Abstract: Methods, apparatus, systems and articles of manufacture for distributed use of a machine learning model are disclosed. An example edge device includes a model partitioner to partition a machine learning model received from an aggregator into private layers and public layers. A public model data store is implemented outside of a trusted execution environment of the edge device. The model partitioner is to store the public layers in the public model data store. A private model data store is implemented within the trusted execution environment. The model partitioner is to store the private layers in the private model data store.

Abstract: In one example an apparatus comprises a memory and a processor to create, from a first deep neural network (DNN) model, a first plurality of DNN models, generate a first set of adversarial examples that are misclassified by the first plurality of deep neural network (DNN) models, determine a first set of activation path differentials between the first plurality of adversarial examples, generate, from the first set of activation path differentials, at least one composite adversarial example which incorporates at least one intersecting critical path that is shared between at least two adversarial examples in the first set of adversarial examples, and use the at least one composite adversarial example to generate a set of inputs for a subsequent training iteration of the DNN model. Other examples may be described.

Abstract: Methods, apparatus, systems and articles of manufacture for distributed use of a machine learning model are disclosed. An example edge device includes a model partitioner to partition a machine learning model received from an aggregator into private layers and public layers. A public model data store is implemented outside of a trusted execution environment of the edge device. The model partitioner is to store the public layers in the public model data store. A private model data store is implemented within the trusted execution environment. The model partitioner is to store the private layers in the private model data store.

Abstract: Methods, apparatus, systems and articles of manufacture for federated training of a neural network using trusted edge devices are disclosed. An example system includes an aggregator device to aggregate model updates provided by one or more edge devices. The one or more edge devices to implement respective neural networks, and provide the model updates to the aggregator device. At least one of the edge devices to implement the neural network within a trusted execution environment.

Abstract: Embodiments discussed herein may be generally directed to systems and techniques to generate a quality score based on an observation and an action caused by an actor agent during a testing phase. Embodiments also include determining a temporal difference between the quality score and a previous quality score based on a previous observation and a previous action, determining whether the temporal difference exceeds a threshold value, and generating an attack indication in response to determining the temporal difference exceeds the threshold value.

Abstract: A method comprises receiving in a governor device, from a plurality of data owner devices, metadata for one or more datasets maintained by the plurality of data owner devices, registering the metadata for the one or more datasets with the governor device, in response to a request from an aggregator, providing at least a portion of the metadata for the one or more datasets to the aggregator, receiving, from the aggregator, a compute plan to be implemented by the plurality of data owner devices, distributing at least a portion of the compute plan to the plurality of data owner devices, in response to receiving, from the plurality of data owner devices, a verification report and a certification for an enclave, binding the enclave to a host device, and providing the compute plan to the plurality of data owner devices.

Abstract: The present invention discloses a secure ML pipeline to improve the robustness of ML models against poisoning attacks and utilizing data provenance as a tool. Two components are added to the ML pipeline, a data quality pre-processor, which filters out untrusted training data based on provenance derived features and an audit post-processor, which localizes the malicious source based on training dataset analysis using data provenance.

Abstract: Embodiments discussed herein may be generally directed to systems and techniques to generate a quality score based on an observation and an action caused by an actor agent during a testing phase. Embodiments also include determining a temporal difference between the quality score and a previous quality score based on a previous observation and a previous action, determining whether the temporal difference exceeds a threshold value, and generating an attack indication in response to determining the temporal difference exceeds the threshold value.

Abstract: Embodiments discussed herein may be generally directed to systems and techniques to generate a quality score based on an observation and an action caused by an actor agent during a testing phase. Embodiments also include determining a temporal difference between the quality score and a previous quality score based on a previous observation and a previous action, determining whether the temporal difference exceeds a threshold value, and generating an attack indication in response to determining the temporal difference exceeds the threshold value.

Abstract: In one example an apparatus comprises a memory and a processor to create, from a first deep neural network (DNN) model, a first plurality of DNN models, generate a first set of adversarial examples that are misclassified by the first plurality of deep neural network (DNN) models, determine a first set of activation path differentials between the first plurality of adversarial examples, generate, from the first set of activation path differentials, at least one composite adversarial example which incorporates at least one intersecting critical path that is shared between at least two adversarial examples in the first set of adversarial examples, and use the at least one composite adversarial example to generate a set of inputs for a subsequent training iteration of the DNN model. Other examples may be described.

Abstract: Methods, apparatus, systems and articles of manufacture for distributed use of a machine learning model are disclosed. An example edge device includes a model partitioner to partition a machine learning model received from an aggregator into private layers and public layers. A public model data store is implemented outside of a trusted execution environment of the edge device. The model partitioner is to store the public layers in the public model data store. A private model data store is implemented within the trusted execution environment. The model partitioner is to store the private layers in the private model data store.

