Automatically adjusting data access policies in data analytics
From a first model parameter, an autoencoder network is generated. A reconstruction error for the autoencoder network is measured, the reconstruction error comprising a difference between an input to the autoencoder network and a corresponding output from the autoencoder network, the input to the autoencoder network comprising a portion of an initial set of data. The reconstruction error and a confidence score corresponding to a complexity level of the autoencoder network are aggregated into a level of difficulty score of the autoencoder network. From the level of difficulty score and an initial data access policy level corresponding to the initial set of data, a derived data access policy level corresponding to the initial data access policy level is generated, the derived data access policy level enforcing access to a transformed set of data generated by applying a transformation to the initial set of data.
1 . A computer-implemented method comprising:
generating, from a first model parameter, an autoencoder network;
generating, by applying a transformation to an input, and corresponding to a portion of an initial set of data of the input, an output from the autoencoder network;
producing a reconstructed input by performing an inverse transformation operation on the output from the autoencoder network;
measuring a reconstruction error for the autoencoder network, the measuring the reconstruction error comprising computing a difference between the input to the autoencoder network and the reconstructed input;
aggregating, into a level of difficulty score of the autoencoder network, the reconstruction error and a complexity level of the autoencoder network, the level of difficulty score representing difficulty of using the output to recover the input;
generating, from the level of difficulty score and an initial data access policy level enforcing access to the initial set of data, a derived data access policy level corresponding to the initial data access policy level, the derived data access policy level enforcing access to an output set of data of the output from the autoencoder network wherein the generating comprises:
lowering, relative to the initial data access policy level, the derived data access policy level, the lowering performed responsive to determining that the level of difficulty score is above a predetermined score; and
setting the derived data access policy level for the output set of data set corresponding to a portion of the output from the autoencoder network.
2 . The computer-implemented method of claim 1 , further comprising:
training, using a training subset of the initial set of data, the autoencoder network.
3 . The computer-implemented method of claim 2 , wherein the training is performed to minimize a reconstruction error of the autoencoder network.
4 . The computer-implemented method of claim 2 , wherein the training is performed to minimize a difference between an output of an encoder portion of the autoencoder network and a transformed set of training data, the transformed set of training data generated by applying the transformation to the training subset.
5 . The computer-implemented method of claim 1 , further comprising:
measuring, for the autoencoder network, the complexity level.
6 . The computer-implemented method of claim 1 , further comprising:
generating, from the level of difficulty score, a set of model parameters, a second model parameter in the set of model parameters comprising a variation from the first model parameter;
generating, from the set of model parameters, a set of autoencoder networks;
measuring a model-specific reconstruction error of each autoencoder network in the set of autoencoder networks, the model-specific reconstruction error comprising a difference between an input to an autoencoder network in the set of autoencoder networks and a corresponding output from the autoencoder network in the set of autoencoder networks, the input to the autoencoder network in the set of autoencoder networks comprising the portion of the initial set of data; and
aggregating, into a level of difficulty score of the set of autoencoder networks, the model-specific reconstruction error of each autoencoder network and a complexity level corresponding to each model-specific reconstruction error.
7 . The computer-implemented method of claim 1 , wherein the first model parameter comprises a number of hidden layers in an encoder portion of the autoencoder network and a number of hidden layers in a decoder portion of the autoencoder network.
8 . The computer-implemented method of claim 1 , wherein the first model parameter comprises a number of dimensions in an output of an encoder portion of the autoencoder network.
9 . The computer-implemented method of claim 1 , wherein the generating comprises setting, responsive to determining that the level of difficulty score is below a second predetermined score, the derived data access policy level to the initial data access policy level.
