61 skills found · Page 1 of 3
jscad / OpenJSCAD.orgJSCAD is an open source set of modular, browser and command line tools for creating parametric 2D and 3D designs with JavaScript code. It provides a quick, precise and reproducible method for generating 3D models, and is especially useful for creating ready-to-print 3D designs.
freelunchtheorem / Conditional Density EstimationPython and torch-based package implementing various parametric and nonparametric methods for conditional density estimation
himanshub1007 / Alzhimers Disease Prediction Using Deep Learning# AD-Prediction Convolutional Neural Networks for Alzheimer's Disease Prediction Using Brain MRI Image ## Abstract Alzheimers disease (AD) is characterized by severe memory loss and cognitive impairment. It associates with significant brain structure changes, which can be measured by magnetic resonance imaging (MRI) scan. The observable preclinical structure changes provides an opportunity for AD early detection using image classification tools, like convolutional neural network (CNN). However, currently most AD related studies were limited by sample size. Finding an efficient way to train image classifier on limited data is critical. In our project, we explored different transfer-learning methods based on CNN for AD prediction brain structure MRI image. We find that both pretrained 2D AlexNet with 2D-representation method and simple neural network with pretrained 3D autoencoder improved the prediction performance comparing to a deep CNN trained from scratch. The pretrained 2D AlexNet performed even better (**86%**) than the 3D CNN with autoencoder (**77%**). ## Method #### 1. Data In this project, we used public brain MRI data from **Alzheimers Disease Neuroimaging Initiative (ADNI)** Study. ADNI is an ongoing, multicenter cohort study, started from 2004. It focuses on understanding the diagnostic and predictive value of Alzheimers disease specific biomarkers. The ADNI study has three phases: ADNI1, ADNI-GO, and ADNI2. Both ADNI1 and ADNI2 recruited new AD patients and normal control as research participants. Our data included a total of 686 structure MRI scans from both ADNI1 and ADNI2 phases, with 310 AD cases and 376 normal controls. We randomly derived the total sample into training dataset (n = 519), validation dataset (n = 100), and testing dataset (n = 67). #### 2. Image preprocessing Image preprocessing were conducted using Statistical Parametric Mapping (SPM) software, version 12. The original MRI scans were first skull-stripped and segmented using segmentation algorithm based on 6-tissue probability mapping and then normalized to the International Consortium for Brain Mapping template of European brains using affine registration. Other configuration includes: bias, noise, and global intensity normalization. The standard preprocessing process output 3D image files with an uniform size of 121x145x121. Skull-stripping and normalization ensured the comparability between images by transforming the original brain image into a standard image space, so that same brain substructures can be aligned at same image coordinates for different participants. Diluted or enhanced intensity was used to compensate the structure changes. the In our project, we used both whole brain (including both grey matter and white matter) and grey matter only. #### 3. AlexNet and Transfer Learning Convolutional Neural Networks (CNN) are very similar to ordinary Neural Networks. A CNN consists of an input and an output layer, as well as multiple hidden layers. The hidden layers are either convolutional, pooling or fully connected. ConvNet architectures make the explicit assumption that the inputs are images, which allows us to encode certain properties into the architecture. These then make the forward function more efficient to implement and vastly reduce the amount of parameters in the network. #### 3.1. AlexNet The net contains eight layers with weights; the first five are convolutional and the remaining three are fully connected. The overall architecture is shown in Figure 1. The output of the last fully-connected layer is fed to a 1000-way softmax which produces a distribution over the 1000 class labels. AlexNet maximizes the multinomial logistic regression objective, which is equivalent to maximizing the average across training cases of the log-probability of the correct label under the prediction distribution. The kernels of the second, fourth, and fifth convolutional layers are connected only to those kernel maps in the previous layer which reside on the same GPU (as shown in Figure1). The kernels of the third convolutional layer are connected to all kernel maps in the second layer. The neurons in the fully connected layers are connected to all neurons in the previous layer. Response-normalization layers follow the first and second convolutional layers. Max-pooling layers follow both response-normalization layers as well as the fifth convolutional layer. The ReLU non-linearity is applied to the output of every convolutional and fully-connected layer.  