Abstract: Methods, apparatus, systems and articles of manufacture for federated training of a neural network using trusted edge devices are disclosed. An example system includes an aggregator device to aggregate model updates provided by one or more edge devices. The one or more edge devices to implement respective neural networks, and provide the model updates to the aggregator device. At least one of the edge devices to implement the neural network within a trusted execution environment.

Abstract: Embodiments discussed herein may be generally directed to systems and techniques to generate a quality score based on an observation and an action caused by an actor agent during a testing phase. Embodiments also include determining a temporal difference between the quality score and a previous quality score based on a previous observation and a previous action, determining whether the temporal difference exceeds a threshold value, and generating an attack indication in response to determining the temporal difference exceeds the threshold value.

Abstract: Technologies are provided in embodiments to manage an authentication confirmation score. Embodiments are configured to identify, in absolute session time, a beginning time and an ending time of an interval of an active user session on a client. Embodiments are also configured to determine a first value representing a first subset of a set of prior user sessions, where the prior user sessions of the first subset were active for at least as long as the beginning time. Embodiments can also determine a second value representing a second subset of the set of prior user sessions, where the prior user sessions of the second subset were active for at least as long as the ending time. Embodiments also determine, based on the first and second values, a decay rate for the authentication confidence score of the active user session. In some embodiments, the set is based on context attributes.

Abstract: Technologies are provided in embodiments to manage an authentication confirmation score. Embodiments are configured to identify, in absolute session time, a beginning time and an ending time of an interval of an active user session on a client. Embodiments are also configured to determine a first value representing a first subset of a set of prior user sessions, where the prior user sessions of the first subset were active for at least as long as the beginning time. Embodiments can also determine a second value representing a second subset of the set of prior user sessions, where the prior user sessions of the second subset were active for at least as long as the ending time. Embodiments also determine, based on the first and second values, a decay rate for the authentication confidence score of the active user session. In some embodiments, the set is based on context attributes.

Abstract: Technologies are provided in embodiments to manage an authentication confirmation score. Embodiments are configured to identify, in absolute session time, a beginning time and an ending time of an interval of an active user session on a client. Embodiments are also configured to determine a first value representing a first subset of a set of prior user sessions, where the prior user sessions of the first subset were active for at least as long as the beginning time. Embodiments can also determine a second value representing a second subset of the set of prior user sessions, where the prior user sessions of the second subset were active for at least as long as the ending time. Embodiments also determine, based on the first and second values, a decay rate for the authentication confidence score of the active user session. In some embodiments, the set is based on context attributes.

Abstract: Technologies are provided in embodiments to manage an authentication confirmation score. Embodiments are configured to identify, in absolute session time, a beginning time and an ending time of an interval of an active user session on a client. Embodiments are also configured to determine a first value representing a first subset of a set of prior user sessions, where the prior user sessions of the first subset were active for at least as long as the beginning time. Embodiments can also determine a second value representing a second subset of the set of prior user sessions, where the prior user sessions of the second subset were active for at least as long as the ending time. Embodiments also determine, based on the first and second values, a decay rate for the authentication confidence score of the active user session. In some embodiments, the set is based on context attributes.

Abstract: Technologies are provided in embodiments to manage an authentication confirmation score. Embodiments are configured to identify, in absolute session time, a beginning time and an ending time of an interval of an active user session on a client. Embodiments are also configured to determine a first value representing a first subset of a set of prior user sessions, where the prior user sessions of the first subset were active for at least as long as the beginning time. Embodiments can also determine a second value representing a second subset of the set of prior user sessions, where the prior user sessions of the second subset were active for at least as long as the ending time. Embodiments also determine, based on the first and second values, a decay rate for the authentication confidence score of the active user session. In some embodiments, the set is based on context attributes.

Abstract: A system is provided to determine whether to re-authenticate a user based on identification parameter measurements of low power sensors. According to an embodiment of the invention, a system may include a processor that includes analysis logic to determine whether to re-authenticate the user based on parameter values received from at least one of one or more agents. The system may also include authentication logic to re-authenticate the user that includes a confirmation of whether the user is authenticated based on input received from one or more authentication sensors. Other embodiments are described and claimed.

Abstract: A method of enforcing a virtual corporate boundary may include a client device requesting sensitive content from a network site on a server device responsive to a user's interaction with the client device. The server device can determine whether the user and/or client device are permitted to access the sensitive content. A secure element on the client device can establish a session key between the server device and the client device. The server device can render the sensitive content and send it to the client device, which can display the content to the user.