SecML
latest

User Guide

  • 1. Machine Learning
    • 1.1. Training of Classifiers and Visualization of Results
      • 1.1.1. Creation and visualization of a simple 2D dataset
      • 1.1.2. Training of classifiers
      • 1.1.3. Visualization of the decision regions of the classifiers
      • 1.1.4. Training other classifiers
    • 1.2. Neural Networks with PyTorch
      • 1.2.1. Classifying blobs
  • 2. Adversarial Machine Learning
    • 2.1. Evasion Attacks against Machine Learning
      • 2.1.1. Crafting Adversarial Examples
      • 2.1.2. Security evaluation of a classifier
    • 2.2. Transferability of Evasion Attacks
      • 2.2.1. Generation of the Adversarial Examples
      • 2.2.2. Analysis of Transferability
    • 2.3. Poisoning Attacks against Machine Learning models
      • 2.3.1. Generation of Poisoning Samples
    • 2.4. Evasion and Poisoning Attacks on MNIST dataset
      • 2.4.1. Training of the classifier
      • 2.4.2. Evasion attack with MNIST dataset
      • 2.4.3. Poisoning attack with MNIST dataset
    • 2.5. Evasion Attacks against Neural Networks on MNIST dataset
      • 2.5.1. Crafting Evasion Attacks
    • 2.6. Evasion Attacks on ImageNet
      • 2.6.1. Load the pretrained model
      • 2.6.2. Load and classify an image
      • 2.6.3. Run the attack
    • 2.7. Deep Neural Rejection
      • 2.7.1. Dataset creation
      • 2.7.2. Training DNR
      • 2.7.3. Attacking DNR
      • 2.7.4. Plotting attack results
    • 2.8. Using cleverhans within SecML
      • 2.8.1. Training the model
      • 2.8.2. Preparing the attacks
      • 2.8.3. Running the attacks
    • 2.9. Testing attacks against RobustBench models
      • 2.9.1. Load RobustBench models inside SecML
      • 2.9.2. Computing evasion attacks
    • 2.10. Using Foolbox attack classes within SecML
      • 2.10.1. Training the model
      • 2.10.2. Crafting the Adversarial Examples
        • 2.10.2.1. Projected Gradient Descent (L2)
        • 2.10.2.2. Carlini-Wagner Attack
      • 2.10.3. Using the generic wrapper
      • 2.10.4. Crafting Adversarial Example on the MNIST Dataset
  • 3. Explaining Machine Learning
    • 3.1. Feature-based explanations
      • 3.1.1. Training of the classifier
      • 3.1.2. Compute the explanations
      • 3.1.3. Visualize results
    • 3.2. Prototype-based explanation
      • 3.2.1. Training of the classifier
      • 3.2.2. Compute the influential training prototypes
      • 3.2.3. Visualize results
  • 4. Applications
    • 4.1. Evasion Attacks on ImageNet (Computer Vision)
      • 4.1.1. Load data
      • 4.1.2. Run the attack
      • 4.1.3. Visualize and check the attack optimization
    • 4.2. Android Malware Detection
      • 4.2.1. Training and testing the detector
      • 4.2.2. Explaining the decisions
      • 4.2.3. Crafting Adversarial Examples
      • 4.2.4. Security Evaluation

Core & Data Structures

  • secml.core
    • CCreator
    • attr_utils
    • constants
    • decorators
    • exceptions
    • type_utils
  • secml.array
    • CArray
    • array_utils
  • secml.data
    • secml.data.loader
      • CDataLoader
      • CDataLoaderCIFAR
      • CDataLoaderICubWorld
      • CDataLoaderImgClients
      • CDataLoaderImgFolders
      • CDataLoaderLFW
      • CDataLoaderMNIST
      • CDataLoaderPyTorch
      • CDataLoaderSkLearn
      • CDataLoaderSvmLight
      • CDataLoaderTorchDataset
      • loader_utils
    • secml.data.selection
      • CPrototypesSelector
      • CPSBorder
      • CPSCenter
      • CPSKMedians
      • CPSRandom
      • CPSSpanning
    • secml.data.splitter
      • CDataSplitter
      • CDataSplitterKFold
      • CDataSplitterLabelKFold
      • CDataSplitterOpenWorldKFold
      • CDataSplitterShuffle
      • CDataSplitterStratifiedKFold
      • CTrainTestSplit
      • CChronologicalSplitter
    • CDataset
    • CDatasetHeader
    • CDatasetPyTorch
    • data_utils

