Abstract: Described is a method for logistic transport monitoring of an IoT device utilizing token generation for payload data for features and peripheral functionality associated with the IoT device being shipped between an origin location and a target location. The method include generating a pre-shipment token prior to loading of the IoT device onto a logistic transporter and regenerating a token subsequent to loading of the IoT device onto the logistic transporter, where a validated pre-shipment token and a matching regenerated token indicate a faultless operational status of the IoT device. The method includes iteratively regenerating tokens during the shipment of the IoT device, where a mismatch between an iteratively generated token and the regenerated token indicate an operational fault of the IoT device indicating the IoT device was damaged during the shipment between the origin location and the target location.

Abstract: An electronic structure includes a first substrate having a first under bump metallization (UBM) region and a second UBM region formed thereon. One or more solder bumps is deposited onto the first UBM region. A downstop formed on the second UBM region is wider, shallower and more rigid than any one of the solder bumps formed on the first UBM region. A second substrate is joined to the first substrate by the one or more solder bumps located on the first UBM region, and a height of the downstop limits a distance between at least one of the first substrate and the second substrate, or between an object and at least one of the first substrate and the second substrate.

Abstract: An integrated circuit includes a field effect transistor (FET) and a phase change memory (PCM) cell. The PCM cell includes a heater, wherein a bottom surface of the heater is at or below a top surface of the FET.

Abstract: A stored energy discharge apparatus for discharging electrical charge from a first component prior to being installed into a second component. The apparatus includes at least one mounting component adapted to a desired voltage potential used in the second component, and at least one discharge component adapted to receive the first component and discharge electrical charge from the first component to the second component through the at least one mounting component prior to the first component being installed in the second component.

Abstract: In an approach to improve converting conversation to user stories, embodiments capture keywords from a captured discussion, and identify the probability of the keywords being an object attribute or action behavior. Further, responsive to identifying, based on the probability, that the keywords are the object attribute or the action behavior, embodiments determine that the object attribute or the action behavior are not new to a first user story. Additionally, embodiments determine that the attribute or the action behavior are associated with an existing object in a first user story, and update the first user story with the attribute or the action behavior.

Abstract: An approach for training a recurrent neural network to create a model for anomaly detection in the topology of a network is disclosed. The approach comprises, creating an embedding vector for each resource in the network based on applying an embedding algorithm to each resource of the network. A feature vector is then created for each change to a resource in the network based on one or more properties of the change. A recurrent neural network can thus be trained with the embedding vectors and the feature vectors to create a model for anomaly detection in the topology of the network.

Abstract: A temperature-aware task scheduling method, system, and computer program product for facilitating a task in a multi-graphical processing unit (GPU) environment, includes executing the task to a GPU in the multi-GPU environment based on a thermal characteristic of the GPU as compared to the other GPUs in the multi-GPU environment.

Abstract: In an approach for smart test data workload generation, a processor receives a plurality of expected image frames for a user interface application to be tested. The plurality of expected image frames is pre-defined and represents a series of workflows and operations of the user interface application to be expected based on a design requirement. A processor calculates a first set of hash-values for each corresponding expected image frame. A processor samples the user interface application with a frequency to a plurality of testing image frames during a test run on the user interface application. A processor calculates a second set of hash-values for each sampled testing image frame. A processor compares the first set of hash-values to the second set of hash-values. A processor verifies that the second set of hash-values matches the first set of hash-values.

Abstract: Mechanisms are provided for performing machine learning training of a computer model. A perturbation generator generates a modified training data comprising perturbations injected into original training data, where the perturbations cause a data corruption of the original training data. The modified training data is input into a prediction network of the computer model and processing the modified training data through the prediction network to generate a prediction output. Machine learning training is executed of the prediction network based on the prediction output and the original training data to generate a trained prediction network of a trained computer model. The trained computer model is deployed to an artificial intelligence computing system for performance of an inference operation.

