A method for detecting an anomaly in resource utilization observed within a cloud computing platform includes observing an actual resource utilization for a customer of the cloud computing platform; determining a historical utilization distribution for the customer that defines values of a resource utilization metric across repeated instances of a time interval and sub-intervals therein; identifying a temporal location of an anomaly detection period within the time interval; filtering the historical utilization distribution to construct a distribution of relevant values of the resource utilization metric; and computing, based on the distribution of relevant values, a resource utilization projection for the customer. An anomaly risk metric in calculated and fit to an anomaly classification to determine the presence of a data usage anomaly. The data usage anomaly is confirmed the using a language model trained on prior data usage anomalies over prior time intervals tagged with actual unauthorized access.
H04L 41/16 - Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
2.
DIGITAL RIGHTS MANAGEMENT ARCHITECTURE FOR ARTIFICIAL INTELLIGENCE MODELS
Various technologies pertaining to digital rights management (DRM) for artificial intelligence (AI) models are provided. In an example, a client computing device comprises a first processor and one or more second processors. Subsequent to transmitting a request for access to a computer-implemented AI model, the client computing device receives an encryption key, AI model certification information, and an AI model payload, wherein the AI model payload comprises an encrypted AI model. The client computing device validates the AI model payload based upon the AI model certification information and decrypts the AI model payload. The client computing device stores the decrypted AI model payload in a second memory where a first memory stores a DRM application. Responsive to a request to execute the AI model stored in the second memory, the computing device causes execution of the AI model by the one or more second processors.
The virtually divided input trackpad disclosed herein is designed to provide enhanced ergonomic user interaction with digital interfaces. The virtually divided input trackpad comprises two virtually separated functional areas, each of which may be dedicated to distinct functionalities such as object movement and object rotation. This arrangement allows for simultaneous two-handed operation, offering users an intuitive, efficient, and accessible way of controlling digital environments, which is especially beneficial for users with specific accessibility needs.
Aspects of embodiments of the present disclosure relate to a field programmable gate array (FPGA) configured to implement an exponential function data path including: an input scaling stage including constant shifters and integer adders to scale a mantissa portion of an input floating-point value by approximately log2 e to compute a scaled mantissa value, where e is Euler's number; and an exponential stage including barrel shifters and an exponential lookup table to: extract an integer portion and a fractional portion from the scaled mantissa value based on the exponent portion of the input floating-point value; apply a bias shift to the integer portion to compute a result exponent portion of a result floating-point value; lookup a result mantissa portion of the result floating-point value in the exponential lookup table based on the fractional portion; and combine the result exponent portion and the result mantissa portion to generate the result floating-point value.
G06F 7/483 - Computations with numbers represented by a non-linear combination of denominational numbers, e.g. rational numbers, logarithmic number system or floating-point numbers
Cross-correlating identifiers are created and stored in a security asset database for corresponding security assets. The security asset database is used with the cross-correlating identifiers to perform functions for the underlying security assets without exposing the security assets. One function is a scan of network resources for potential exposure of the security assets. A key generation algorithm used to generate the cross-correlating identifiers is applied to any identified potential security data to generate a corresponding reference identifier. A determination is made whether a security asset comprises a security risk that exceeds a predetermined threshold based on at least a determination of whether the cross-correlating identifier for the security asset matches any reference identifiers generated for the potential security data, as well as based on other data stored in the security asset database for the security asset.
H04L 9/32 - Arrangements for secret or secure communicationsNetwork security protocols including means for verifying the identity or authority of a user of the system
The present disclosure proposes a method, apparatus, and computer-readable medium for model checkpoint saving based on multi-tier storage. During a training of a machine learning model performed through a Graphics Processing Unit (GPU) in a target node, a checkpoint to be saved of the machine learning model may be identified from a GPU memory that directly exchanges data with the GPU. The checkpoint may be saved from the GPU memory to a central processing unit (CPU) memory that directly exchange data with a CPU in the target node. The checkpoint may be saved from the CPU memory to a non-transitory memory, the non-transitory memory including at least one of: a local non-transitory memory in the target node, a neighbor non-transitory memory in a neighbor node of the target node, and a remote non-transitory memory located remotely from the target node.
Disclosed herein is the compound CCC═C(F)CF. The compound can be provided as a pure isomer of either the E-isomer or Z-isomer, or as a diastereomeric mixture of both isomers (i.e., an E/Z mixture). The boiling point and dielectric constant of the compound enable effective use of the compound as a coolant in a two-phase immersion cooling system. A two-phase immersion cooling system incorporating the compound as coolant can include: an immersion tank configured to contain the coolant and to contain a component capable of generating heat; and a condenser configured to receive vaporized coolant and to condense the vaporized coolant back to liquid form.
A wearable device comprises an exteriorly positioned first electrode and a reporting capacitor. The first electrode forms a first side of the reporting capacitor, and a second side of the reporting capacitor is formed by skin of a user when the wearable device is worn. An oscillator is configured to output a signal to drive the first electrode at a first frequency. The oscillator is configured such that changes in capacitance at the reporting capacitor adjust the signal output by the oscillator from the first frequency to a second frequency. A frequency-to-voltage converter is configured to generate a voltage representation of the second frequency. A controller determines a change between the first frequency and the second frequency based on the voltage representation and indicates an amount of movement of skin of the user relative to the first electrode based on the determined frequency change.
Systems and methods for service incident detection in a cloud computing platform. According to an example implementation, the incident detection system retrieves a single user metric corresponding to a service call to a resource on which the service is dependent and uses unsupervised anomaly detection to detect anomalies indicative of a service incident. Detected anomalies include an anomaly score indicating a level of anomality. Additionally, a supervised learning classifier is trained and used to filter/classify the anomaly detection results based on features corresponding to the anomaly score. The features are learned based on characteristic dimensions, distribution, and statistics of anomaly scores of the user metric at different resolution/aggregation levels. Anomaly detection results are classified as an incident or not an incident. A report is generated for a determined incident.