Machine Learning

  • secml.ml
    • secml.ml.classifiers
      • secml.ml.classifiers.multiclass
        • CClassifierMulticlass
        • CClassifierMulticlassOVA
        • CClassifierMulticlassOVO
      • secml.ml.classifiers.secure
        • CClassifierSecSVM
      • secml.ml.classifiers.reject
        • CClassifierReject
        • CClassifierRejectThreshold
        • CClassifierDNR
      • secml.ml.classifiers.loss
        • CLoss
        • CLossCrossEntropy
        • CLossEpsilonInsensitive
        • CLossHinge
        • CLossLogistic
        • CLossSquare
        • CSoftmax
      • secml.ml.classifiers.regularizer
        • CRegularizer
        • CRegularizerElasticNet
        • CRegularizerL1
        • CRegularizerL2
      • CClassifier
      • CClassifierLinear
      • CClassifierSkLearn
      • CClassifierDecisionTree
      • CClassifierKNN
      • CClassifierLogistic
      • CClassifierNearestCentroid
      • CClassifierRandomForest
      • CClassifierRidge
      • CClassifierSGD
      • CClassifierSVM
      • CClassifierDNN
      • CClassifierPyTorch
      • clf_utils
    • secml.ml.features
      • secml.ml.features.normalization
        • CNormalizer
        • CNormalizerLinear
        • CNormalizerMeanStd
        • CNormalizerMinMax
        • CNormalizerUnitNorm
        • CNormalizerDNN
      • secml.ml.features.reduction
        • CReducer
        • CLDA
        • CPCA
      • CPreProcess
    • secml.ml.kernels
      • CKernel
      • CKernelChebyshevDistance
      • CKernelEuclidean
      • CKernelHistIntersect
      • CKernelLaplacian
      • CKernelLinear
      • CKernelPoly
      • CKernelRBF
    • secml.ml.peval
      • secml.ml.peval.metrics
        • CMetric
        • CMetricAccuracy
        • CMetricAUC
        • CMetricAUCWMW
        • CMetricConfusionMatrix
        • CMetricF1
        • CMetricMAE
        • CMetricMSE
        • CMetricPartialAUC
        • CMetricPrecision
        • CMetricRecall
        • CRoc
        • CMetricTestError
        • CMetricTPRatFPR
        • CMetricFNRatFPR
        • CMetricTHatFPR
        • CMetricTPRatTH
        • CMetricFNRatTH
      • CPerfEvaluator
      • CPerfEvaluatorXVal
      • CPerfEvaluatorXValMulticlass
    • secml.ml.stats
      • CDensityEstimation
      • CDistributionGaussian
  • secml.adv
    • secml.adv.attacks
      • secml.adv.attacks.evasion
        • CAttackEvasion
        • CAttackEvasionPGD
        • CAttackEvasionPGDLS
        • CAttackEvasionPGDExp
        • CAttackEvasionCleverhans
        • secml.adv.attacks.evasion.foolbox
          • CAttackEvasionFoolbox
          • CFoolboxBasicIterative
          • CFoolboxL2CarliniWagner
          • CFoolboxL2DDN
          • CFoolboxDeepfool
          • CFoolboxEAD
          • CFoolboxFGM
          • CFoolboxPGD
      • secml.adv.attacks.poisoning
        • CAttackPoisoning
        • CAttackPoisoningLogisticRegression
        • CAttackPoisoningRidge
        • CAttackPoisoningSVM
      • CAttack
    • secml.adv.seceval
      • CSecEval
      • CSecEvalData
  • secml.optim
    • secml.optim.function
      • CFunction
      • CFunctionLinear
      • CFunctionQuadratic
      • CFunctionRosenbrock
      • CFunctionThreeHumpCamel
      • CFunctionBeale
      • CFunctionMcCormick
    • secml.optim.optimizers
      • secml.optim.optimizers.line_search
        • CLineSearch
        • CLineSearchBisect
        • CLineSearchBisectProj
      • COptimizer
      • COptimizerPGD
      • COptimizerPGDLS
      • COptimizerPGDExp
      • COptimizerScipy
    • secml.optim.constraints
      • CConstraint
      • CConstraintBox
      • CConstraintL1
      • CConstraintL2
  • secml.model_zoo
    • load_model

