Abstract: Aspects of the present invention disclose a method, computer program product, and system for determining engagement levels of an audience and providing associated recommendations to a presenter. The method includes one or more processors receiving sensor data from sensors located in an audience area during a presentation. The method further includes one or more processors determining an engagement level for a first individual in the audience area utilizing sensor data from sensors associated with the first individual. The method further includes one or more processors determining whether the determined engagement level for the first individual indicates that the first individual is not engaged with the presentation. In response to determining that the determined engagement level of the first individual indicates that the first individual is not engaged with the presentation, the method further includes one or more processors determining recommendations to increase engagement of the first individual.

Abstract: A method, a structure, and a computer system for digital twin management. The exemplary embodiments may include gathering data corresponding to one or more bodily tissues and generating one or more digital twin versions of the one or more bodily tissues. In addition, the exemplary embodiments may further include identifying a most appropriate digital twin version of the one or more digital twin versions based on a health and a compatibility of the one or more digital twin versions, and distributing the most appropriate digital twin version.

Abstract: Techniques facilitating mitigation of dangerous activities are provided. In one example, a computer-implemented method can include detecting, by a system operatively coupled to a processor, a trigger event based on a determination that a security risk level satisfies a defined condition associated with the security risk level. The security risk level can be associated with an item of interest and a monitored behavior related to the item of interest. The computer-implemented method can also include implementing, by the system, an action that mitigates an impact of the security risk level. The action can be implemented as the trigger event is detected and the action can be determined based on the item of interest and the monitored behavior.

Abstract: A method, system and apparatus of for guiding a physically impaired individual during an emergency, including receiving emergency data, determining or predicting the reactions of other people in response to the emergency data, interpreting in real-time the reactions of the other people to physically impaired person according to the impaired person profile, determining a type and modality of aid services to the physically impaired, and prompting the physically impaired person through one or more devices for aid services.

Abstract: An example operation may include one or more of receiving user profile identification data associated with a user profile, creating a smart contract on a blockchain with the user profile identification data, validating an identity of the user profile based on the user profile identification data, performing a predictive analysis by the smart contract to determine one or more future tasks to be performed by a user device associated with the user profile, generating one or more tokens associated with the user profile, the one or more tokens include access rights for the user device to perform the one or more future tasks, and storing the one or more tokens in the blockchain.

Abstract: Methods and systems for operating a machine learning system are described. In an example, a device can receive an image and assign a set of pixels in the image as a digital representation of a livestock. The device can further train a machine learning model using the digital representation. The device can further run the machine learning model to generate prediction data relating to the livestock. The device can further generate output data relating to at least one activity among a livestock value chain. The at least one activity can correspond to a process to generate a commodity based on the livestock.

Abstract: A signal representing a sound can be received. A machine learning model can be run to identify that the sound triggers a reaction in a user hearing the sound. A preventive action can be automatically activated to mitigate the reaction. The user's reactions can be monitored. Responsive to determining that the user's reaction has been mitigated or suppressed, the preventive action can be deactivated. The machine learning model can be retrained using at least the signal as new training data.

Abstract: Improved quantification, detection, and characterizing of out-of-distribution (OOD) of a set of inputs that are generated by an alternative process such as anomalies, outliers, adversarial attacks, input errors can be provided with a combination of detection under perturbations and subset scanning algorithms. A first set of activations is extracted from nodes in a hidden layer of a neural network for an input. Noise is added to the input. A second set of activations is extracted from nodes in the hidden layer of a neural network for the noised input. A difference between the first set of activations and the second set of activations is determined. The difference is compared with a difference computed using in-distribution samples. Based on the comparison, an anomaly score for the input is determined. Multiple inputs can be processed. An iterative ascent algorithm finds out-of-distribution input and internal nodes with anomalous activations.

Abstract: An image file is to be posted on a social media web site is automatically detected, the image file containing at least an image of a first individual with a background image depicting a location determined to be unsafe. A risk level associated with the location is determined. Based on the risk level, an action is automatically activated, which prevents at least a second individual from taking a photograph of the second individual at the location.

Abstract: The present invention is a system and method that generates a Table of Contents (TOC) customized to the user knowledge about the concept(s) in the user query and the specific context and preferences of the user. The invention identifies search concepts within the search queries, receives search results, and splits the search results into one or more result segments. In a preferred embodiment, a correlation strength between concepts in one or more of the result segments and the user search query, along with reference to the user knowledge, context, and/or preferences determine which result segments are selected in sequenced to form the TOC.