H04L 41/5061 - Network service management, e.g. ensuring proper service fulfilment according to agreements characterised by the interaction between service providers and their network customers, e.g. customer relationship management
H04L 41/5074 - Handling of user complaints or trouble tickets
10.
DYNAMICALLY SUBSTITUTING A MODIFIED QUERY BASED ON PERFORMANCE ANALYSIS
The disclosure herein describes analyzing queries and dynamically modifying those queries based on the analysis. An indication that a query is to be executed by a first process is detected. It is determined that an analysis results data store does not include an active analysis result for the query using a query identifier of the query and, as a result, a modified instance of the query is generated using a modification pattern. The query and the modified instance of the query are analyzed based on a performance metric using a second process that is independent of the first process. An active analysis result of the query is recorded based on the analysis, wherein the analysis result indicates whether future executions of the query should be modified using the modification pattern. Further, in some examples, analysis results expire, such that associated queries are reanalyzed to generate active analysis results periodically.
This document relates to generative machine learning. Users can provide input relating to a content generation task. A generative machine learning model can be prompted to suggest concept tags based on the received user input. Then, users can select a concept tag and a value for that concept tag. A content item can be generated for the user using the generative machine learning model, where the content generation is based on the value for the selected concept tag. Over time, additional content tags can be updated with user-designated values while updating the generated content accordingly. In this manner, user generation of content with a generative machine learning model can be guided via received user inputs.
G06F 3/0482 - Interaction with lists of selectable items, e.g. menus
G06F 3/04845 - Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range for image manipulation, e.g. dragging, rotation, expansion or change of colour
G06T 11/60 - Editing figures and textCombining figures or text
12.
Reducing Computational Burden in a Generative Model through Target Vocabulary Constraints
A technique generates a response based on a query using a generative model. The technique includes encoding the query into a shortlist embedding and a sequence of response-part embeddings. The technique uses the shortlist embedding to identify a reduced-size vocabulary that is relevant to the query, selected from a larger target vocabulary of tokens. The technique then identifies at least one group of ranked tokens associated with a corresponding response-part embedding. The group of ranked tokens is selected from the reduced-size vocabulary. The technique then constructs a part of a response in a manner that is constrained by the group of ranked tokens. Such constraint helps reduce computational burden and latency. Some implementations construct the response non-autoregressively in a single pass, while others perform this operation autoregressively in plural passes. Some implementations of the target vocabulary include plural-word tokens, each including two or more words.
A delay optimizer includes circuits for detecting delay between the received data and the received clock, such as in an integrated circuit having separate dies coupled to communication via an interconnect that includes a clock channel and a data channel. The electrical characteristics of the clock channel and the data channel (including on-chip buffers) may introduce significant differences in the delay between the received clock and the received data coupled with the effects of clock jitter, inter-symbol interference, and duty-cycle distortion that may introduce significant sampling errors in the sampled data. The delay optimizer operates to detect the delay and optimize a sampling clock in real-time and on a continuous basis using the difference between the data edge of the received data and the received clock edge to determine an optimal delay value to sample the received data to produce the sampled data with the desired very-low bit-error-rate (BER).
Disclosed herein is the compound CCC=C(F)CF. The compound can be provided as a pure isomer of either the E-isomer or Z-isomer, or as a diastereomeric mixture of both isomers (i.e., an E/Z mixture). The boiling point and dielectric constant of the compound enable effective use of the compound as a coolant in a two-phase immersion cooling system. A two-phase immersion cooling system incorporating the compound as coolant can include: an immersion tank configured to contain the coolant and to contain a component capable of generating heat; and a condenser configured to receive vaporized coolant and to condense the vaporized coolant back to liquid form.
Techniques are provided for improving chat response generation using knowledge graph-based data retrieval. A chat system receives a user message and identifies relevant entities by performing a web search. A first generative language model receives a prompt containing the chat history, identified entities, and a knowledge graph schema defining entity types and relationships. The first model generates a structured query targeting specific entity attributes in the knowledge graph. After executing the query to retrieve targeted entity data, a second generative language model receives the retrieved data and user message to generate a contextually relevant response. The system enables precise control over grounding data by using the knowledge graph schema to specify exactly which entity attributes to retrieve, avoiding excessive or irrelevant information while maintaining comprehensive responses. This approach improves upon conventional database solutions by allowing flexible, relationship-aware queries that retrieve diverse yet focused entity information based on conversational context.
Various security mechanisms are considered for detecting and mitigating potential security risks posed by generative artificial intelligence. In one example, a generative model prompt is separated into a meta prompt part and an input prompt part, which in turn are separately encoded. Based on the resulting meta prompt embedding vector and input prompt embedding vector, the prompt is identified as anomalous, which in turn triggers an appropriate security action. In one example implementation, the meta prompt embedding vector is used to classify the prompt (e.g. by application type or application flow type), and the input prompt embedding vector is used for context-aware anomaly detection, using a class assigned to the prompt based on its meta prompt embedding vector. In another example implementation, the prompt is identified as anomalous based on distance between the meta prompt and input prompt embedding vector.
17.
EMBEDDING DYNAMIC CONTENT IN VIDEO DATA ALLOWING REAL-TIME INTERACTION VIA CONFIGURATION CHANGES DURING RENDERING
A method for embedding dynamic content into video data. The method comprises detecting, by a video editor, an indication that a piece of dynamic content is intended be rendered in a dynamic video and determining at least one dynamic variable associated with the piece of dynamic content. The method further comprises detecting a selection of a selected dynamic variable of the at least one dynamic variable and determining a dynamic interaction associated with the selected dynamic variable. The method further comprises generating the dynamic video to comprise interactive dynamic content based on the piece of dynamic content and a static video, and labeling, in metadata of a dynamic video using a backwards-compatible video format, that the dynamic interaction is configured to be performed on the interactive dynamic content to modify the selected dynamic variable during playback of the dynamic video.
This document relates to secure storage techniques that can be employed in a multi-tenant environment. In the disclosed implementations, a host CPU can perform data at rest encryption on data that is written to an SSD. By offloading the data at rest encryption from the SSD to the CPU, the complexity and power consumption of the SSD can be reduced. Furthermore, redundant cryptographic operations can be mitigated, because the SSD and CPU do not necessarily perform link encryption on data payloads themselves. Rather, link encryption can be limited to associated command data used to configure transfers of encrypted data payloads.