10 . A computer program product for automatically adjusting a data access policy, the computer program product comprising: one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions comprising:
program instructions to generate, from a first model parameter, an autoencoder network;
program instructions to generate, by applying a transformation to an input, and corresponding to a portion of an initial set of data of the input, an output from the autoencoder network;
program instructions to produce a reconstructed input by performing an inverse transformation operation on the output from the autoencoder network;
program instructions to measure a reconstruction error for the autoencoder network, the measuring the reconstruction error comprising computing a difference between an input to the autoencoder network and the reconstructed input;
program instructions to aggregate, into a level of difficulty score of the autoencoder network, the reconstruction error and a complexity level of the autoencoder network, the level of difficulty score representing difficulty of using the output to recover the input;
program instructions to generate, from the level of difficulty score and an initial data access policy level enforcing access to the initial set of data, a derived data access policy level corresponding to the initial data access policy level, the derived data access policy level enforcing access to an output set of data of the output from the autoencoder network wherein the generating comprises:
lowering, relative to the initial data access policy level, the derived data access policy level, the lowering performed responsive to determining that the level of difficulty score is above a predetermined score; and
setting the derived data access policy level for the output set of data set corresponding to a portion of the output from the autoencoder network.
11 . The computer program product of claim 10 , further comprising:
program instructions to train, using a training subset of the initial set of data, the autoencoder network, wherein the training is performed to minimize a reconstruction error of the autoencoder network.
12 . The computer program product of claim 11 , wherein the training is performed to minimize a difference between an output of an encoder portion of the autoencoder network and a transformed set of training data, the transformed set of training data generated by applying the transformation to the training subset.
13 . The computer program product of claim 10 , further comprising:
program instructions to measure, for the autoencoder network, the complexity level.
14 . The computer program product of claim 10 , further comprising:
program instructions to generate, from the level of difficulty score, a set of model parameters, a second model parameter in the set of model parameters comprising a variation from the first model parameter;
program instructions to generate, from the set of model parameters, a set of autoencoder networks;
program instructions to measure a model-specific reconstruction error of each autoencoder network in the set of autoencoder networks, the model-specific reconstruction error comprising a difference between an input to an autoencoder network in the set of autoencoder networks and a corresponding output from the autoencoder network in the set of autoencoder networks, the input to the autoencoder network in the set of autoencoder networks comprising the portion of the initial set of data; and
program instructions to aggregate, into a level of difficulty score of the set of autoencoder networks, the model-specific reconstruction error of each autoencoder network and a complexity level corresponding to each model-specific reconstruction error.
15 . The computer program product of claim 10 , wherein the first model parameter comprises a number of hidden layers in an encoder portion of the autoencoder network and a number of hidden layers in a decoder portion of the autoencoder network.
16 . The computer program product of claim 10 , wherein the first model parameter comprises a number of dimensions in an output of an encoder portion of the autoencoder network.
17 . The computer program product of claim 10 , wherein the stored program instructions are stored in the at least one of the one or more storage media of a local data processing system, and wherein the stored program instructions are transferred over a network from a remote data processing system.
18 . The computer program product of claim 10 , wherein the stored program instructions are stored in the at least one of the one or more storage media of a server data processing system, and wherein the stored program instructions are downloaded over a network to a remote data processing system for use in a computer readable storage device associated with the remote data processing system.
19 . The computer program product of claim 10 , wherein the computer program product is provided as a service in a cloud environment.
20 . A computer system comprising one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, the stored program instructions comprising:
program instructions to generate, from a first model parameter, an autoencoder network;
program instructions to generate, by applying a transformation to an input, and corresponding to a portion of an initial set of data of the input, an output from the autoencoder network;
program instructions to produce a reconstructed input by performing an inverse transformation operation on the output from the autoencoder network;
program instructions to measure a reconstruction error for the autoencoder network, the measuring the reconstruction error comprising computing a difference between an input to the autoencoder network and the reconstructed input;
program instructions to aggregate into a level of difficulty score of the autoencoder network, the reconstruction error and a complexity level of the autoencoder network, the level of difficulty score representing difficulty of using the output to recover the input;
program instructions to generate, from the level of difficulty score and an initial data access policy level enforcing access to the initial set of data, a derived data access policy level corresponding to the initial data access policy level, the derived data access policy level enforcing access to an output set of data of the output from the autoencoder network wherein the generating comprises:
lowering, relative to the initial data access policy level, the derived data access policy level, the lowering performed responsive to determining that the level of difficulty score is above a predetermined score; and
setting the derived data access policy level for the output set of data set corresponding to a portion of the output from the autoencoder network.