The first convolutional layer filters the 224x224x3 input image with 96 kernels of size 11x11x3 with a stride of 4 pixels (this is the distance between the receptive field centers of neighboring neurons in a kernel map). The second convolutional layer takes as input the (response-normalized and pooled) output of the first convolutional layer and filters it with 256 kernels of size 5x5x48. The third, fourth, and fifth convolutional layers are connected to one another without any intervening pooling or normalization layers. The third convolutional layer has 384 kernels of size 3x3x256 connected to the (normalized, pooled) outputs of the second convolutional layer. The fourth convolutional layer has 384 kernels of size 3x3x192 , and the fifth convolutional layer has 256 kernels of size 3x3x192. The fully-connected layers have 4096 neurons each. #### 3.2. Transfer Learning Training an entire Convolutional Network from scratch (with random initialization) is impractical[14] because it is relatively rare to have a dataset of sufficient size. An alternative is to pretrain a Conv-Net on a very large dataset (e.g. ImageNet), and then use the ConvNet either as an initialization or a fixed feature extractor for the task of interest. Typically, there are three major transfer learning scenarios: **ConvNet as fixed feature extractor:** We can take a ConvNet pretrained on ImageNet, and remove the last fully-connected layer, then treat the rest structure as a fixed feature extractor for the target dataset. In AlexNet, this would be a 4096-D vector. Usually, we call these features as CNN codes. Once we get these features, we can train a linear classifier (e.g. linear SVM or Softmax classifier) for our target dataset. **Fine-tuning the ConvNet:** Another idea is not only replace the last fully-connected layer in the classifier, but to also fine-tune the parameters of the pretrained network. Due to overfitting concerns, we can only fine-tune some higher-level part of the network. This suggestion is motivated by the observation that earlier features in a ConvNet contains more generic features (e.g. edge detectors or color blob detectors) that can be useful for many kind of tasks. But the later layer of the network becomes progressively more specific to the details of the classes contained in the original dataset. **Pretrained models:** The released pretrained model is usually the final ConvNet checkpoint. So it is common to see people use the network for fine-tuning. #### 4. 3D Autoencoder and Convolutional Neural Network We take a two-stage approach where we first train a 3D sparse autoencoder to learn filters for convolution operations, and then build a convolutional neural network whose first layer uses the filters learned with the autoencoder.  #### 4.1. Sparse Autoencoder An autoencoder is a 3-layer neural network that is used to extract features from an input such as an image. Sparse representations can provide a simple interpretation of the input data in terms of a small number of \parts by extracting the structure hidden in the data. The autoencoder has an input layer, a hidden layer and an output layer, and the input and output layers have same number of units, while the hidden layer contains more units for a sparse and overcomplete representation. The encoder function maps input x to representation h, and the decoder function maps the representation h to the output x. In our problem, we extract 3D patches from scans as the input to the network. The decoder function aims to reconstruct the input form the hidden representation h. #### 4.2. 3D Convolutional Neural Network Training the 3D convolutional neural network(CNN) is the second stage. The CNN we use in this project has one convolutional layer, one pooling layer, two linear layers, and finally a log softmax layer. After training the sparse autoencoder, we take the weights and biases of the encoder from trained model, and use them a 3D filter of a 3D convolutional layer of the 1-layer convolutional neural network. Figure 2 shows the architecture of the network. #### 5. Tools In this project, we used Nibabel for MRI image processing and PyTorch Neural Networks implementation.
d909b / Drnet💉📈 Dose response networks (DRNets) are a method for learning to estimate individual dose-response curves for multiple parametric treatments from observational data using neural networks.
wbuntine / Topic ModelsTopic modelling software using non-parametric methods
ShelvanLee / XFEM# XFEM_Fracture2D ### Description This is a Matlab program that can be used to solve fracture problems involving arbitrary multiple crack propagations in a 2D linear-elastic solid based on the principle of minimum potential energy. The extended finite element method is used to discretise the solid continuum considering cracks as discontinuities in the displacement field. To this end, a strong discontinuity enrichment and a square-root singular crack tip enrichment are used to describe each crack. Several crack growth criteria are available to determine the evolution of cracks over time; apart from the classic maximum tension (or hoop-stress) criterion, the minimum total energy criterion and the local symmetry criterion are implemented implicitly with respect to the discrete time-stepping. ### Key features * *Fast:* The stiffness matrix and the force vector (i.e. the equations' system) and the enrichment tracking data structures are updated at each time step only with respect to the changes in the fracture topology. This ultimately results in the major part of the computational expense in the solution to the linear system of equations rather than in the post-processing of the solution or in the assembly and updating of the equations. As Matlab offers fast and robust direct solvers, the computational times are reasonably fast. * *Robust.