# Module: tf_privacy
|---| | [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/privacy/blob/v0.9.0/privacy/__init__.py) | TensorFlow Privacy library. ## Modules [`restart_query`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/restart_query) module: Implements DPQuery interface for restarting the states of another query. [`tree_aggregation`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/tree_aggregation) module: Tree aggregation algorithm. [`v1`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/v1) module: TensorFlow Privacy library v1 imports. ## Classes [`class DNNClassifier`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DNNClassifier): DP version of `tf.estimator.DNNClassifier`. [`class DPModel`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPModel): DP subclass of [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model). [`class DPQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPQuery): Interface for differentially private query mechanisms. [`class DPSequential`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPSequential): DP subclass of [`tf.keras.Sequential`](https://www.tensorflow.org/api_docs/python/tf/keras/Sequential). [`class DiscreteGaussianSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DiscreteGaussianSumQuery): Implements DPQuery for discrete Gaussian sum queries. [`class DistributedDiscreteGaussianSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DistributedDiscreteGaussianSumQuery): Implements DPQuery for discrete distributed Gaussian sum queries. [`class DistributedSkellamSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DistributedSkellamSumQuery): Implements DPQuery interface for discrete distributed sum queries. [`class GaussianSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/GaussianSumQuery): Implements DPQuery interface for Gaussian sum queries. [`class GenericDPAdagradOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/GenericDPAdagradOptimizer): Differentially private subclass of class `tf.keras.optimizers.legacyAdagrad`. [`class GenericDPAdamOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/GenericDPAdamOptimizer): Differentially private subclass of class `tf.keras.optimizers.legacyAdam`. [`class GenericDPSGDOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/GenericDPSGDOptimizer): Differentially private subclass of class `tf.keras.optimizers.legacySGD`. [`class NestedQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/NestedQuery): Implements DPQuery interface for structured queries. [`class NoPrivacyAverageQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/NoPrivacyAverageQuery): Implements DPQuery interface for an average query with no privacy. [`class NoPrivacyQuantileEstimatorQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/NoPrivacyQuantileEstimatorQuery): Iterative process to estimate target quantile of a univariate distribution. [`class NoPrivacySumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/NoPrivacySumQuery): Implements DPQuery interface for a sum query with no privacy. [`class NormalizedQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/NormalizedQuery): `DPQuery` for queries with a `DPQuery` numerator and fixed denominator. [`class QAdaClipTreeResSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/QAdaClipTreeResSumQuery): `DPQuery` for tree aggregation queries with adaptive clipping. [`class QuantileAdaptiveClipSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/QuantileAdaptiveClipSumQuery): `DPQuery` for Gaussian sum queries with adaptive clipping. [`class QuantileEstimatorQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/QuantileEstimatorQuery): DPQuery to estimate target quantile of a univariate distribution. [`class RestartQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/restart_query/RestartQuery): `DPQuery` for `SumAggregationDPQuery` with a `reset_state` function. [`class SumAggregationDPQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/SumAggregationDPQuery): Base class for DPQueries that aggregate via sum. [`class TreeCumulativeSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/TreeCumulativeSumQuery): Returns private cumulative sums by clipping and adding correlated noise. [`class TreeRangeSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/TreeRangeSumQuery): Implements dp_query for accurate range queries using tree aggregation. [`class TreeResidualSumQuery`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/TreeResidualSumQuery): Implements DPQuery for adding correlated noise through tree structure. [`class VectorizedDPKerasAdagradOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/VectorizedDPKerasAdagradOptimizer): Vectorized differentially private subclass of given class [`tf.keras.optimizers.Adagrad`](https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/experimental/Adagrad). [`class VectorizedDPKerasAdamOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/VectorizedDPKerasAdamOptimizer): Vectorized differentially private subclass of given class [`tf.keras.optimizers.Adam`](https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/Adam). [`class VectorizedDPKerasSGDOptimizer`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/VectorizedDPKerasSGDOptimizer): Vectorized differentially private subclass of given class [`tf.keras.optimizers.SGD`](https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/experimental/SGD). ## Functions [`DPFTRLTreeAggregationOptimizer(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPFTRLTreeAggregationOptimizer): Returns a `DPOptimizerClass` `cls` using the `TreeAggregationQuery`. [`DPKerasAdagradOptimizer(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPKerasAdagradOptimizer): Returns a `DPOptimizerClass` `cls` using the `GaussianSumQuery`. [`DPKerasAdamOptimizer(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPKerasAdamOptimizer): Returns a `DPOptimizerClass` `cls` using the `GaussianSumQuery`. [`DPKerasSGDOptimizer(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/DPKerasSGDOptimizer): Returns a `DPOptimizerClass` `cls` using the `GaussianSumQuery`. [`compute_dp_sgd_privacy(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/compute_dp_sgd_privacy): Compute epsilon based on the given hyperparameters. [`compute_dp_sgd_privacy_statement(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/compute_dp_sgd_privacy_statement): Produces a privacy report summarizing the DP guarantee. [`compute_rdp_single_tree(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/compute_rdp_single_tree): Computes RDP of the Tree Aggregation Protocol for a single tree. [`compute_rdp_tree_restart(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/compute_rdp_tree_restart): Computes RDP of the Tree Aggregation Protocol for Gaussian Mechanism. [`compute_zcdp_single_tree(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/compute_zcdp_single_tree): Computes zCDP of the Tree Aggregation Protocol for a single tree. [`linearly_separable_labeled_examples(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/linearly_separable_labeled_examples): Generates num_examples labeled examples using separator given by weights. [`logistic_dpsgd(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/logistic_dpsgd): Trains and validates private logistic regression model via DP-SGD. [`logistic_objective_perturbation(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/logistic_objective_perturbation): Trains and validates differentially private logistic regression model. [`make_dp_model_class(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/make_dp_model_class): Given a subclass of [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model), returns a DP-SGD version of it. [`make_gaussian_query_optimizer_class(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/make_gaussian_query_optimizer_class): Returns a differentially private optimizer using the `GaussianSumQuery`. [`make_keras_generic_optimizer_class(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/make_keras_generic_optimizer_class): Returns a differentially private (DP) subclass of `cls`. [`make_keras_optimizer_class(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/make_keras_optimizer_class): Returns a differentially private optimizer using the `GaussianSumQuery`. [`make_vectorized_keras_optimizer_class(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/make_vectorized_keras_optimizer_class): Given a subclass of [`tf.keras.optimizers.Optimizer`](https://www.tensorflow.org/api_docs/python/tf/keras/optimizers/Optimizer), returns a vectorized DP-SGD subclass of it. [`single_layer_softmax_classifier(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/single_layer_softmax_classifier): Trains a single layer neural network classifier with softmax activation. [`synthetic_linearly_separable_data(...)`](https://www.tensorflow.org/responsible_ai/privacy/api_docs/python/tf_privacy/synthetic_linearly_separable_data): Generates synthetic train and test data for logistic regression.


| ## Other Members || |---|---| | **version** | `'0.9.0'` |