Abstract: Distributed machine learning employs a central fusion server that coordinates the distributed learning process. Preferably, each of set of learning agents that are typically distributed from one another initially obtains initial parameters for a model from the fusion server. Each agent trains using a dataset local to the agent. The parameters that result from this local training (for a current iteration) are then passed back to the fusion server in a secure manner, and a partial homomorphic encryption scheme is then applied. In particular, the fusion server fuses the parameters from all the agents, and it then shares the results with the agents for a next iteration. In this approach, the model parameters are secured using the encryption scheme, thereby protecting the privacy of the training data, even from the fusion server itself.

Abstract: Generating an environment information map from a wireless communication system having directional communication capabilities. A plurality of features are extracted and logged resultant from a communication link attempt at multiple beam directions and training at least one machine learning model based on the plurality of extracted features from the first beam direction and the at least one additional beam directions to infer an environment information map of the area between the first transmitter and the receiver.

Abstract: An output time-series of a cell of a neural network is captured. A subset of a set of data points of the output time-series is consolidated into a singular data point. The singular data point is fitted in a data representation to form a quantified aggregated data point. The quantified aggregated data point is included in an intermediate time-series. Using the intermediate time-series as an input at an intermediate layer of the neural network, an anonymized output time-series is produced from the neural network.

Abstract: Generating an environment information map from a wireless communication system having directional communication capabilities. A plurality of features are extracted and logged resultant from a communication link attempt at multiple beam directions and training at least one machine learning model based on the plurality of extracted features from the first beam direction and the at least one additional beam directions to infer an environment information map of the area between the first transmitter and the receiver.

Abstract: A path computing method, system, and computer program product, include extracting unpleasant data from a database to create a multivariate spatia-temporal density function, collecting a tolerance level of a user, and computing a path for the user based on the tolerance level and the density function.

Abstract: Techniques for performing root cause analysis in dynamic software testing via probabilistic modeling are provided. In one example, a computer-implemented method comprises initializing, by a system operatively coupled to a processor, a threshold value, a defined probability value, and a counter value. The computer-implemented method also includes, in response to determining, by the system, that a probability value assigned to a candidate payload of one or more candidate payloads exceeds the defined probability value, and in response to determining, by the system, that the counter value exceeds the threshold value: determining, by the system, that a match exists between the candidate payload and an input point based on an application of the candidate payload to the input point resulting in a defined condition, wherein the one or more candidate payloads are represented by population data accessed by the system.

Abstract: Method and apparatus for exchanging corpora via a data broker are provided. One example method generally includes receiving, at the data broker from a holder of a first corpus application, a coreset for the first corpus and transmitting the coreset to a set of data providers. The method further includes receiving, from a first data provider of the set of data providers, a value with respect to the coreset of a second corpus associated with the first data provider and transmitting, from the data broker to the holder of the first corpus, the value. The method further includes receiving, at the data broker from the holder of the first corpus, a request to receive the second corpus and receiving the second corpus from the first data provider. The method further includes validating the value of the second corpus and transmitting the second corpus to the holder of the first corpus.

Abstract: A first set of data associated with one or more data stores is received. A distance from a representation of a subset of the first set of data to at least a second representation of another set of data in vector space is identified. In response to the identifying of the distance, the first set of data is anonymized. The anonymizing includes adding noise to at least some of the first set of data.

Abstract: Neural network protection mechanisms are provided. The neural network protection engine receives a pre-trained neural network computer model and forward propagates a dataset through layers of the pre-trained neural network computer model to compute, for each layer of the pre-trained neural network computer model, inputs and outputs of the layer. For at least one layer of the pre-trained neural network computer model, a differentially private distillation operation is performed on the inputs and outputs of the at least one layer to generate modified operational parameters of the at least one layer. The modified operational parameters of the at least one layer obfuscate aspects of an original training dataset used to train the pre-trained neural network computer model, present in original operational parameters of the at least one layer. The neural network protection engine generates a privatized trained neural network model based on the modified operational parameters.

Abstract: A method is provided for transmitter authentication including generating a noise vector using a generative adversarial network generator model, wherein a signature of a first transmitter is embedded into a signal output by the first transmitter based at least on the noise vector; and using the signature to identify the first transmitter.

Abstract: A method (and structure and computer product) to encrypt plaintext data into ciphertext data includes encrypting, using a processor on a computer, plaintext into corresponding ciphertext, using a Property Preserving Encryption (PPE) protocol in which a predefined property is maintained when plaintext values are encrypted into ciphertext values. The predefined property is randomly flipped during encryption to reverse the predefined property in the corresponding ciphertext node. An indication of whether the predefined property has been maintained or reversed is stored as the state of encryption.