Abstract: An operation performed with a computer cursor on a graphical user interface (GUI) can be detected, which operation copies content pointed to by the computer cursor into a memory buffer. An analysis can be performed to determine a risk associated with the operation. Based on the analysis, a change in a GUI element of the GUI can be made or caused.

Abstract: Context associated with content may be received. An additional content may be generated for insertion into the content. The additional content can be created to be within the context associated with the content and based on a likely responsiveness of the user to the additional content, the additional content referring to an item. The additional content preserves continuity and/or semantics in the context of the content.

Abstract: Amount of distance and/or time spent driving off-road by a vehicle during a specified drive may be detected. Readings from a location device and accelerometer sensor of the vehicle may be compared with historical readings of other vehicles taking similar drives, and for example, with similar driving behaviors, to detect off-road driving by the vehicle.

Abstract: An emergency vehicle that is displaying emergency lights detects an approaching vehicle, and assesses road conditions, intensity of the emergency lights at the approaching vehicle, a response of the approaching vehicle to the emergency lights, and a state variable related to a driver of the approaching vehicle. The emergency vehicle then obtains a likelihood value that the approaching vehicle will initiate an accident, by supplying information about the road conditions, the intensity of the emergency lights at the approaching vehicle, the response of the approaching vehicle, and the state variable as inputs to a neural network trained to correlate accidents to such information. In response to the likelihood value exceeding a threshold value, the emergency vehicle modifies the emergency lights.

Abstract: Embodiments for intelligent media data filtering in an Internet of Things (IoT) computing environment by a processor. User preferences for the media data may be determined according to a plurality of identified contextual factors. At least a portion of the media data may be filtered according to one or more user classifications.

Abstract: Systems and methods for reminding a user or a group of users about one or more items related to an upcoming event. The upcoming event is identified based on user data and a context for determining a list of items useful for the upcoming event is obtained. The list of items for the identified upcoming event is determined, using the context, and item reminder enhancement factors are determined in response to the upcoming event and the determined list of items. Content selection models are determined based on the context, the user data, and the item reminder enhancement factors. Conversational content for engaging the user or group of users is generated by applying the content selection models to a plurality of potential content. A conversation trigger signal is determined, and the conversational content is presented to the user via a computing device in response to receiving the conversation trigger signal.

Abstract: Methods and systems for crop grading and crop management. One or more images of crops are obtained and one or more crop related features are at least one of identified or extracted from the one or more images. A crop health status is determined based on the one or more crop related features, an environmental context, a growth stage of the crop, and a farm cohort by using a computerized deep learning system to perform an automated growth stage analysis. One or more actions are at least one of recommended, triggered, and performed.

Abstract: An apparatus, for example a primary device, implements a method that includes receiving promoted content for display at the primary device; predicting a user's behavior with respect to the promoted content by estimating the user's context and preferences, using a neural network; and in response to the prediction of the user's behavior, automatically demoting the promoted content.

Abstract: In an aspect, during a presentation of a presentation material, viewers of the presentation material can be monitored. Based on the monitoring, new content can be determined for insertion into the presentation material. The new content can be automatically inserted to the presentation material in real time. In another aspect, during the presentation, a presenter of the presentation material can be monitored. The presenter's speech can be intercepted and analyzed to detect a level of confidence. Based on the detected level of confidence, the presenter's speech can be adjusted and the adjusted speech can be played back automatically, for example, in lieu of the presenter's original speech that is intercepted.

Abstract: A processor-implemented method records and maintains a record of browser events in a blockchain using a peer-to-peer network. One or more processors detect one or more browser events for a browser on a computer. The processor(s) transmit transactions associated with the browser event(s) to a peer-to-peer network of devices that create a blockchain, which includes one or more blocks that describe the browser event(s), where the blockchain records and maintains a record, of browser events, that includes records of uniform resource locators (URLs) browsed by a user, bookmarks added to the browser, and search terms searched by the user. The processor(s) perform a vulnerability analysis to determine how vulnerable the computer and the browser are to a malicious attack, and adjust a frequency of transmitting the transactions from the computer to the peer-to-peer network of devices according to how vulnerable the computer and the browser are to the malicious attack.