Abstract: A computer implemented method manages zero-day vulnerabilities in an application package having a set of components. The computer ingests data about potential vulnerabilities from a plurality of data sources. Using a set of machine learning models, the computer predicts a vulnerability based on of the data that was ingested. The computer performs a code analysis of the set of components to identify a possibility of the vulnerability impacting the application package. The computer generates a recommendation to resolve the vulnerability based on the code analysis and the data that was ingested. The computer manages the recommendation in a private blockchain.

Abstract: A computer-implemented method, system and computer program product for identifying and recommending code snippets to be reused by a software developer. Multiple sources (e.g., social media messages) are monitored for source code development difficulties. In response to detecting a source code development difficulty, the source code which is the source of the detected source code development difficulty is labeled with a struggle tag, which may indicate a category of the source code. Furthermore, source code snippets from various sources (e.g., source code snippets publicly available on the Internet) are mined, identified and labeled with tags, which may indicate a category of the source code snippet. Relevant source code snippets are identified to address the software developer's struggles based on matching one or more of these tags with the struggle tag, which are later ranked as to which are the best to address the software developer's struggles in writing source code.

Abstract: Embodiments of the invention provide a method, system and computer program product for viewport selection for hypervideo presentation. In a method for viewport selection for hypervideo presentation, a multiplicity of different hypervideos, for example 360° hypervideos, are played back to an end user and end user interactions by the end user with each of the different hypervideos are recorded. Then, an end user profile for the end user is computed from the recorded end user interactions so as to specify a particular viewport. Finally, in response to a directive by the end user to view a new hypervideo, the end user profile is retrieved and the particular viewport identified so that the new hypervideo is played back to the end user utilizing the particular viewport.

Abstract: One embodiment provides a method, including: obtaining, from an application used by users to converse with other users, a conversation corresponding to a user, wherein the conversation includes an indication of an intent to purchase a plurality of retail services; identifying at least one micro-service corresponding to each of the plurality of retail services, wherein the micro-service allows purchase from the corresponding retail service; sequencing the micro-services, thereby sequencing purchases from the plurality of retail services through the micro-services, wherein the sequencing comprises (i) determining a sequence of purchases and (ii) identifying, for each of the determined sequences, a financial benefit to the user; and providing a recommendation comprising a purchasing sequence to the user, wherein the recommendation comprises the sequence of purchases that result in the maximum financial benefit to the user.

Abstract: A computer-implemented method, system and computer program product for identifying and recommending code snippets to be reused by a software developer. Multiple sources (e.g., social media messages) are monitored for source code development difficulties. In response to detecting a source code development difficulty, the source code which is the source of the detected source code development difficulty is labeled with a struggle tag, which may indicate a category of the source code. Furthermore, source code snippets from various sources (e.g., source code snippets publicly available on the Internet) are mined, identified and labeled with tags, which may indicate a category of the source code snippet. Relevant source code snippets are identified to address the software developer's struggles based on matching one or more of these tags with the struggle tag, which are later ranked as to which are the best to address the software developer's struggles in writing source code.

Abstract: One embodiment provides a method, including: receiving at least two challenge test mechanisms of different challenge test modalities, wherein a challenge test mechanism comprises a challenge portion of a challenge-response test for distinguishing between a human operator and a computer; receiving challenge test operators for combining the at least two challenge test mechanisms; generating a composite challenge task by combining the at least two challenge test mechanisms using the identified challenge test operators; identifying any errors in the composite challenge task by running the composite challenge task; evaluating the composite challenge task to determine (i) a challenge difficulty for a human operator and (ii) a challenge difficulty for a computer; and implementing the composite challenge task if (i) no errors are identified at the composite challenge task analyzer, (ii) the challenge difficulty for a human operator is below a predetermined threshold, and (iii) the challenge difficulty for a computer

Abstract: A computer-implemented method for context-based, fine-grained data access control to microservice data is provided. The method may include retrieving a sensor data snapshot corresponding to the microservice data from a sensor node array of a microservice platform, and encrypting the sensor data snapshot according to a functional encryption scheme to generate an encrypted sensor data snapshot. The method may further include receiving a registration request from a user device, detecting an occurrence of the user-defined event based on sensory event data corresponding to the user event data, and generating a restricted-access functional decryption key in response to detecting the occurrence of the user-defined event. The restricted-access functional decryption key may be communicated to the user device for decryption of the encrypted sensor data snapshot according to an access control policy corresponding to a user-defined event associated with the user device.

Abstract: One embodiment provides a method, including: obtaining, from an application used by users to converse with other users, a conversation corresponding to a user, wherein the conversation includes an indication of an intent to purchase a plurality of retail services; identifying at least one micro-service corresponding to each of the plurality of retail services, wherein the micro-service allows purchase from the corresponding retail service; sequencing the micro-services, thereby sequencing purchases from the plurality of retail services through the micro-services, wherein the sequencing comprises (i) determining a sequence of purchases and (ii) identifying, for each of the determined sequences, a financial benefit to the user; and providing a recommendation comprising a purchasing sequence to the user, wherein the recommendation comprises the sequence of purchases that result in the maximum financial benefit to the user.

Abstract: A computer-implemented method for context-based, fine-grained data access control to microservice data is provided. The method may include retrieving a sensor data snapshot corresponding to the microservice data from a sensor node array of a microservice platform, and encrypting the sensor data snapshot according to a functional encryption scheme to generate an encrypted sensor data snapshot. The method may further include receiving a registration request from a user device, detecting an occurrence of the user-defined event based on sensory event data corresponding to the user event data, and generating a restricted-access functional decryption key in response to detecting the occurrence of the user-defined event. The restricted-access functional decryption key may be communicated to the user device for decryption of the encrypted sensor data snapshot according to an access control policy corresponding to a user-defined event associated with the user device.

Abstract: One embodiment provides a method, including: receiving at least two challenge test mechanisms of different challenge test modalities, wherein a challenge test mechanism comprises a challenge portion of a challenge-response test for distinguishing between a human operator and a computer; receiving challenge test operators for combining the at least two challenge test mechanisms; generating a composite challenge task by combining the at least two challenge test mechanisms using the identified challenge test operators; identifying any errors in the composite challenge task by running the composite challenge task; evaluating the composite challenge task to determine (i) a challenge difficulty for a human operator and (ii) a challenge difficulty for a computer; and implementing the composite challenge task if (i) no errors are identified at the composite challenge task analyzer, (ii) the challenge difficulty for a human operator is below a predetermined threshold, and (iii) the challenge difficulty for a computer

Abstract: A method for placing a linked hyper-video within a hyper-video. The method includes identifying one or more points of interest in the hyper-video, detecting one or more features of the hyper-video, extracting at least one of the one or more features, and displaying the linked hyper-video, within the hyper-video, based on the one or more identified points of interest and based on the extracted at least one of the one or more features. The method further includes extracting one or more features from the hyper-video that depict one or more viewing characteristics of a user. The method includes analyzing the one or more features from the hyper-video to identify a low-saliency area of the hyper-video, a low-frequency region within the low-saliency area, and a homogenous sub-region within the low-frequency region. The method further includes placing a linked hyper-video at the homogenous sub-region within the low-frequency region of the hyper video.