An electronic device hinge includes a first body, a second body, and a link. The link is rotatable relative to the first body around a first pivot point and rotatable relative to the second body around a second pivot point. The first pivot point has a first rotational resistance and the second pivot point has a second rotational resistance that is different from the first rotational resistance. The hinge further includes a third body that is selectively positionable relative to the first body in a first configuration. The third body limits a first rotational range of motion around the first pivot point when positioned in the first configuration.
Aspects of the technology disclosed herein related to a distributed architecture for securely delivering AI models and/or training data sets to client devices for local use. The distributed includes a licensing server that controls access to and decryption of the models. The licensing server controls the distribution of licensing packages for the different models delivered by the distribution server. The client device transmits a license request to the licensing server. The licensing request may include device-level details about the client device itself, and such details may be provided in a secure, trusted manner, such as through a hardware root of trust (HROT) of the client device. If the details in the license request satisfy the security requirements for the model, a license package for the model is delivered to the client device. The license package includes a license for the model and a decryption key for the model.
Techniques for dynamically compressing neural network models enable efficient deployment on resource-constrained devices. A compression manager monitors device metrics and application requirements to determine when compression is needed. Model-specific compression instructions guide sequential operations including quantization, layer fusion, pruning, and aggressive pruning, with each operation's parameters specified per layer. Compression metadata preserves information needed for potential restoration. When resources become constrained, the system progressively applies compression while maintaining critical model capabilities. As resources become available, compressed components can be selectively restored using stored metadata. The compression level adapts automatically based on real-time conditions, optimizing the balance between model size and performance.
A structured query language (SQL) query including an SQL intrinsic function is processed using native code including single instruction multiple data (SIMD) or single instruction multiple thread (SIMT) processor instructions for execution on a processor having native parallelism. The native code is compiled from an implementation of the SQL intrinsic function in a platform-independent source code. The compiling comprises compiling the source code to generate a platform-independent intermediate representation (IR) of the source code. The IR is optimized for improved performance through parallelization. The optimized IR is lowered to generate the native code.
Embodiments of the present disclosure include techniques for moving data between electronic system components using buffers. A digital processor stores data generated in response to a series of commands in a first buffer. Commands are received with a reference to a second buffer. The digital processor tracks the last location that a data result was stored in the first buffer. When a data result fills the buffer, the remaining data is automatically stored in the second buffer. Downstream devices may empty full buffers. A client may receive an indication that a buffer is empty and subsequently send commands with a reference to empty buffer.
This document relates to secure storage techniques that can be employed in a multi-tenant environment. In the disclosed implementations, a host CPU can perform data at rest encryption on data that is written to an SSD via an interposer. By offloading the data at rest encryption from the interposer to the CPU, the complexity and power consumption of the interposer can be reduced. Furthermore, redundant cryptographic operations can be mitigated, because the interposer and CPU do not necessarily perform link encryption on data payloads themselves. Rather, link encryption can be limited to associated command data used to configure transfers of encrypted data payloads.
H04L 9/06 - Arrangements for secret or secure communicationsNetwork security protocols the encryption apparatus using shift registers or memories for blockwise coding, e.g. D.E.S. systems
25.
DETERMINING GENERATIVE SEARCH RESULTS DOCUMENT FOR QUERIES USING GENERATIVE ARTIFICIAL INTELLIGENCE MODELS AND ARBITRATION MODELS
This disclosure describes utilizing a generative document system to create generative search results documents using generative artificial intelligence (AI) models and dynamically determining which one of the generative search results documents to provide in response to a search query. For example, in response to receiving a search query, the generative document system obtains search link results (e.g., website links and corresponding grounding information) for the search query and utilizes this information with multiple generative AI models to generate various types of generative search results documents. Additionally, the generative document system generates and utilizes a generative document arbitration model to determine, based on the search link results, which of the generative search results documents to provide in response to the search query.
Examples are disclosed that relate to performing reasoning processes using generative language models. One disclosed example provides a method of performing a spatial reasoning task. The method comprises, iteratively, at a reasoner agent, receiving query results from a retriever agent, and based upon the query results, generating a reasoner prompt. The method further comprises inputting the reasoner prompt into a reasoner language model, receiving a reasoner output from the reasoner language model, and sending a query to a retriever agent. The method further comprises, at the retriever agent, receiving the query from the reasoner agent, generating a retriever prompt, and inputting the retriever prompt into a retriever language model. The method further comprises receiving an output from the retriever language model, querying scene data, and receiving one or more results of the query, and sending the one or more results of the query to the reasoner agent.
A set of geometric shapes to be applied by a machine learning model to objects identified in image data is defined. A learning rate of the machine learning model is updated in response to external events. The machine learning model is used to estimate spatial parameters for each of the objects identified in the image data. The spatial parameters are estimated by fitting the objects to the set of geometric shapes. Updates to the spatial parameters are temporally integrated. A spatial estimate of the objects identified in the image data is generated.
The automatic generation of synthetic training data that can be used to train a language model to generate code examples following a code language based on a natural language input. Thus, new language models may be created, or existing language models may be fine-tuned, to adapt to automatically generate code without having to manually generate bulk quantities of training data. Rather, a many-to-many grammar mapping is navigated to generate training data. Specifically, the many-to-many grammar mapping maps code grammar to natural grammar. Then, each training data is generated by navigating the many-to-many grammar mapping definition to generate a mapping of a respective code expression to a respective natural language expression.
A method, computer program product, and computing system for executing a plurality of requests to process data using a trained machine learning model. An anomalous pattern of requests including at least a threshold amount of out-of-domain data is identified from the plurality of requests. A potential model inversion attack is detected based upon, at least in part, identifying the anomalous pattern of requests.