Abstract: Embodiments are directed to elliptic curve cryptography scalar multiplications in a generic field with heavy pipelining between field operations. A bit width is determined of operands in data to be processed by a modular hardware block. It is checked whether the bit width of the operands matches a fixed bit width of the modular hardware block. In response to there being a match, the modular hardware block processes the operands. In response to there being a mismatch, the operands are modified to be accommodated by the fixed bit width of the modular hardware block.

Abstract: In an approach to meeting summarization for disconnected participants, one or more computer processors detect a start of a collaborative meeting. One or more computer processors monitor a connection to the collaborative meeting of each of one or more participants. One or more computer processors determine a first participant of the one or more participants has disconnected from the collaborative meeting. One or more computer processors record the collaborative meeting. One or more computer processors determine the first participant has re-connected with the collaborative meeting. One or more computer processors synthesize, a first summary of a portion of the collaborative meeting missed by the first participant. One or more computer processors to synthesize, a second summary of a portion of the collaborative meeting recorded after the first participant re-connected with the collaborative meeting. One or more computer processors present the first summary and the second summary to the first participant.

Abstract: A semiconductor component includes an area of dielectric material extending below an uppermost surface of a substrate. The semiconductor component further includes a trench formed so as to extend from above the uppermost surface of the substrate into the area of dielectric material. The semiconductor component further includes a non-metal liner coating interior surfaces of the trench. The semiconductor component further includes a metal liner coating interior surfaces of the non-metal liner. The semiconductor component further includes a power rail formed in the trench in direct contact with at least one of the metal liner or the non-metal liner such that the power rail extends into the area of dielectric material and above the uppermost surface of the substrate.

Abstract: An approach to harvesting electrical energy from a user blinking is disclosed. The approach uses an apparatus that includes a piezoelectric film attaching to a user's eyelid, a wire connecting the piezoelectric film to an electronic component; and an antenna surrounding an eye of the user connecting to the electronic component. The approach includes harvesting electrical energy from a user blinking with the piezoelectric film and the antenna wirelessly transmitting the electrical energy to smart contact lens worn by the user.

Abstract: Embodiments are disclosed for a method for a security model. The method includes identifying a plurality of primary semantic relationships between a plurality of initial incident artifacts for a security domain based on a plurality of historical incidents. The method further includes identifying a plurality of parsed incident artifacts from a security encyclopedia based on the initial incident artifacts. Additionally, the method includes determining a plurality of secondary semantic relationships between the parsed incident artifacts based on a natural language processing of the security encyclopedia. Also, the method includes determining a plurality of influence directions corresponding to the secondary semantic relationships based on the secondary semantic relationships and the historical incidents.

Abstract: An example operation may include one or more of detecting, by a blockchain (BC) node, a contract update on a blockchain, determining, by the BC node, at least one charge affected by the contract update, executing, by the BC node, a smart contract to re-generate an invoice based on the at least one charge, and recording the invoice on a ledger of the blockchain.

Abstract: An example operation may include one or more of connecting, by a migration server, to a legacy system configured to store legacy data, identifying, by the migration server, artifacts of the legacy system subject to a migration to a blockchain, generating, by the migration server, transactions that correspond to the artifacts of the legacy system, mapping the artifacts of the legacy system to artifacts of the blockchain based on the transactions, and executing, by the migration server, a smart contract to migrate the legacy data corresponding to the mapped artifacts into the blockchain.

Abstract: Mechanisms for performing intelligent federated machine learning (ML) model updates are provided. A plurality of ML model updates, and a plurality of dataset sketch commitment data structures (sketches), are received from a plurality of participant computing systems. Each sketch provides statistical characteristics of a corresponding local dataset used by a corresponding participant to train a local ML model. A potentially malicious participant identification operation is performed based on an analysis of the plurality of sketches to identify one or more potentially malicious participants based on differences in sketches. ML model updates received from participant computing systems identified as potentially malicious participants are discarded to thereby generate a modified set of updates. The federated ML computer model is updated based on the modified set of updates.