* Suitable for multiple crack propagations with intersections. Furthermore, the stress intensity factors are computed robustly via the interaction integral approach (with the inclusion of the terms to account for crack surface pressure, residual stresses or strains). The minimum total energy criterion and the principle of local symmetry are implemented implicitly in time. The energy release rates are computed based on the stiffness derivative approach using algebraic differentiation (rather than finite differencing of the potential energy). On the other hand, the crack growth direction based on the local symmetry criterion is determined such that the local mode-II stress intensity factor vanishes; the change in a crack tip kink angle is approximated using the ratio of the crack tip stress intensity factors. * *Easy to run.* Each job has its own input files which are independent form those of all other jobs. The code especially lends itself to running parametric studies. Various results can be saved relating to the fracture geometry, fracture mechanics parameters, and the elastic fields in the solid domain. Extensive visualisation library is available for plotting results. ### Instructions 1. Get started by running the demo to showcase some of the capabilities of the program and to determine if it can be useful for you. At the Matlab's command line enter: ```Matlab >> RUN_JOBS.m ``` This will execute a series of jobs located inside the *jobs directory* `./JOBS_LIBRARY/`. These jobs do not take very long to execute (around 5 minutes in total). 2. Subsequently, you can pick one of the jobs inside `./JOBS_LIBRARY/` by defining the job title: ```Matlab >> job_title = 'several_cracks/edge/vertical_tension' ``` 3. Then you can open all the relevant scripts for this job as follows: ```Matlab >> open_job ``` The following input scripts for the *job* will be open in the Matlab's editor: 1. `JOB_MAIN.m`: This is the job's main script. It is called when executing `RUN_JOB` (or `RUN_JOBS`) and acts like a wrapper. Notably, it can serve as a convenient interface to run parametric studies and to save intermediate simulation results. 2. `Input_Scope.m`: This defines the scope of the simulation. From which crack growth criteria to use, to what to compute and what results to show via plots and/or movies. To put it simply, the script is a bunch of "switches" that tell the program what the user wants to be done. 3. `Input_Material.m`: Defines the material's elastic properties in different regions or layers (called "phases") of the computational domain. Moreover, it defines the fracture toughness of the material (assumed to be constant in all material phases). 4. `Input_Crack.m`: Defines the initial crack geometry. 5. `Input_BC.m`: Defines boundary conditions, such as displacements, tractions, crack surface pressure (assumed to be constant in all cracks), body loads (e.g. gravity, pre-stress or pre-strain). 6. `Mesh_make.m`: In-house structured mesh generator for rectangular domains using either linear triangle or bilinear quadrilateral elements. It is possible to mesh horizontal layers using different mesh sizes. 7. `Mesh_read.m`: Gmsh based mesh reader for version-1 mesh files. Of course you can use your own mesh reader provided the output variables are of the correct format (see later). 8. `Mesh_file.m`: Specifies the mesh input file (.msh). At the moment, only Gmsh mesh files of version-1 are allowed. ### Mesh_file.m A mesh file needs to be able to output the following data or variables: * `mNdCrd`: Node coordinates, size = `[nNdStd, 2]` * `mLNodS`: Element connectivities, size = `[nElemn,nLNodS]` * `vElPhz`: Element material phase (or region) ID's, size = `[nElemn,1]` * `cBCNod`: cell of boundary nodes, cell size = `{nBound,1}`, cell element size = `[nBnNod,2]` Example mesh files are located in `./JOBS_LIBRARY/`. Gmsh version-1 file format is described [here](http://www.manpagez.com/info/gmsh/gmsh-2.4.0/gmsh_60.php). ### Additional notes * global variables are defined in `.\Routines_AuxInput\Declare_Global.m` * External libraries are `.\Other_Libs\distmesh` and `.\Other_Libs\mesh2d` ### References Two external meshing libraries are used for the local mesh refinement and remeshing at the crack tip during crack propagation or prior to a crack intersection with another crack or with a boundary of the domain. Specifically, these libraries, which are located in `.\Other_Libs\`, are the following: * [*mesh2d*](https://people.sc.fsu.edu/~jburkardt/m_src/mesh2d/mesh2d.html) by Darren Engwirda * [*distmesh*](http://persson.berkeley.edu/distmesh/) by Per-Olof Persson and Gilbert Strang. ### Issues and Support For support or questions please email [sutula.danas@gmail.com](mailto:sutula.danas@gmail.com). ### Authors Danas Sutula, University of Luxembourg, Luxembourg. If you find this code useful, we kindly ask that you consider citing us. * [Minimum energy multiple crack propagation](http://hdl.handle.net/10993/29414)
yikang-li / PasteGANAn pytorch implementation of our NeurIPS paper of PasteGAN: A Semi-Parametric Method to Generate Image from Scene Graph
derrynknife / SurPyvalA Python package for survival analysis. The most flexible survival analysis package available. SurPyval can work with arbitrary combinations of observed, censored, and truncated data. SurPyval can also fit distributions with 'offsets' with ease, for example the three parameter Weibull distribution.
stk-kriging / StkThe STK is a (not so) Small Toolbox for Kriging. Its primary focus is on the interpolation/regression technique known as kriging, which is very closely related to Splines and Radial Basis Functions, and can be interpreted as a non-parametric Bayesian method using a Gaussian Process (GP) prior.