A system for providing a content holding layer includes a processing system and memory storing instructions that, when executed by the processing system, cause the system to display a content holding element on a desktop of an operating system, receive an input for adding a content item to the content holding element, obtain content item data of the content item, add the content item data to a database associated with the content holding element, display a content holding application on the desktop, display a representation of the content item in the content holding application, determine one or more recommended actions applicable to the content item based at least in part of the content item data, and display, one or more selectable elements for executing the one or more recommended actions.
G06F 9/451 - Execution arrangements for user interfaces
G06F 3/0482 - Interaction with lists of selectable items, e.g. menus
G06F 3/0484 - Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range
31.
KNOWLEDGE DISTILLATION USING HYBRID LOSS FUNCTION FOR DIFFERENT SAMPLE TYPES
A technique is described for training a student model based on a larger teacher model. The training includes generating a loss measure having a contrastive combination of two parts. The first part is based on a forward measure of divergence between teacher-generated and student-generated probability distributions, which, in turn, are based on teacher-generated samples. The second part is based on a reverse measure of divergence between student-generated and teacher-generated probability distributions, which, in turn, are based on student-generated samples. The technique then updates parameters of the student model based on the loss. In some implementations, the first part of the loss is generated using forward Kullback-Leibler (KL) divergence, and the second part of the loss is generated using reverse KL divergence. The technique also involves dynamically updating hyper-parameters during training.
Head-level key-value (KV) cache compression may be performed by allocating memory to individual heads of a multi-head attention model based on importance scores assigned to the individual heads of a multi-head attention model. The importance scores may be calculated based on both retrieval and reasoning attributes of the individual heads of the multi-head attention model. The retrieval attributes and reasoning attributes of the individual heads of the multi-head attention model may be determined based on an updated Needle-in-a-Haystack test that incorporates a Retrieval-Reasoning example. It should be appreciated that the reasoning attributes of the individual heads are different than attention scores associated with the multi-head attention model.
Disclosed solutions provide for interactive video editing and playback with three dimensional (3D) object manipulation. Examples enable video players to display both an underlying static video along with a 3D object as dynamic content. The video editor presents a settings editor that enables the creator of the video to specify the ability of viewers to interact with the 3D object. The video viewer exposes settings for the dynamic content to enable users to reconfigure the display of the 3D dynamic content, making the video rendering an interactive experience. Use of references (e.g., URLs) within the dynamic content enables videos distributed in the new format updateable and correctable, such that information that is subject to change may be kept current, and informational errors introduced at the time of the video production may be corrected – without requiring creation and distribution of a substitute video file.
G11B 27/031 - Electronic editing of digitised analogue information signals, e.g. audio or video signals
G11B 27/32 - IndexingAddressingTiming or synchronisingMeasuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording on separate auxiliary tracks of the same or an auxiliary record carrier
Storage format solutions are disclosed for interactive video editing and playback that provide backwards compatibility for legacy players. Examples enable newer video players, that are able to extract dynamic content from the new video file format, to display both the underlying static video along with the dynamic content (according to a timeline within metadata stored in the new video file format), whereas legacy players display the static video. Some examples expose settings for the dynamic content to enable newer players to reconfigure the display of the dynamic content, making the video rendering an interactive experience. Use of references (e.g., URLs) within the dynamic content enables videos distributed in the new format updateable and correctable, such that information that is subject to change may be kept current, and informational errors introduced at the time of the video production may be corrected – without requiring creation and distribution of a substitute video file.
G11B 27/031 - Electronic editing of digitised analogue information signals, e.g. audio or video signals
G11B 27/32 - IndexingAddressingTiming or synchronisingMeasuring tape travel by using information detectable on the record carrier by using information signals recorded by the same method as the main recording on separate auxiliary tracks of the same or an auxiliary record carrier
H04N 9/82 - Transformation of the television signal for recording, e.g. modulation, frequency changingInverse transformation for playback the individual colour picture signal components being recorded simultaneously only
To avoid inconsistencies caused by clock skew across data storage tiers, a point-in-time restore service creates a restore account, restores a partition of the database to its state at the designated point-in-time, including data from that time. A restored value is read from the data and a list is retrieved of segments written to a disaggregated storage tier by the existing partition. The disaggregated storage tier includes a first segment with a first value and a second segment with a second value, the values indicating their creation order. A first comparison is made between the restored value and the first value, and a second comparison is made between the restored value and the second value. Based on these comparisons, the first segment but not the second segment is copied to the restore account.
G06F 16/27 - Replication, distribution or synchronisation of data between databases or within a distributed database systemDistributed database system architectures therefor
36.
SPECIALIZED SUB-TASK MODELS USED TO PERFORM A TARGET TASK
Embodiments of the disclosed technologies are capable of deploying a sequence of sub-task models to perform a target task. A query is received that includes a digital content item with a criterion. A task responsive to the query is determined. A first sub-task and second sub-task are generated from the task. The first sub-task includes a classification task related to a user and the criterion. The second sub-task includes a content generation task related to the classification task. The first sub-task is performed by determining a classification for the user with respect to the criterion. The second sub-task is performed by determining a natural text explanation for the classification. The classification and the natural language text explanation are presented via a user interface.
A technique transforms an original model into an in-place mixture-of-experts model. To accomplish this, the technique first identifies at least one group of subnetworks that have different task-processing capabilities. The subnetworks are associated with respective groups of parameters. The technique then produces a router-supplemented model that includes a router that is capable of selecting a subset of the subnetworks to be used in processing a particular instance of input information. The router determines when a particular subnetwork should be selected based on a combination of two score parts. A first score part is based on token-related hidden state information, and a second score part is based on an assessed saliency of the particular subnetwork. The technique then fine-tunes the router-supplemented model, to produce the mixture-of-experts model. In inference, the mixture-of-experts model selects among the group of subnetworks using the router in a resource-efficient and low-latency manner.
A delay optimizer includes circuits for detecting delay between the received data and the received clock, such as in an integrated circuit having separate dies coupled to communication via an interconnect that includes a clock channel and a data channel. The electrical characteristics of the clock channel and the data channel (including on-chip buffers) may introduce significant differences in the delay between the received clock and the received data coupled with the effects of clock jitter, inter-symbol interference, and duty-cycle distortion that may introduce significant sampling errors in the sampled data. The delay optimizer operates to detect the delay and optimize a sampling clock in real-time and on a continuous basis using the difference between the data edge of the received data and the received clock edge to determine an optimal delay value to sample the received data to produce the sampled data with the desired very-low bit-error-rate (BER).
