images</li> <li>run settings</li> <li>performance results</li> <li>pytorch model parameters</li> </ul> <p>See README.txt for more information about the file strucutre.</p>
Fig. 1 in Testing the robustness of transmission network models to predict ectoparasite loads. One lizard, two ticks and four years
Fig. 1. Transmission networks generated with (a) a short time window of infection; and (b) a long time window of infection, from the GPS location data of the lizards in the study population in 2010. Nodes represent individual lizards and edges between nodes are directed towards the lizard that is at risk of infection. The edges are weighted as described in the main text and the thicker the line the more weight is associated with that edge.
Improving Robustness of Deep Neural Networks for Aerial Navigation by Incorporating Input Uncertainty
<p>CEA covered the scenario of UAV navigation through a set of gates with unknown locations using a DNN-based navigation model. The implemented navigation model uses two DL components (perception and control), and uses (Bayesian) uncertainty estimation methods to capture the uncertainty (confidence) associated with the predictions of each component. The safety requirements in the UAV mission are related to the confidence (uncertainty) associated with the predictions from these components. CEA observed and analysed the uncertainty from each DNN under specific situations that can pose a risk to the UAV mission. Then, the observations were used to define STL rules to track the confidence of the DNN-based navigation system. Finally, mitigation behaviours (e.g., hover, land, DNN-based autonomous flight) are triggered depending on the satisfaction (or violation) of the STL rules. Moreover, the proposed ROS2-based architecture for safe navigation contributed to the definition and improvement of the COMP4DRONES reference architecture, showing in practice how the proposed safety monitoring architecture relates and integrates with the components from other system functions.</p>
Group and individual social network metrics are robust to changes in resource distribution in experimental populations of forked fungus beetles
Open the record for dataset details and reuse information.
Robust semi-automatic vessel tracing in the human retinal image by an instance segmentation neural network
Open the record for dataset details and reuse information.
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>
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>
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>
A multi-uncertainty-set-based robust transmission expansion planning model using an efficient linear AC network
<p>The file uploaded provides input data for the paper "A multi-uncertainty-set-based robust transmission expansion planning model using an efficient linear AC network".</p>
Network files and Python code used in "Designing a sector-coupled European energy system robust to 60 years of historical weather data"
<p><strong>Description</strong></p> <p>This repository contains data presented in the paper <a href="https://www.nature.com/articles/s41467-024-54853-3" target="_blank" rel="noopener">Designing a sector-coupled European energy system robust to 60 years of historical weather data</a>. It contains the derived metrics (.csv) files from a:</p> <ol> <li>joint capacity and dispatch optimization with weather years (design years) from 1960 to 2021 as input</li> <li>dispatch optimization of the 62 capacity layouts using weather years (operational years) different from the design year.</li> </ol> <p>All results from (1) are found in "Capacity_optimization.zip" and results from (2) are found in "Dispatch_optimization.zip".</p> <p>The resulting network files (both from the capacity and dispatch optimization) are located <a href="https://anon.erda.au.dk/cgi-sid/ls.py?share_id=DuGvDWlkeI">here</a>.</p> <p>We also provide the Python code used to derive the metrics and to create the visualizations included in the paper. This is located in "Jupyter_notebooks". The Jupyter notebooks refer to Python scripts located <a href="https://github.com/ebbekyhl/multi-weather-year-assessment">here</a>.</p> <p><strong>Revisions:</strong></p> <p>This version includes the following additions compared to the previous versions: </p> <ul> <li>Timeseries of nodal loads for all years</li> <li>Timeseries of nodal heat pump Coefficient of Performance (COP) </li> <li>Nodal capacity and hourly capacity factors </li> </ul>
Supplementary Data and Software for "Robust NLoS Localization in 5G mmWave Networks: Data-based Methods and Performance"
<p>The file includes supplementary data for "Robust NLoS Localization in 5G mmWave Networks: Data-based Methods and Performance". If you would like to re-use the software provided please cite the following two items.</p> <p>R. Klus, J. Talvitie, J. Equi, G. Fodor, J. Torsner, and M. Valkama, “Robust NLoS Localization in 5G mmWave Networks: Data-based Methods and<br>Performance,” IEEE Transactions on Vehicular Technology, 2024.</p> <p>R Klus et al. (2024). Supplementary materials for “Robust NLoS Localization in 5G mmWave Networks: Data-based Methods and Performance”. version v1, 25.06.2024, [Online]. Available: https://doi.org/10.5281/zenodo.12204892</p> <p>If you have any questions about this package, please do not hesitate to contact Roman Klus (roman.klus@tuni.fi).</p>
Mapping trait versus species turnover reveals spatiotemporal variation in functional redundancy and network robustness in a plant‐pollinator community
<p>1. Functional overlap among species (redundancy) is considered important in shaping competitive and mutualistic interactions that determine how communities respond to environmental change. Most studies view functional redundancy as static, yet traits within species – which ultimately shape functional redundancy – can vary over seasonal or spatial gradients. We therefore have limited understanding of how trait turnover within and between species could lead to changes in functional redundancy or how loss of traits could differentially impact mutualistic interactions depending on where and when the interactions occur in space and time.</p> <p>2. Using an Arctic bumblebee community as a case study, and 1,277 individual measures from 14 species over three annual seasons, we quantified how inter- and intraspecific body-size turnover compared to species turnover with elevation and over the season. Coupling every individual and their trait with a plant visitation, we investigated how grouping individuals by a morphological trait or by species identity altered our assessment of network structure and how this differed in space and time. Finally, we tested how the sensitivity of the network in space and time differed when simulating extinction of nodes representing either morphological trait similarity or traditional species groups. This allowed us to explore the degree to which trait-based groups increase or decrease interaction redundancy relative to species-based nodes.</p> <p>3. We found that i) groups of taxonomically and morphologically similar bees turn over in space and time independently from each other, with trait turnover being larger over the season; ii) networks composed of nodes representing species versus morphologically similar bees were structured differently; and iii) simulated loss of bee trait groups caused faster coextinction of bumblebee species and flowering plants than when bee taxonomic groups were lost. Crucially, the magnitude of these effects varied in space and time, highlighting the importance of considering spatiotemporal context when studying the relative importance of taxonomic and trait contributions to interaction network architecture.</p> <p>4. Our finding that functional redundancy varies spatiotemporally demonstrates how considering the traits of individuals within networks is needed to understand the impacts of environmental variation and extinction on ecosystem functioning and resilience.</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.