Abstract: Systems and techniques that facilitate policy-driven views of datasets are provided. In various embodiments, a system can comprise a data component that can access a dataset. In various instances, the system can further comprise a view policy component that can respectively correlate a set of security credentials to a set of view instructions associated with the dataset. In various embodiments, the system can further comprise an input component that can receive as input a first security credential that is listed in the set of security credentials. In various aspects, the system can further comprise an output component that can identify a first view instruction that is listed in the set of view instructions associated with the dataset, wherein the first view instruction is respectively correlated to the first security credential. In various cases, the output component can generate as output a view of the dataset based on the first view instruction.

Abstract: One or more computer processors transmit a machine learning model and an associated loss function to a worker, wherein the worker isolates private data. The one or more computer processors receive a plurality of encrypted gradients computed utilizing the transmitted machine learning model, the associated loss function, and the isolated private data. The one or more computer processors generate a plurality of adversarial perturbations, wherein the plurality of adversarial perturbations includes true perturbations and false perturbations. The one or more computer processors obfuscate the generated plurality of adversarial perturbations. The one or more computer processors transmit the obfuscated adversarial perturbations to the worker. The one or more computer processors harden the machine learning model utilizing the transmitted obfuscated adversarial perturbations and the private data.

Abstract: Disclosed are techniques for linking information about individual entities across multiple datasets. A target dataset with some information corresponding to at least one attribute of an entity is received. Semantic processing is performed on the target dataset to extract semantic representations of the information and corresponding attributes, which is utilized to search at least one other dataset for additional information that is absent from the target dataset, corresponding to at least one attribute of the entity, which are used to augment the target dataset with additional information corresponding to the entity. This is repeated iteratively, with each subsequent iteration including semantic representations of information found in the searches of previous iterations until no additional information about the entity is found when searching the multiple datasets with semantic representations of the now augmented target dataset.

Abstract: Embodiments of the invention are directed to techniques that include receiving a query intended for a targeted database and determining that the query is from an unauthorized user. A response is returned to the unauthorized user generated by a model, the response being dynamically generated to fulfill the query. The model is configured to generate responses consistent with any previous responses returned to the unauthorized user.

Abstract: Embodiments include a method for one or more processors to receive an organic dataset and a domain knowledge base. The one or more processors identify private data entities present within the organic dataset. The one or more processors determine statistical properties of the private data entities identified within the organic dataset. The one or more processors create a plurality of test data templates by removing the private data entities from the organic dataset. The one or more processors select from the domain knowledge base, synthetic data entities that match a data type of the removed private data entities, respectively, and align with the statistical properties of the private data entities, and the one or more processors generate synthetic test data by inserting, respectively, the synthetic data entities of the matching data type for the removed private data entities in the test data templates.

Abstract: A method, computer system, and a computer program product for published content protection is provided. The present invention may include receiving a content file from a content management system (CMS). The present invention may include extracting a feature from the received content file. The present invention may include transforming, using an adversarial generation algorithm, the received content file into an adversarial content file. The present invention may include returning the adversarial content file to the CMS. The returned adversarial content file may represent an equivalent of the received content file to a content consumer. The present invention may include preventing an application of the returned adversarial content file in at least one machine learning task based on the adversarial noise included in the returned adversarial content file.

Abstract: A system and method for implementing a marketplace for risk assessed smart contracts issuers and execution providers in blockchains in a computer environment are presented. A smart contract, issued by one or more blockchain nodes to a secondary blockchain and accepted by a primary blockchain, may be executed following a risk assessment to recursive call attack vulnerabilities of the smart contract according to a reputation of both the one or more blockchain nodes and the one or more execution nodes.

Abstract: A first copy of a heterogeneous program payload is provided to a first computing device. The heterogeneous program payload contains an unencrypted component and a set of one or more encrypted components. The set of encrypted components corresponding to a set of one or more programs. The unencrypted component of the heterogeneous program payload includes loader program code configured to receive a first license key. The loader program code is configured to, in response to receiving the first license key, perform a decryption action against the set of encrypted components of the heterogeneous program payload.