Abstract: From data of a first interaction environment, user activity data while using the first interaction environment, and user biometric data while using the first interaction environment, a first cognitive load index is generated. A first directed acyclic graph corresponding to the first cognitive load index is generated. Using a graph convolutional network, the first directed acyclic graph, and a second directed acyclic graph generated from a second cognitive load index, a set of cognitive overload causation factors is generated. Using the set of cognitive overload causation factors, the first interaction environment is adjusted.

Abstract: Data access requests are grouped into a plurality of clusters according to an activity count within each of a plurality of time periods. Based on a time separating two clusters in the plurality of clusters, a time window in which a data access policy applies is defined. Using the plurality of data access requests, a forecasting model is trained to predict a volume of future data access requests, the training resulting in a trained forecasting model. A data access policy effective during the time window and conditioned on a threshold related to the volume of future data access requests is generated. Responsive to determining that a new data request meets a criterion of the data access policy, processing of the new data request is allowed.

Abstract: A method, apparatus and computer program product for privacy-preserving homomorphic inferencing using one-hot data representations. In this approach, a client interacting with a cloud-based server submits one-hot maps of a Chinese Remainder Theorem (CRT)-based representation of an data element, and the server expands these maps in an online phase to obtain a full one-hot map for the element. After the server obtains the full one-hot map, it performs an operation, e.g., a comparison operation associated with inferencing on a decision tree, on the one-hot map under homomorphic encryption, and in response generates a result. The result is provided back to the client.

Abstract: In an approach to obfuscation of sensitive information through non-visual feedback, responsive to receiving a signal from a user to start a session requiring input of sensitive data, one or more dummy character positions are determined, where the one or more dummy character positions represent characters that are not part of the sensitive data. Responsive to a current character position is any dummy character position of the one or more dummy character positions, the user is signaled that a next input character will be ignored, where the user is signaled using non-visual feedback. Responsive to receiving the next input character from the user, where the current character position is any dummy character position of the one or more dummy character positions, the next input character is ignored.

Abstract: An example computer-implemented method for mirroring memory in a disaggregated memory clustered environment is provided. The method includes assigning, by a hypervisor, a disaggregated memory to a virtual machine comprising a remote disaggregated memory, the virtual machine being one node of a cluster of the disaggregated memory clustered environment. The method further includes allocating, by a disaggregated memory manager, a mirrored memory for the remote disaggregated memory to mirror the remote disaggregated memory on an alternate node of the cluster of the disaggregated memory clustered environment. The method further includes responsive to a memory access occurring, maintaining, by the disaggregated memory manager, the mirrored memory. The method further includes, responsive to detecting a memory allocation adjustment, modifying, by the disaggregated memory manager, memory usage across the cluster.

Abstract: Disclosed herein is a virtual cache and method in a processor for supporting multiple threads on the same cache line. The processor is configured to support virtual memory and multiple threads. The virtual cache directory includes a plurality of directory entries, each entry is associated with a cache line. Each cache line has a corresponding tag. The tag includes a logical address, an address space identifier, a real address bit indicator, and a per thread validity bit for each thread that accesses the cache line. When a subsequent thread determines that the cache line is valid for that thread the validity bit for that thread is set, while not invalidating any validity bits for other threads.

Abstract: A redistribution layer for an integrated circuit package is provided. The redistribution layer includes a first conductive layer and a second layer disposed directly on the first conductive layer. The first conductive layer has a resistivity of less than 3.6*10?8 ?·m and has a thickness of greater than or equal to 1 ?m. The second layer includes tungsten. An integrated circuit package is also provided that includes the redistribution layer electrically connecting a first integrated circuit of the first integrated circuit package to a first input/output of a frame of the integrated circuit package. The frame is connected to the first integrated circuit. A method for manufacturing a redistribution layer is also provided.

