Abstract: An embodiment includes computing a computational cost of a job, using a first machine learning algorithm. The job may include an original amount of an egress of data from a cloud computing environment. The embodiment includes determining, using a second machine learning algorithm, the amount of the egress of data corresponding to the job has a computer business criticality that exceeds a threshold level of business criticality. The embodiment includes analyzing a current egress plan used in computing the computational cost of the job. The embodiment includes reconfiguring the current plan to a second egress plan to reduce the computational cost of the job. The embodiment includes implementing a second plan such responsive to execution of the job, the second plan causes data egress behavior to change from the original egress of data behavior. A modified egress behavior causes an effective reduction in the egress cost of the job.

Abstract: A computer-implemented method, according to one embodiment, includes: receiving a preliminary password key from a user, and in response to receiving the preliminary password, determining whether a cursor of a graphical user interface is located in a password entry field. In response to determining that the cursor of the graphical user interface is not located in the password entry field, a warning is issued to the user. Additionally, any inputs that are received from the user are discarded in response to determining that the cursor of the graphical user interface is not located in the password entry field.

Abstract: An embodiment predicts a plurality of strategies associated with a cloud workload. The embodiment presents the predicted plurality of strategies to a user for selection of a strategy. The embodiment predicts a first technology for the cloud workload having a first technology strategic potential value based on the selected strategy. The embodiment predicts a first deployment for the cloud workload having a first deployment strategic potential value based on the selected strategy. The embodiment identifies a second technology for the cloud workload having a second technology strategic potential value based on the selected strategy. The embodiment identifies a second deployment for the cloud workload having a second deployment strategic potential value based on the selected strategy. The embodiment transmits a strategic alert.

Abstract: An embodiment optimizes a live stream event by annotating a script with markers for a live stream event. The embodiment defines an element within the live stream event. The embodiment defines a performance milestone within the element. The embodiment associates a first action with the performance milestone. The embodiment constructs a trigger within the script. The embodiment associates the trigger with a corresponding second action. The embodiment monitors the live stream event for the performance milestone and actuates the first action in response to the performance milestone. The embodiment monitors the live stream event for the trigger and actuates the second action in response to the trigger. The first and the second actions are real time changes to the live stream event.

Abstract: A method for identifying video clips for inclusion in a targeted communication is disclosed. In one embodiment, such a method includes receiving video footage comprising multiple clips. The method receives multiple information streams that are associated with the video footage and synchronizes the information streams with the video footage. The method applies weights to the information streams, aggregates the weighted information streams, and identifies peaks therein. Clips are then selected from the video footage that correspond to the peaks for inclusion in a targeted communication. The method analyzes metrics from the targeted communication to provide feedback in order to optimize the weights. In certain embodiments, optimizing the weights leads to selecting different clips for inclusion in the targeted communication. A corresponding system and computer program product are also disclosed.

Abstract: An embodiment establishes a system behavior database based at least in part on behavior data received from a system, wherein the system behavior database comprises a set of historical observations of behavior of the system. The embodiment samples current behavior data from the system. The embodiment generates a current observation based on the current behavior data that was sampled. The embodiment compares the current observation to each historical observation of the set of historical observations to determine a divergence between the current observation and each historical observation. The embodiment compares the divergence to a divergence threshold, and upon a determination that the divergence exceeds the divergence threshold, detects an anomaly in the system, updates the set of historical observations to include the current observation as a new historical observation, and performs a responsive action within the system based on the detected anomaly.

Abstract: A method for training and fine-tuning an artificial intelligence model is disclosed. In one embodiment, such a method distributes, across multiple chiplets of a package, functionality associated with a deep neural network. The method implements, within a first set of chiplets, frozen layers of the deep neural network. By contrast, the method implements, within a second set of chiplets, trainable layers of the deep neural network. The number of chiplets in the second set may be smaller than the number of chiplets in the first set and may consist of a single chiplet in some embodiments. In certain embodiments, the second set of chiplets has one or more of additional memory capacity and additional processing capacity compared to the first set of chiplets in order to train and fine tune the trainable layers. A corresponding apparatus is also disclosed.

Abstract: An embodiment extracts, from a natural language document, using a first language model, a set of cause-effect pairs, each cause-effect pair comprising a pair of phrases, each phrase comprising a portion of the natural language document. An embodiment constructs a first graph representing the set of cause-effect pairs, each node in the first graph representing a phrase in the set of phrases, each edge in the first graph representing a cause-effect relationship between nodes connected by an edge. An embodiment measures a graph similarity between the first graph and a second graph constructed from phrases extracted from the natural language document. An embodiment incorporates, into a knowledge base responsive to determining that the graph similarity is above an acceptance threshold, the first graph.

Abstract: An embodiment sends, at a first time, from a boundary server to a client system in response to a challenge request, a challenge specifying a computational problem to be solved by the client system, the boundary server specified in a challenge list sent to the client system. An embodiment receives, at a second time, at the boundary server, a challenge response from the client system, the challenge response comprising a solution to the computational problem. An embodiment generates, at the boundary server, a certificate encoding an elapsed time between the first time and the second time, the certificate usable by the client system to prove a location of the client system. An embodiment sends, from the boundary server to the client system, the certificate.

Abstract: A privacy-preserving method of accessing a model as a service (e.g., a language model) receives an input prompt authored in a first domain having data intended to be maintained private. In response, and using a conversion model, the input prompt authored in the first domain is converted to a second domain to create a converted input prompt. The converted input prompt preserves a semantic context of the input prompt. The converted input prompt is then delivered to the model as a service, which returns a response in the second domain. Using the conversion model, the response is then converted back into the first domain to create a converted response. The converted response has the semantic context of the input prompt. The converted response is provided as a reply to the input prompt. The conversion model may be trained using data collected from user interactions with the model as a service.