H03K 5/14 - Arrangements having a single output and transforming input signals into pulses delivered at desired time intervals by the use of delay lines
H03K 5/00 - Manipulation of pulses not covered by one of the other main groups of this subclass
H03K 5/135 - Arrangements having a single output and transforming input signals into pulses delivered at desired time intervals by the use of time reference signals, e.g. clock signals
39.
ASYNCHRONOUS SERVING ARCHITECTURE FOR CUSTOMIZED CONTENT ITEMS
Custom content generation techniques for connection networking are described. A method comprises receiving a first signal indicating an entity session associated with an entity identifier, retrieving a first content item associated with the entity identifier from a memory cache, presenting the first content item in a first content slot of a first section of a graphical user interface (GUI) in response to the first signal, wherein the first section is in a rendered section of the GUI, generating a second content item associated with the entity identifier using a generative artificial intelligence model in response to the first signal, determining whether the second content item is received, and assigning the second content item to a second content slot of a second section of the GUI when the second content item is received, wherein the second section is in a non-rendered section of the GUI.
A computing system for fragment-based quantum mechanical calculation of protein properties is provided. A processor implements a protein fragmentation module that separates a computer-readable polypeptide sequence into a plurality of data units. For each subsequence of three adjacent amino acids in the polypeptide sequence, a first amino acid, a second amino acid, and a third amino acid are identified, each amino acid having a respective main chain including an amino group, a carbon, and a carboxyl group, and a side chain attached to the alpha carbon. The protein fragmentation module generates a data unit representing a first alpha carbon, a first carboxyl group, a second amino group, a second alpha carbon, a second carboxyl group, a second side chain, a third amino group, and a third alpha carbon, and stores the generated data unit in the memory.
G16B 5/00 - ICT specially adapted for modelling or simulations in systems biology, e.g. gene-regulatory networks, protein interaction networks or metabolic networks
G06N 10/80 - Quantum programming, e.g. interfaces, languages or software-development kits for creating or handling programs capable of running on quantum computersPlatforms for simulating or accessing quantum computers, e.g. cloud-based quantum computing
G16B 40/00 - ICT specially adapted for biostatisticsICT specially adapted for bioinformatics-related machine learning or data mining, e.g. knowledge discovery or pattern finding
41.
DIGITAL PHASE-LOCKED LOOPS (PLL) INCLUDING CLOSED-LOOP TIME-TO-DIGITAL CONVERTER (TDC) GAIN CALIBRATION CIRCUITS AND RELATED METHODS
In a calibrated digital phase-locked-loop (DPLL) circuit, during a normal operating mode, a control value provided to a digitally controlled oscillator (DCO) is updated by a feedback circuit to keep an output clock generated by the DCO synchronized with a reference clock. The feedback circuit includes a time-to-digital converter (TDC) circuit to measure a phase difference as a time interval. In a calibration operating mode of the calibrated DPLL circuit, calibration of a resolution of a time measurement of the time interval measured by the TDC is performed in the feedback circuit while the control value provided to the DCO is kept constant. Calibrating the TDCs in each of the DPLLs in an integrated circuit (IC) to a nominal resolution in this manner improves synchronization of the clock domains. In some examples, the TDC circuit is a Vernier type circuit and calibration sets a delay difference to a nominal resolution.
H03L 7/107 - Details of the phase-locked loop for assuring initial synchronisation or for broadening the capture range using a variable transfer function for the loop, e.g. low pass filter having a variable bandwidth
G04F 10/00 - Apparatus for measuring unknown time intervals by electric means
H03L 7/085 - Details of the phase-locked loop concerning mainly the frequency- or phase-detection arrangement including the filtering or amplification of its output signal
H03L 7/099 - Details of the phase-locked loop concerning mainly the controlled oscillator of the loop
42.
SYSTEMS AND METHODS FOR ISOLATING FAULTS IN DIE-TO-DIE INTERCONNECTS
Systems and methods for isolating faults in die-to-die interconnects are provided. A method includes providing a first transmission path, along a die-to-die interconnect, from a transmitter associated with a first die to an asynchronous buffer associated with a second die. The method further includes providing a second transmission path from voltage reference circuitry associated with the second die to the asynchronous buffer associated with the second die. The method further includes simultaneously enabling both the first transmission path and the second transmission path to allow the asynchronous buffer to receive inputs from both the transmitter associated with the first die and the voltage reference circuitry associated with the second die, such that the inputs received by the asynchronous buffer are indicative of: (1) no failure in the die-to-die interconnect, (2) an open failure in the die-to-die interconnect, or (3) a short failure in the die-to-die interconnect.
The disclosed concepts relate to contextualization of generative language models. In some implementations, a linked entity database is populated with entity resource identifiers of entities extracted from a search log by an entity linker. A contextualized prompt data structure is generated based on the linked entity database, e.g., by including linked entity context information in the contextualized prompt data structure. A response to the contextualized prompt data structure is received, where the response is conditioned on the linked entity context information.
Systems and methods are provided for handing off execution of an application from a local computing device to a cloud-based computing device. The disclosed technology is directed to determining whether and when to initiate handing off the execution of the application based on monitoring resource consumption of the local computing device. When the application is not previously installed on the cloud-based computing device, the local computing device transmits an application installer executable to the cloud-based computing device for enabling use of the same application on the cloud-based computing device.
A computing device (10) including a system-on-a-chip (SoC) (12). The SoC includes a plurality of logic circuit blocks, including synchronization blocks (22) and hardware accelerator blocks (24). A synchronization block is configured to receive a wait request (30) from a first hardware accelerator block. The wait request includes one or more semaphores (32) and one or more wait threshold values (34). The synchronization block is configured to store the wait request. The synchronization block is configured to receive, from a signal source block (42), a signal request (50) that indicates a semaphore included among the one or more semaphores in the wait request. In response to receiving the signal request, the synchronization block is configured to update the semaphore. The synchronization block is configured to determine that the updated value (52) of the semaphore has reached the wait threshold value and to transmit a wait completion response (54) to at least the first hardware accelerator block.