Abstract: A method, computer system, and a computer program product are provided for restoring an interrupted communication session. In one embodiment, the methodology comprises obtaining communication interface and network information from a user required to communicate from a user device with another device. The user communication is then captured between a current session using the user device and another device. The context of the captured communication is then determined, and a token is associated with it. The information relating to the current session and its context as well as the associated token are stored. This stored information is used to re-establish a new session when the current session is terminated or interrupted prior to resolution. The new session uses the stored information to reestablish communication at the exact process stage where interruption occurred.

Abstract: At least one knowledge source for use in updating a first problem-solution data set can be selected. From data provided by the at least one knowledge source, a plurality of dialogs can be automatically generated. Existing dialogs in the first problem-solution data can be updated set using the plurality of dialogs that are generated.

Abstract: Methods, computer program products, and systems are presented.

Abstract: A system and method to tie a removable component to a host device. A first pairing key is stored into a security module on a host device such as a server rack. A removable component is inserted into the server rack for the first time. In response to this first insertion the first pairing key is burned into the removable component using a plurality of physically modifiable internal components. The server rack/security module receives a request form the removable component to operate on the server rack, the request includes a burned in pairing key. The security module compares the received pairing key with the first pairing key and permits operation of the removable component in response to a match between the received pairing key and the first pairing key.

Abstract: Systems, methods and/or computer program products maximizing answer coherence of virtual agents responding to multi-part user queries, using relationships among extracted chunks corresponding to support document(s). Relationships are extracted by mining entities from text and images of support document(s) and mapping entities to an otology. Induced subgraphs for chunks, based on extracted entities are created and compared. Depending on amounts of overlap and/or relationships between subgraphs, relationships among chunks are identified. Chunks relationships can be complementary, contained within one another, instances of one another, or not related at all. For a given multi-part query, a query answer graph is created comprising answer nodes corresponding to potential answer candidates associated with each chunk.

Abstract: A one-shot neural architecture search method referred to as MergeNAS by merging different types of convolutions into a single operation. This mergence approach reduces the search cost to roughly half a GPU-day as well as alleviates the over-fitting problem caused by a traditional differentiable architecture search (DARTS) approach by reducing the number of redundant parameters.

Abstract: A semiconductor structure may include a resistive random access memory device embedded between an upper metal interconnect and a lower metal interconnect in a backend structure of a chip. The resistive random access memory may include a first electrode and a second electrode separated by a dielectric film. A portion of the dielectric film directly above the first electrode may be crystalline. The semiconductor structure may include a stud below and in electrical contact with the first electrode and the lower metal interconnect and a dielectric layer between the upper metal interconnect and the lower metal interconnect. The dielectric layer may separate the upper metal interconnect from the lower metal interconnect. The crystalline portion of the dielectric film may include grain boundaries that extend through an entire thickness of the dielectric film. The crystalline portion of the dielectric film may include grains.

Abstract: Data transformation and data quality checking is provided by reading data from a source datastore and storing the data into memory, performing in-memory processing of the data stored in memory, where the data is maintained in-memory for performance of the in-memory processing thereof, and where the in-memory processing includes performing one or more transformations on the data stored in memory, in which the data stored in memory is transformed and stored back into the memory and applying one or more data quality rules to the data stored in-memory, and based on performing the in-memory processing of the data stored and maintained in memory for the in-memory processing, loading to a target datastore at least some of the data processed by the in-memory processing.

Abstract: A processor may receive first sensor data associated with detected movement of a user in an environment. The processor may determine, using an AI model, based on the first sensor data, that the movement of the user is associated with a first object. The processor may retrieve sensory data associated with the first object from a repository. In some embodiments, the sensory data may include at least one of auditory, olfactory, and haptic data associated with historical user interactions with the first object. The processor may provide a first sensory signal to the user to indicate a location of the first object relative to a first location of the user.