Abstract: A method to handle insufficient on-chip memory capacity in decompressors is disclosed. In one embodiment, such a method includes executing, by a decompressor configured to decompress data, an instruction configured to copy data from a source position within a data stream to a destination position within the data stream. The method determines whether the source position currently resides within an on-chip buffer of the decompressor. In the event the source position does not currently reside within the on-chip buffer of the decompressor, the method writes arbitrary placeholder data to the destination position and adds the instruction to a patch buffer. At a later point in time, the method retrieves the instruction from the patch buffer and executes the instruction by retrieving the data from the source position and overwriting the arbitrary placeholder data at the destination position with the data. A corresponding system and computer program product are also disclosed.

Abstract: An embodiment parses, by a host plugin, a host content view into a plurality of host resources. The embodiment identifies, by an attendee plugin, an attendee interaction with a host resource in the plurality of host resources. The embodiment transmits, by the attendee plugin, based on the attendee interaction, a resource request for the host resource. The embodiment transmits, by the host plugin, based on the resource request, a host resource content associated with the host resource. The embodiment executes, by the attendee plugin, based on the host resource content, a response action for presenting the host resource content.

Abstract: A mitigation system protects data in a data store that is not yet encrypted by a successful ransomware attack against encryption. Data is stored in the data store as a set of versions identified in a data tree, and a version can only be updated by writing a new version to the tree. Access controls to prevent modification of the tree are also in place. Following an attack, a restore function is executed to attempt recovery. This function computes an entropy delta that compares an entropy of an encrypted version, with an entropy of versions of the data not yet encrypted. Based on the computed entropy deltas, the restore function identifies a latest clear version of the data, and a restore operation is then is initiated with respect to this version.

Abstract: A method for countering adversarial attacks on deep neural networks is disclosed. In one embodiment, such a method includes observing, by a first system, actual traffic behavior within a transportation network. The method classifies, by a deep neural network of a second system, a traffic sign for regulating traffic within the transportation network. The method determines whether a conflict exists between the actual traffic behavior and expected traffic behavior based on the traffic sign. If a conflict is deemed to exist, the method adjusts training data of the deep neural network with respect to classifying the traffic sign. In certain embodiments, the first system includes one or more forerunner vehicles and the second system is an autonomous vehicle. The one or more forerunner vehicles may be configured to travel ahead of the autonomous vehicle when navigating the transportation network. A corresponding system and computer program product are also disclosed.

Abstract: An embodiment detects, by an outlier detector, a first potential outlier in a data structure. The embodiment determines, by the outlier detector, whether the first potential outlier is a first outlier based on a first threshold. The embodiment applies, by an outlier corrector, responsive to determining the first potential outlier is a first outlier, a data quality rule to the first outlier. The embodiment detects, by the outlier detector, a second potential outlier in the data structure. The embodiment decreases, by the outlier detector, the first threshold to a second threshold. The embodiment determines, by the outlier detector, whether the second potential outlier is a second outlier based on the second threshold. The embodiment applies, by the outlier corrector, responsive to determining the second potential outlier is a second outlier, the data quality rule to the second outlier.

Abstract: An embodiment constructs a knowledge graph based on data received from a system wherein the knowledge graph comprises a plurality of entities and relationships among the plurality of entities which represent a context information of the knowledge graph. The embodiment assigns a unique identifier to a designated entity. The embodiment expresses in natural language the unique identifier, an associated label, an associated property name, an associated property value, an associated relationship label and an associated relationship property name. The embodiment trains, using a machine learning algorithm, a foundation model responsive to inputting the natural language expression and applying a prompt comprising a cryptographic criterion, wherein the foundation model is trained to solve a task associated with the cryptographic criterion.

Abstract: A computer system performs automated inventory management. A service level is determined, using a predictive model, for a count of inventory units for one or more sites of a plurality of sites, wherein the predictive model is provided with a reorder point and reorder quantity as input. In response to determining that the service level does not satisfy a threshold value, the reorder point is adjusted. The adjusted reorder point is determined to yield a service level, according to the predictive model, that satisfies the threshold value. Instructions are automatically transmitted to cause a reorder of inventory units in response to determining that a count of inventory units is below the adjusted reorder point. Embodiments of the present invention further include a method and program product for performing automated inventory management in substantially the same manner described above.

Abstract: A semiconductor structure, and a method of making the same includes a fin extending upward from a substrate, an epitaxially grown bottom source/drain region in direct contact with the substrate and a bottom portion of the fin. A bottom surface and sidewalls of a metal silicide layer are in direct contact with the epitaxially grown bottom source/drain region. A bottom spacer is located above and in direct contact with the metal silicide layer and a portion of the epitaxially grown bottom source/drain region not covered by the metal silicide layer, the bottom spacer surrounding the fin.

Abstract: Use a computerized trained graph neural network model to classify an input instance with a predicted label. With a computerized graph neural network interpretation module, compute a gradient-based saliency matrix based on the input instance and the predicted label, by taking a partial derivative of class prediction with respect to an adjacency matrix of the model. With a computerized user interface, obtain user input responsive to the gradient-based saliency matrix. Optionally, modify the trained graph neural network model based on the user input; and re-classify the input instance with a new predicted label based on the modified trained graph neural network model.

Abstract: A method, a structure, and a computer system for unidirectional text comparison. The exemplary embodiments may include determining a first similarity score between a first text string and a second text string, and computing an error term between the first text string and the second text string, wherein the error term incorporates a directionality of the first text string and the second text string. The exemplary embodiments may further include determining a second similarity score based on the first similarity score and the error term.