Systems and methods to provide vulnerability detection and management in a cloud computing system according to examples. More specifically, a vulnerability detection system receives access logs of a package repository and records information that links packages downloaded from the package repository to the computing assets that downloaded the packages. The vulnerability detection system evaluates an inventory of the package repository against a report of identified vulnerabilities (e.g., Common Vulnerabilities and Exposures (CVEs)). When a vulnerable package is identified in the inventory, the vulnerability detection system removes the vulnerable package from the package repository. The vulnerability detection system further determines affected assets and contact information of corresponding users and provides notifications to the users. In some examples, a mitigation or remediation is determined and provided to the users.
G06F 21/55 - Detecting local intrusion or implementing counter-measures
G06F 21/57 - Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
47.
ISOLATED PLATFORM FOR RESPONSIBLE ARTIFICIAL INTELLIGENCE
An isolated platform for responsible AI is described. In various examples, a method is performed by a computing device. Output is generated by executing at least part of an application in an isolated virtual machine with an input-output virtualized accelerator, where the application is an artificial intelligence application. The isolated virtual machine is used to check the output. In response to the check being successful, the output is returned. In response to the check being unsuccessful, an error message is returned.
G06F 21/53 - Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems during program execution, e.g. stack integrity, buffer overflow or preventing unwanted data erasure by executing in a restricted environment, e.g. sandbox or secure virtual machine
G06F 9/455 - EmulationInterpretationSoftware simulation, e.g. virtualisation or emulation of application or operating system execution engines
G06F 21/56 - Computer malware detection or handling, e.g. anti-virus arrangements
G06F 21/57 - Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
48.
INFERENCE ACCELERATION OF A MODEL USING AN IN-PLACE MIXTURE-OF-EXPERTS
A technique transforms an original model into an in-place mixture-of-experts model. To accomplish this, the technique first identifies at least one group of subnetworks that have different task-processing capabilities. The subnetworks are associated with respective groups of parameters. The technique then produces a router-supplemented model that includes a router that is capable of selecting a subset of the subnetworks to be used in processing a particular instance of input information. The router determines when a particular subnetwork should be selected based on a combination of two score parts. A first score part is based on token-related hidden state information, and a second score part is based on an assessed saliency of the particular subnetwork. The technique then fine-tunes the router-supplemented model, to produce the mixture-of-experts model. In inference, the mixture-of-experts model selects among the group of subnetworks using the router in a resource-efficient and low-latency manner.
An example may, at a first device, input an entity embedding for an entity and an item embedding for a plurality of items to a scoring function. The entity embedding and the item embedding are pre-computed using a first language model. At the first device, a retrieval score for the entity and a first item of the plurality of items is computed. The retrieval score is used to identify an item subset of the plurality of items. A ranking prompt is input to a second language model. The second language model and the first language model have a common parameter value. The second language model generates a ranking score for the entity and a second item in response to the ranking prompt. An online system uses the ranking score to include or exclude the second item from a presentation of digital content to the entity via a device.
Artificial intelligence techniques for connection networking are described. A method comprises receiving a request for a set of content items for a content feed, generating a set of metrics for a first set of content items of a first type and a second set of candidate content items of a second type using a machine learning model, selecting a first content item of the first type from the first set of content items and a second content item of the second type from the second set of content items based on the set of metrics using a blending algorithm to form a blended set of content items, allocating the first content item and the second content item from the blended set of content items to multiple slots in the content feed, and presenting the blended set of content items within the content feed on a GUI of a device.
A model output evaluator may, for each ground truth text of ground truth texts, link response entities of a language model output text and ground truth entities of the ground truth text to corresponding ontology entities of an ontology that includes the set of ontology entities and edges connecting the ontology entities. The evaluator may, for each ground truth text, determine a ground truth text score based on traversal distances within the ontology between each linked response entity and one or more linked ground truth entities of the ground truth text, wherein the traversal distances are calculated based on a number of edges traversed within the ontology between the linked response entity and the one or more linked ground truth entities. The evaluator may classify the output text of the language model based on at least one of the ground truth text scores satisfying a classification condition.
Disclosed solutions provide for interactive video editing and playback with three dimensional (3D) object manipulation. Examples enable video players to display both an underlying static video along with a 3D object as dynamic content. The video editor presents a settings editor that enables the creator of the video to specify the ability of viewers to interact with the 3D object. The video viewer exposes settings for the dynamic content to enable users to reconfigure the display of the 3D dynamic content, making the video rendering an interactive experience. Use of references (e.g., URLs) within the dynamic content enables videos distributed in the new format updateable and correctable, such that information that is subject to change may be kept current, and informational errors introduced at the time of the video production may be corrected—without requiring creation and distribution of a substitute video file.
G06F 3/04845 - Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range for image manipulation, e.g. dragging, rotation, expansion or change of colour
G06F 3/04847 - Interaction techniques to control parameter settings, e.g. interaction with sliders or dials
Storage format solutions are disclosed for interactive video editing and playback that provide backwards compatibility for legacy players. Examples enable newer video players, that are able to extract dynamic content from the new video file format, to display both the underlying static video along with the dynamic content (according to a timeline within metadata stored in the new video file format), whereas legacy players display the static video. Some examples expose settings for the dynamic content to enable newer players to reconfigure the display of the dynamic content, making the video rendering an interactive experience. Use of references (e.g., URLs) within the dynamic content enables videos distributed in the new format updateable and correctable, such that information that is subject to change may be kept current, and informational errors introduced at the time of the video production may be corrected—without requiring creation and distribution of a substitute video file.