Abstract: Methods and systems of determining a data protection level of a dataset are described. In an example, a processor may encode a dataset and generate a network model of the encoded dataset. The processor may sort a set of edges of the network model based on a descending order of costs of the set of edges. The processor may determine a flow for a first edge among the sorted edges, the first edge may be an edge associated with the least cost. The processor may performing the determining of flows for the other edges in accordance with the descending order of the sorted edges. The processor may determine a metric based on the determined flows of the sorted edges and based on the costs of the sorted edges. The processor may compare the metric with a threshold to determine a level of data protection provided by the encoded dataset.

Abstract: Embodiments of the invention are directed to a computer-implemented method of distributed learning using a fusion-based approach. The method includes determining data statistics at each system node of a plurality of system nodes, wherein each system node respectively comprises an artificial intelligence model. The method further includes determining a set of control and coordination instructions for training each artificial intelligence model at each system node of the plurality of system nodes. The method further includes directing an exchange of data between the plurality of system nodes based on the data statistics of each system node of the plurality of system nodes. The method further includes fusing trained artificial intelligence models from the plurality of system nodes into a fused artificial intelligence model, wherein the trained artificial intelligence models are trained using the set of control and coordination instructions.

Abstract: A GAN includes a first device and a second device. A discriminator model in the first device is trained to discriminate samples from a transmitter in the first device from samples from other transmitters, by collaborating by the first device with the second device to train the discriminator model to discriminate between samples from its transmitter and spoofed samples received from a generator model in the second device and to train the generator model in the second device to produce more accurate spoofed samples received by the first device during the training. The training results in a trained discriminator model, which is distributed to another device for use by the other device to discriminate samples received by the other device in order to perform authentication of the transmitter in the first device. The other device performs authentication of the transmitter of the first device using the distributed model.

Abstract: A method for targeted advertisement includes transmitting a pre-filter to the user device, responsive to contextual information from a user device, to determine, using a processor, one or more inferences based on physical browsing information, collected at the user device, in compliance with one or more privacy policies of the user. The method also includes receiving one or more inferences determined by the pre-filter from the user device and transmitting one or more targeted advertisements to the user device based on one or more inferences.

Abstract: A method includes training, using a first set of training data, to produce a machine learning model to generate an output based on an input. In an embodiment, the method includes training, using a second set of training data, to produce a second model to generate the output based on the input. In an embodiment, the method includes receiving a query to explain a decision-making process of the machine learning model. In an embodiment, the method includes producing, in response to the query, an explanation of the decision-making process of the second model.

Abstract: Methods and systems for targeted advertisement include transmitting a pre-filter to a user device, responsive to contextual information supplied by the user device to determine one or more inferences based on physical browsing information, collected at the user device, in compliance with one or more privacy policies of the user. One or more targeted advertisements are determined, using a processor, based on the one or more inferences. The one or more targeted advertisements are transmitted to the user device.

Abstract: Distributed machine learning employs a central fusion server that coordinates the distributed learning process. Preferably, each of set of learning agents that are typically distributed from one another initially obtains initial parameters for a model from the fusion server. Each agent trains using a dataset local to the agent. The parameters that result from this local training (for a current iteration) are then passed back to the fusion server in a secure manner, and a partial homomorphic encryption scheme is then applied. In particular, the fusion server fuses the parameters from all the agents, and it then shares the results with the agents for a next iteration. In this approach, the model parameters are secured using the encryption scheme, thereby protecting the privacy of the training data, even from the fusion server itself.

Abstract: A method is provided for transmitter authentication including generating a noise vector using a generative adversarial network generator model, wherein a signature of a first transmitter is embedded into a signal output by the first transmitter based at least on the noise vector; and using the signature to identify the first transmitter.

Abstract: Techniques for synthesizing security exploits via self-amplifying deep learning are provided. In one example, a computer-implemented method can comprise generating, by a system operatively coupled to a processor, a probabilistic model based on an evaluation of one or more first payloads included in a first group of payloads. The computer implemented method can also comprise determining, by the system, based on the probabilistic model, that at least one first payload from the first group of payloads is invalid. Additionally, the computer implemented method can comprise, generating, by the system, a second group of payloads based on removing the at least one invalid first payload from the first group of payloads.