Abstract: Embodiments herein facilitate a rule-based anonymization of an original dataset. The system may include a processor including a data privacy evaluator and a rules engine. The data privacy evaluator may receive at least one anonymized dataset corresponding to a predefined strategy of anonymization. The at least one anonymized dataset may include a variation from the original dataset by at least one of a privacy metric and a consistency metric. The data privacy evaluator may evaluate the at least one anonymized dataset and may generate a final output value based on a first output and a second output. The processor may assess the final output value with respect to a predefined threshold through the rules engine. If the final output value may be equal or higher than the predefined threshold, the system may permit an access to the anonymized dataset.

Abstract: The disclosed system and method provide an artificial intelligence (AI) model trained using clustering and reinforcement learning. Data in a dataset can be loaded for de-identification, along with data scope answers. The data can be audited, and once audited, the audited data and the data scope answers can be provided to a strategy recommendation engine including the trained AI model. The engine can determine a cluster corresponding to the dataset and assesses strategies for data de-identification based on the determined cluster. The strategies can be ranked and provided as output, providing the ability to better de-identify the dataset by indicating which techniques will be the most effective. Additionally, the system and method can automatically implement a top-ranked strategy satisfying certain criteria as a determined optimal approach for data de-identification. Clustering and reinforcement learning may efficiently and automatically glean information from unlabeled data.

Abstract: Methods, systems and apparatus, including computer programs encoded on computer storage medium, for machine unlearning. In one aspect a method includes receiving a request to remove a client dataset from a machine learning model, the model being associated with noise sensitivities determined during training of the model on respective client datasets including the client; and in response to receiving the request: identifying, from stored noise sensitivities of the client, a most recent training iteration that produced a noise sensitivity that is below a predetermined threshold that is based on a noise standard deviation and predefined target privacy parameters; updating parameters of the model, comprising adding noise to model parameters for the most recent training iteration; and performing subsequent iterations of training of the model, wherein the model is initialized with the updated parameters and the subsequent iterations train the model on datasets excluding the dataset owned by the client.

Abstract: Systems and methods for smart incentivization for achieving collaborative machine learning are disclosed. A system receives local model parameters from plurality of client devices in a network, for global model corresponding to collaborative machine learning. The system determines an optimum score for each client device using pre-trained Conditional Variational Auto Encoder (CVAE), based on local model parameter. The system computes contribution score for each client device by determining relative distance value of optimum score corresponding to each client device with optimum score corresponding to another client device from the plurality of client devices, and a global model optimum score of global model. The system updates global model with local model parameter received from the selected set of client devices of the plurality of client devices corresponding to good class, average class, and bad class.

Abstract: Examples of a system for usage control of a personal data are described. The system may obtain an input image including a first face of a person. Further, the system may compute a usage control matrix based on the input image, at least one usage control function, and a predefined criteria. The pre-defined criteria may be associated with at least one of: a data usage policy, a face matching probability related to matching of the face present in the input image, and a face recognition probability related to a recognition of an identity of the person. Furthermore, by using the input image and the usage control matrix, the system may transform the input image to a usage-controlled image. Furthermore, the system may verify a matching of the face present in the input image with a second face present in the usage-controlled image.

Abstract: A system and method for facilitating a rule-based anonymization of an original dataset is disclosed. The system may include a processor including a data privacy evaluator and a rules engine. The data privacy evaluator may receive at least one anonymized dataset corresponding to a predefined strategy of anonymization. The at least one anonymized dataset may include a variation from the original dataset by at least one of a privacy metric and a consistency metric. The data privacy evaluator may evaluate the at least one anonymized dataset and may generate a final output value based on a first output and a second output. The processor may assess the final output value with respect to a predefined threshold through the rules engine. If the final output value may be equal or higher than the predefined threshold, the system may permit an access to the anonymized dataset.

Abstract: Examples of a system for usage control of a personal data are described. The system may obtain an input image including a first face of a person. Further, the system may compute a usage control matrix based on the input image, at least one usage control function, and a predefined criteria. The pre-defined criteria may be associated with at least one of: a data usage policy, a face matching probability related to matching of the face present in the input image, and a face recognition probability related to a recognition of an identity of the person. Furthermore, by using the input image and the usage control matrix, the system may transform the input image to a usage-controlled image. Furthermore, the system may verify a matching of the face present in the input image with a second face present in the usage-controlled image.

Abstract: A device may receive, from an augmented reality device, speech data identifying augmented reality interactions with an augmented reality robot. The device may receive, from a camera, video data identifying movements of a real robot based on the augmented reality interactions with the augmented reality robot. The device may process the speech data to obtain first action data identifying first actions that the real robot is to perform. The device may process the video data to obtain second action data identifying second actions that the real robot is to perform. The device may process the first action data and the second action data to generate pseudocode. The device may transform the pseudocode into code. The device may cause the real robot to execute the code, wherein executing the code causes the real robot to perform the first actions and the second actions.