A method for embedding dynamic content into video data. The method comprises detecting, by a video editor, an indication that a piece of dynamic content is intended be rendered in a dynamic video and determining at least one dynamic variable associated with the piece of dynamic content. The method further comprises detecting a selection of a selected dynamic variable of the at least one dynamic variable and determining a dynamic interaction associated with the selected dynamic variable. The method further comprises generating the dynamic video to comprise interactive dynamic content based on the piece of dynamic content and a static video, and labeling, in metadata of a dynamic video using a backwards-compatible video format, that the dynamic interaction is configured to be performed on the interactive dynamic content to modify the selected dynamic variable during playback of the dynamic video.
H04N 21/472 - End-user interface for requesting content, additional data or servicesEnd-user interface for interacting with content, e.g. for content reservation or setting reminders, for requesting event notification or for manipulating displayed content
H04N 21/431 - Generation of visual interfacesContent or additional data rendering
Techniques are provided for improving chat response generation using knowledge graph-based data retrieval. A chat system receives a user message and identifies relevant entities by performing a web search. A first generative language model receives a prompt containing the chat history, identified entities, and a knowledge graph schema defining entity types and relationships. The first model generates a structured query targeting specific entity attributes in the knowledge graph. After executing the query to retrieve targeted entity data, a second generative language model receives the retrieved data and user message to generate a contextually relevant response. The system enables precise control over grounding data by using the knowledge graph schema to specify exactly which entity attributes to retrieve, avoiding excessive or irrelevant information while maintaining comprehensive responses. This approach improves upon conventional database solutions by allowing flexible, relationship-aware queries that retrieve diverse yet focused entity information based on conversational context.
A technique is described for training a student model based on a larger teacher model. The training includes generating a loss measure having a contrastive combination of two parts. The first part is based on a forward measure of divergence between teacher-generated and student-generated probability distributions, which, in turn, are based on teacher-generated samples. The second part is based on a reverse measure of divergence between student-generated and teacher-generated probability distributions, which, in turn, are based on student-generated samples. The technique then updates parameters of the student model based on the loss. In some implementations, the first part of the loss is generated using forward Kullback-Leibler (KL) divergence, and the second part of the loss is generated using reverse KL divergence. The technique also involves dynamically updating hyper-parameters during training.
A method is presented for mixed-precision matrix multiplication. The method comprises decomposing an n-bit weight matrix into a series of n one-bit matrices. An m-bit activation matrix is received, where m>n. A lookup table of partial matrix multiplication results is generated based on permutations of groups of the m-bit activation matrix and the n one-bit matrices. An input weight matrix is received, the input weight matrix having a depth of g bits. For each bit of the input weight matrix, partial matrix multiplication results are retrieved from the lookup table using g-bit groups of the input weight matrix as indices. The retrieved partial matrix multiplication results are aggregated into a final mixed-precision matrix multiplication result.
The described technology provides a method including receiving a request from an agent for accessing a cogran and for allocating an agent ID to one of a plurality of SFT entries in a snoop filter (SFT), performing a tag lookup function for a tag of the cogran in the SFT to find a matched SFT entry, wherein the matched SFT entry is tracking the tag of the cogran, determining the number n of agents being tracked by the matched SFT entry, and in response to determining that the number n of agents being tracked by the matched entry is above a threshold, storing a DVT index in the tracking_info field of the matched SFT entry, wherein the DVT index selects a DVT entry in a disaggregated vector table (DVT), wherein the selected DVT entry is configured to hold a tracking vector for tracking the agents that have cached the cogran for the matched SFT entry.
Methods, systems, and computer program products are provided that implement vector-related requests using query operators. For example, a system includes a database, a parser, a converter, an optimizer, and an execution engine. The database is associated with query operators that are non-vector-specific. The database stores a table with vector embeddings. The parser is configured to parse a request indicating a vector operation associated with the vector embeddings. The converter is configured to convert the vector operation into a logical operator tree comprising a representation of the request as a logical flow of the query operators, enabling the vector operation without vector-specific executable code or operators. The optimizer is configured to convert the logical operator tree into an executable plan. The execution engine is configured to execute the executable plan against the table with vector embeddings.
The present disclosure relates to methods and systems that generate a confidence score for the generated large language model (LLM) output. The methods and systems use the text of the input provided to the LLM and the text from the generated LLM output to produce a feature vector that encodes a readability of the text from the input and the text of the LLM output. The feature vector is used to determine a corresponding confidence score for the generated LLM output. The confidence score is used to evaluate a quality of the generated LLM output.
Systems and methods are disclosed herein for compressing a prompt. In an example system, an importance score listing is obtained that includes a score indicative of an importance of a plurality of dataset keywords. From the importance score listing, a keyword importance score is identified for a plurality of keywords in a current text fragment, such as a text fragment to be compressed. A set of placeholders in an abstract prompt template is populated based on the current text fragment. The current text fragment is compressed based on the importance of the plurality of keywords in the current text fragment to generate a compressed text fragment. In an example, the compressed text fragment is included in the prompt for transmission to a computing entity, such as a large language model of a generative question-answering system.
The present disclosure includes an image restoration system that efficiently and accurately produces high-quality images captured under low-light and/or low-quality environmental conditions. To illustrate, when a user is in a low-lit environment and participating in a video stream, the image restoration system enhances the quality of the image by dynamically re-lighting the user's face. Moreover, it significantly enhances the image quality to the extent that other users viewing the video stream are unaware of the poor environmental conditions of the user. In addition, the image restoration system creates and utilizes an image restoration machine-learning model to improve the quality of low-quality images by re-lighting and restoring them in real time. Various implementations combine an autoencoder model with a distortion classifier model to create the image restoration machine-learning model.
A code adaptation mechanism automatically integrates the variable names of a pasted source code snippet into variable names defined in a pre-existing partial source code program. The variable names from the pasted source code snippet are replaced with anonymized values. A deep learning model predicts the most likely variable name from the pre-existing partial source code program to replace each anonymized value. The deep learning model is trained on numerous variable usage patterns from various source code programs to learn to predict the most likely mapping of an undefined variable name from the pasted source code snippet to a variable name in the pre-existing partial source code program thereby generating a syntactically and semantically correct program.