Abstract: A semiconductor device, such as an MRAM device, includes a stepped contact within a memory region that has a greater number of different contact structures (i.e., contact structures formed in different fabrication stages) relative to a logic region contact. The stepped contact includes a lower stepped contact and an upper stepped contact. The inclusion of the lower stepped contact allows for a relatively shorter upper stepped contact compared to the logic region contact. The stepped contact may allow the use of a multi-layer encapsulation spacer upon the sidewalls of a memory cell that fill out tight spacing therebetween, which may decrease the propensity of void formation and resulting shorting between neighboring memory cell features.

Abstract: A system may generate a web conference recording of a live web conference among a plurality of participant computing systems. The system may receive, at the end of the live web conference, one or more local recordings from one or more participant computing systems of the plurality of participant computing systems. A system may merge content from the one or more local recordings into the web conference recording, the content from the one or more local recordings corresponding to one or more connectivity issues of a corresponding participant computing system during the live web conference.

Abstract: An architecture for, and techniques for fabricating, a thermal decoupling device are provided. In some embodiments, thermal decoupling device can be included in a thermally decoupled cryogenic microwave filter. In some embodiments, the thermal decoupling device can comprise a dielectric material and a conductive line. The dielectric material can comprise a first channel that is separated from a second channel by a wall of the dielectric material. The conductive line can comprise a first segment and a second segment that are separated by the wall. The wall can facilitate propagation of a microwave signal between the first segment and the second segment and can reduce heat flow between the first segment and the second segment of the conductive line.

Abstract: ML model(s) are created and trained using training data from user(s) to create corresponding trained ML model(s). The training data is in FHE domains, each FHE domain corresponding to an individual one of the user(s). The trained machine learning model(s) are run to perform inferencing using other data from at least one of the user(s). The running of the ML model(s) determines results. The other data is in a corresponding FHE domain of the at least one user. Using at least the results, it is determined which of the following issues is true: the results comprise objectionable material, or at least one of the trained ML model(s) performs prohibited release of information. One or more actions are taken to take to address the issue determined to be true. Methods, apparatus, and computer program product are disclosed.

Abstract: In a method for authenticating a user, a processor receives a login request for an application. A processor directs the login request to a collection page. The collection page may include an authentication script. A processor receives a risk assessment based on an identity authenticated through the authentication script. A processor grants a level of access to the application based on the risk assessment.

Abstract: A semiconductor structure includes a first transistor device comprising a plurality of channel regions. The semiconductor structure further includes a second transistor device comprising a plurality of channel regions. The first transistor device and the second transistor device are disposed in a stacked configuration. The plurality of channel regions of the first transistor device are disposed in a staggered configuration relative to the plurality of channel regions of the second transistor device.

Abstract: Predictive dead store elimination is provided. The method comprises identifying, in a program, a first store operation and a second store operation in a program loop that comprise a store pair with a same loop-invariant base address and determining whether the store pair is a predictive dead store elimination candidate. Responsive to a determination that the store pair is a predictive dead store elimination candidate, the method eliminates the first store operation in each iteration of the program loop, except the last DSRC (dead store recurrence constant) iterations and sinks the first store operation in the last DSRC iterations to after the program loop.

Abstract: A stacked field-effect transistors (FETs) layout and a method for fabrication are provided. The stacked FETs include a buried interconnect within the stacked devices which provides power to buried components without requiring a wired connection from a top of the stacked FET to the buried components. The buried interconnect allows for efficient scaling of the stacked devices without extraneous wiring from a top of the device to each epitaxial region/device within the overall device.

Abstract: Provided are techniques for dynamically creating and playing a scenario replication for an activity on one or more visual and audio devices. One or more patterns and one or more contexts are stored. One or more goals are stored for a user. A current context is identified from data from a plurality of data sources. The current context is matched to a stored context of the one or more contexts. A pattern associated with the matched context is identified. A recommendation of an activity is provided based on the pattern and based on a goal of the one or more goals. A scenario replication is created with at least one of visual elements and audio elements for the recommendation. One or more visual and audio devices are identified. The scenario replication is played on the one or more visual and audio devices.