issaz / Signature Regime DetectionCode accompanying the paper "Pathwise methods for non-parametric online market regime detection and regime clustering for multidimensional and non-Markovian data"
csatzky / Forecasting Realized Volatility Using Supervised LearningTraditionally, volatility is modeled using parametric models. This project focuses on predicting EUR/USD volatility using more flexible, machine-learning methods.
dracula-ybp / Class Shape Transformation MethodAn airfoil parametric method
FrancescoCrecchi / Multiscale Parametric T SNEESANN20 paper code repository. This package is a perplexity-free extension of Parametric t-SNE dimensionality reduction method implemented in `Keras` and compatible with `Scikit-learn`.
HannesPetrenz / RALMPC Linear Uncertain SystemsThis repository contains the MATLAB implementation for the Robust Adaptive Learning Model Predictive Control (RALMPC) framework proposed in the paper Robust MPC for uncertain linear systems. The method is designed for linear systems with parametric uncertainties and additive disturbances performing iterative tasks.
reddyprasade / Machine Learning Interview PreparationPrepare to Technical Skills Here are the essential skills that a Machine Learning Engineer needs, as mentioned Read me files. Within each group are topics that you should be familiar with. Study Tip: Copy and paste this list into a document and save to your computer for easy referral. Computer Science Fundamentals and Programming Topics Data structures: Lists, stacks, queues, strings, hash maps, vectors, matrices, classes & objects, trees, graphs, etc. Algorithms: Recursion, searching, sorting, optimization, dynamic programming, etc. Computability and complexity: P vs. NP, NP-complete problems, big-O notation, approximate algorithms, etc. Computer architecture: Memory, cache, bandwidth, threads & processes, deadlocks, etc. Probability and Statistics Topics Basic probability: Conditional probability, Bayes rule, likelihood, independence, etc. Probabilistic models: Bayes Nets, Markov Decision Processes, Hidden Markov Models, etc. Statistical measures: Mean, median, mode, variance, population parameters vs. sample statistics etc. Proximity and error metrics: Cosine similarity, mean-squared error, Manhattan and Euclidean distance, log-loss, etc. Distributions and random sampling: Uniform, normal, binomial, Poisson, etc. Analysis methods: ANOVA, hypothesis testing, factor analysis, etc. Data Modeling and Evaluation Topics Data preprocessing: Munging/wrangling, transforming, aggregating, etc. Pattern recognition: Correlations, clusters, trends, outliers & anomalies, etc. Dimensionality reduction: Eigenvectors, Principal Component Analysis, etc. Prediction: Classification, regression, sequence prediction, etc.; suitable error/accuracy metrics. Evaluation: Training-testing split, sequential vs. randomized cross-validation, etc. Applying Machine Learning Algorithms and Libraries Topics Models: Parametric vs. nonparametric, decision tree, nearest neighbor, neural net, support vector machine, ensemble of multiple models, etc. Learning procedure: Linear regression, gradient descent, genetic algorithms, bagging, boosting, and other model-specific methods; regularization, hyperparameter tuning, etc. Tradeoffs and gotchas: Relative advantages and disadvantages, bias and variance, overfitting and underfitting, vanishing/exploding gradients, missing data, data leakage, etc. Software Engineering and System Design Topics Software interface: Library calls, REST APIs, data collection endpoints, database queries, etc. User interface: Capturing user inputs & application events, displaying results & visualization, etc. Scalability: Map-reduce, distributed processing, etc. Deployment: Cloud hosting, containers & instances, microservices, etc. Move on to the final lesson of this course to find lots of sample practice questions for each topic!
ThangLe-duc / FEINNIn this study, we propose a novel deep learning model named as the Finite-element-informed neural network (FEI-NN), inspired from finite element method (FEM) for parametric simulation of static problems in structural mechanics.
mustafaseisa / RegimechangeNon-parametric method for estimating regime change in bivariate time series setting.
nifm-gin / DB QMRI👐 This package includes 3 MR fingerprinting methods to reconstruct parametric maps: standard dictionary-based matching and dictionary-based learning using a statistical or a neural network approach.
asalarpour / Point GNOfficial WACV 2025 code for Point-GN: A non-parametric, training-free method for 3D point cloud classification using Gaussian Positional Encoding (GPE). No training, no parameters, state-of-the-art accuracy.
cadema-PoliTO / RECOptThe repository contains a routine that optimizes the operation of a PV system with energy storage for fixed or variable (parametric) sizes for both of them, in the context of collective self-consumption and energy communities in Italy. PV production data are to be provided by the user (PVGIS database can be used), while consumption profiles are generated for an aggregate of households using probabilistic methods.