Explanation

  • secml.explanation
    • CExplainer
    • CExplainerGradient
    • CExplainerGradientInput
    • CExplainerIntegratedGradients
    • CExplainerInfluenceFunctions

Visualization

  • secml.figure
    • CFigure
    • CPlot

Utilities

  • secml.parallel
    • parfor
  • secml.utils
    • CLog
    • c_file_manager
    • pickle_utils
    • download_utils
    • dict_utils
    • list_utils
    • mixed_utils
  • secml.settings
  • secml.testing
    • CUnitTest

References

  • UPDATE GUIDES
    • From 0.8.* to 0.9
      • 1. Configuration file
      • 2. Deprecations
  • CHANGELOG
    • v0.14.1 (22/04/2021)
      • Fixed (3 changes)
      • Removed & Deprecated (1 change)
      • Documentation (2 changes)
    • v0.14 (23/03/2021)
      • Requirements (5 changes)
      • Added (3 changes)
      • Improved (3 changes)
      • Fixed (8 changes)
      • Removed & Deprecated (2 changes)
      • Documentation (10 changes)
    • v0.13 (24/07/2020)
      • Requirements (1 change)
      • Added (4 changes)
      • Improved (12 changes)
      • Changed (6 changes)
      • Fixed (10 changes)
      • Removed & Deprecated (2 changes)
      • Documentation (10 changes)
    • v0.12 (11/03/2020)
      • Added (7 changes)
      • Improved (8 changes)
      • Changed (3 changes)
      • Fixed (10 changes)
      • Removed & Deprecated (5 changes)
      • Documentation (5 changes)
    • v0.11.2 (07/01/2020)
      • Requirements (1 change)
      • Fixed (7 changes)
      • Documentation (2 changes)
    • v0.11.1 (18/12/2019)
      • Fixed (3 changes)
    • v0.11 (02/12/2019)
      • Requirements (2 changes)
      • Added (5 changes)
      • Improved (6 changes)
      • Changed (2 changes)
      • Fixed (8 changes)
      • Removed & Deprecated (4 changes)
      • Documentation (4 changes)
    • v0.10 (29/10/2019)
      • Requirements (1 change)
      • Added (4 changes)
      • Improved (6 changes)
      • Changed (4 changes)
      • Fixed (4 changes)
      • Removed & Deprecated (5 changes)
      • Documentation (5 changes)
    • v0.9 (11/10/2019)
      • Improved (1 change)
      • Changed (1 change)
      • Fixed (1 change)
      • Documentation (2 changes)
      • Deprecations (3 changes)
    • v0.8.1 (05/09/2019)
      • Documentation (2 changes)
    • v0.8 (06/08/2019)
  • ROADMAP

Developer's Guide

  • Contributing
    • Submitting a bug report or feature request
      • Bug report
      • Feature request
  • Code Contributions
    • Development Installation
      • Editable Installation
    • Merge request checklist
    • Coding guidelines
      • Code style
      • Documentation style
      • Packages
      • Modules
      • Classes
      • Tests
  • Extending SecML
    • Abstract Base Classes
    • Creating new extensions
      • CClassifier
        • Standard classifiers (CClassifier)
        • DNN backends (CClassifierDNN)
SecML
  • »
  • Search

© Copyright 2021, PRALab - Pattern Recognition and Applications Lab & Pluribus One s.r.l.. Revision fd905aa5.

Built with Sphinx using a theme provided by Read the Docs.