Abstract: A method includes forming a first nanosheet fin extending vertically from a first region of a substrate corresponding to a logic device and forming a second nanosheet fin extending vertically from a second region of the substrate corresponding to an input/output device. The first nanosheet fin includes first semiconductor channel layers vertically stacked over the first region. The second nanosheet fin includes an alternating sequence of semiconductor sacrificial layers and second semiconductor channel layers vertically stacked over the second region. An encapsulation layer is epitaxially grown along sidewalls of the alternating sequence of semiconductor sacrificial layers and second semiconductor channel layers, and an oxide layer is formed in contact with a top surface of an uppermost second semiconductor channel layer of the alternating sequence of semiconductor sacrificial layers and second semiconductor channel layers and in contact with opposite sidewalls of the encapsulation layer.

Abstract: A method, computer system, and a computer program product for automated agent intent detection enhancement are provided. A first message from a first user is received. The first message is generated during a first conversation between the first user and a first automated agent. A computer produces a second message that includes a same request as the first message but a different language modality than the first message. The second message and the first message are combined to form a combined message. The combined message is input into the first automated agent such that the first automated agent produces an intent classification for the first message.

Abstract: Mechanisms are provided to implement a neural flow attestation engine and perform computer model execution integrity verification based on neural flows. Input data is input to a trained computer model that includes a plurality of layers of neurons. The neural flow attestation engine records, for a set of input data instances in the input data, an output class generated by the trained computer model and a neural flow through the plurality of layers of neurons to thereby generate recorded neural flows. The trained computer model is deployed to a computing platform, and the neural flow attestation engine verifies the execution integrity of the deployed trained computer model based on a runtime neural flow of the deployed trained computer model and the recorded neural flows.

Abstract: A plurality of tickets can be received. A pattern pertaining to an issue with an application can be determined by at least one artificial neural network. The pattern can be indicated among at least a portion of the plurality of tickets. The portion of the plurality of tickets can be grouped into a group comprising a primary ticket and at least one secondary ticket. Automation steps can be accessed from a database. The automation steps can be implemented to resolve the issue for the primary ticket. Responsive to the resolving the issue by implementing the automation steps from the database for the primary ticket, the secondary ticket can be closed.

Abstract: A method, computer system, and a computer program product for video conference optimization is provided. The present invention may include accessing data for a scheduled web conference. The present invention may include generating a visual display template for each of a plurality of participants of the scheduled web conference. The present invention may include monitoring the scheduled web conference. The present invention may include adjusting the visual display template for at least a portion of the plurality of participants during the scheduled web conference.

Abstract: An audio emitter configured to emit a sound creates a high-frequency copy of the sound to be emitted. The high-frequency copy of the sound is superimposed over the sound, resulting in a composite signal. The composite signal is emitted by the emitter. The high-frequency copy is at a frequency inaudible to humans, enabling a receiver to identify the emitter and/or the sound.

Abstract: Embodiments disclosed herein describe semiconductor devices that include semiconductor structures and methods of forming the semiconductor structures. The semiconductor structures may include an upper conductive line, a first lower conductive line laterally insulated by a first lower dielectric region and a second lower dielectric region. The semiconductor structure also includes a lower level via region above the first lower conductive line. The lower level via region includes a dielectric blocking material and a spacer material.

Abstract: A computer-implemented method for learning unknown concepts during natural language processing is disclosed, including identifying a sentence associated with an unknown concept, selecting a first sequential set of sentences from a first document, including the sentence associated with the unknown concept, one sentence prior, and subsequent to the sentence associated with the unknown concept, selecting a second sequential set of sentences from a second document, including a sentence associated with a known concept, and one sentence prior and subsequent to the sentence associated with the known concept, comparing concepts associated with the first sequential set of sentences and second sequential set of sentences, determining whether an inference can be made between the unknown concept associated with the sentence from the first document and the sentence associated with the known concept associated with the sentence from the second document, and tagging the unknown concept associated with the known concept.