Abstract: Various embodiments are provided for providing intelligent data utility constraint information in a computing environment are provided. One or more utility constraints may be extracted from a dataset and a set of queries and applications to infer one or more data transformations to apply to selected data.

Abstract: Quasi-identifiers (QIDs) are detected in a dataset using a set of computing tasks. The dataset has a plurality of records and a set of attributes. An index is generated for the dataset. The index has an indicator for each attribute value of each record in the dataset. Each indicator specifies all the records in the dataset having the same value for the attribute. Each task is assigned an attribute combination and a subset of the plurality of records in the dataset and is passed to a thread for execution on computing resources. The executing task inspects the set of records specified by the index indicator for each attribute value in the attribute combination to produce a result. The result of at least one task identifies a unique record for the associated attribute combination. The attribute combination producing the unique record is a QID.

Abstract: Embodiments are provided for intelligent augmentation of word representations in a computing system by a processor. One or more character shape embeddings may be created in an artificial neural network (ANN) by augmenting one or more word representations based on a set of natural language processing (NLP) data, a set of word embedding vectors, and one or more clustering operations. A shape, a category, and a contextual identify of the one or more word representations may be analyzed used in creating the one or more character shape embeddings.

Abstract: Embodiments for performing consistent data masking in a distributed computing environment by a processor. A dictionary based data masking operation is performed on one or more datasets with causal ordering of the one or more datasets to enable reconstruction of a state of the one or more dictionaries for the one or more datasets.

Abstract: Embodiments for ensemble training in a distributed marketplace in a computing environment. One or more ensemble machine learning models may be provided from a plurality of machine learning models competing within the distributed marketplace that achieve a performance on ensemble training data equal to or greater than a selected performance threshold, wherein the distributed marketplace is a blockchain.

Abstract: Various embodiments are provided for providing data security in a computing environment are provided. Data may be inspected during a write operation or a read operation and selected data from the data may be filtered according to one or more data security policies or rules prior to sending the plurality of data to or receiving the plurality of data from a shared computing file system.

Abstract: A system and method for reducing risk of smart contracts in blockchains in a computer environment are presented. One or more smart contracts may be accepted or rejected from a secondary blockchain to a primary blockchain according to a risk assessment to recursive call attack vulnerabilities.

Abstract: Embodiments are provided for intelligent augmentation of word representations in a computing system by a processor. One or more character shape embeddings may be created in an artificial neural network (ANN) by augmenting one or more word representations based on a set of natural language processing (NLP) data, a set of word embedding vectors, and one or more clustering operations. A shape, a category, and a contextual identify of the one or more word representations may be analyzed used in creating the one or more character shape embeddings.

Abstract: Embodiments for providing optimized machine learning model features using federated learning on distributed data in a computing environment by a processor. Machine learning model features may be learned from one or more data sets extracted from one or more localized machine learning models associated with one or more nodes. The machine learning model features may be aggregated using a centralized machine learning model at a source node. The one or more localized machine learning models may be trained using aggregated machine learning model features provided by the centralized machine learning model.

Abstract: Various embodiments are provided for authenticating an entity in non-secure communication channels via secure out-of-bands channels. An operator may be authenticated to have access to a secured location associated with an entity upon determining the operator retrieved and communicated a unique token, provided by the user and stored at the secured location, back to the user.

Abstract: Various embodiments are provided for linking of anonymized datasets in a computing environment are provided. A number of linking records may be identified between an anonymized dataset and one or more non-anonymized datasets of a knowledge base according to one or more equivalence classes and a generalization level.

Abstract: To improve the security performance of a computer system, a retrieval component executing on at least one hardware processor obtains a list of known compromised passwords. A validation component executing on the at least one hardware processor obtains a specification of a putative password and risk-scores the putative password based at least in part on presence of the putative password in the list of known compromised passwords. The system obtains a specification of an actual password chosen in accordance with the risk score of the putative password. Access to at least one aspect of the computer system is controlled based on the actual chosen password.