A computing system includes a processor and memory storing instructions that, when executed by the processor, cause the processor to perform several acts. The acts include providing a prompt to a generative language model, where the generative language model generates output based upon the prompt, identifies text in the output that is to be associated with a supplemental content item, and assigning a hyperlink to the text in the output. Upon the hyperlink being selected or hovered over, the supplemental content item is displayed.
In an example embodiment, a generator model such as a large language model (LLM) is leveraged to generate embeddings for both pieces of content and users. The embeddings map the pieces of content and the users into the same latent n-dimensional space. The embeddings are then fine-tuned using a two-tower deep neural network, with one of the towers representing users and the other tower representing content. The two-tower deep neural network is trained to optimize the embeddings over some shared goal, such as user engagement with content, and uses information such as user interactions with content in that process. A clustering technique, such as K-nearest neighbor (kNN) can then be used to identify a grouping of top user/content pairs based on similarity between users and content, as reflected in the embeddings. For a given piece of content, therefore, the top users from that cluster can then be recommended as an audience for the content.
Data-at-rest protection for virtual machines includes operating a data protection component within a first privilege context of a guest partition, and operating a guest operating system (OS) within a second privilege context of the guest partition. The data protection component participates in data input/output operations of the guest OS. Based on a data output operation of the guest OS, the data protection component applies a first data protection operation to first data associated with the data output operation; and initiates storage of a first result of the first data protection operation to a data storage device. Based a data input operation of the guest OS, the data protection component applies a second data protection operation to second data associated with the data input operation; and, based on applying the second data protection operation to the second data, communicates an outcome of the data input operation to the guest OS.
Technology is disclosed for improving read accuracy and storage capacity in optical data storage systems. The technology includes writing, with an optical write system, voxels into a voxel lattice having a voxel lattice constant (VLC) of an optical storage medium and reading, with an optical read system, the voxels from the optical storage medium. The optical read system includes optical components that have a point spread function (PSF). The VLC matches a shadow of the PSF either by virtue of the write system setting the VLC to match the shadow of the PSF of the read system or the read system adjusting the optical components to match the shadow of the PSF with the VLC. The system may be tuned by selecting which shadow of the PSF is used for matching with the VLC, the shape of the voxel lattice, the number of bits per voxel, or a combination.
G11B 7/00 - Recording or reproducing by optical means, e.g. recording using a thermal beam of optical radiation, reproducing using an optical beam at lower powerRecord carriers therefor
G11B 3/06 - Determining or indicating position of head
A data processing system implements receiving, from an application of a client device of a user, a request for content to be generated by a language model. The request includes a natural language prompt describing the content to be generated and an identifier of the user. The system further implements executing a query on one or more sample content sources to obtain sample content items authored at least in part by the user, constructing a prompt for the language model based on the natural language prompt and the sample content items, the prompt instructs the language model to mimic the writing style of the user based on the one or more sample content items, providing the prompt as an input to the language model to obtain the content, providing the content to the application of the client device, and causing the application to present the content in the application.
Aspects of the disclosure include methods and systems for modeling negative user experiences with notifications. A method includes collecting notification engagement patterns for a plurality of recipients. The notification engagement patterns each include one or more recipient actions in a sequence ending with a disinterest action. Each notification engagement pattern is assigned to a disinterest class of a plurality of predetermined disinterest classes according to the disinterest action for the respective notification engagement pattern and dynamically labeled training data is generated from the notification engagement patterns. Positive labels are assigned to actions within a respective notification engagement pattern according to the disinterest class of the notification engagement pattern. A model is trained using the dynamically labeled training data to generate disinterest predictions for candidate notifications to be delivered to recipients.
True random number generators including ring oscillator circuits leveraging frequency and phase collapse events are described. An example ring oscillator circuit (ROC) includes a first oscillating loop comprising a first set of inverters, where the first oscillating loop is to oscillate at a first frequency and phase. The ROC further includes a second oscillating loop comprising a second set of inverters, which differs from the first set of inverters either in terms of a type of inverters or a number of inverters. The second oscillating loop is to oscillate at a second frequency and phase, different from the first frequency and phase. The ROC further includes a third oscillating loop to, as a result of mode switching from a first mode of operation into a second mode of operation, oscillate at a third frequency and phase, different from the first frequency and phase and the second frequency and phase.
A computerized method classifies vegetation in geographic regions. Target coordinates that describe a geographic region are obtained and vegetation data associated with the geographic region is obtained. In some examples, the vegetation data includes satellite imagery data of the geographic region. Vegetation index data points are generated using the obtained vegetation data (e.g., Normalized Difference Vegetation Index (NDVI) data points). A signature curve associated with the geographic region is plotted using the generated vegetation index data points and the vegetation in the geographic region is classified using the plotted signature curve. Then, a vegetation management action is caused to be performed in association with the geographic region based on the classified vegetation. The method enables the identification of crops in the geographic location and/or monitoring of forest growth or decline in the geographic location.
Systems and methods to provide vulnerability detection and management in a cloud computing system according to examples. More specifically, a vulnerability detection system receives access logs of a package repository and records information that links packages downloaded from the package repository to the computing assets that downloaded the packages. The vulnerability detection system evaluates an inventory of the package repository against a report of identified vulnerabilities (e.g., Common Vulnerabilities and Exposures (CVEs)). When a vulnerable package is identified in the inventory, the vulnerability detection system removes the vulnerable package from the package repository. The vulnerability detection system further determines affected assets and contact information of corresponding users and provides notifications to the users. In some examples, a mitigation or remediation is determined and provided to the users.
G06F 21/57 - Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
G06F 21/55 - Detecting local intrusion or implementing counter-measures
73.
ROLE-BASED APPROVAL OF A RESOURCE LIFECYCLE EVENT WITH REGARD TO A CLOUD-BASED RESOURCE
Techniques are described herein that are capable of performing role-based approval of a resource lifecycle event with regard to a cloud-based resource. A request to perform a resource lifecycle event with regard to a cloud-based resource is received. A designated approver is identified. The designated approver has authority to grant approval to perform the resource lifecycle event with regard to the cloud-based resource as a result of the designated approver having an approver role, which is authorized to grant the approval. An explicit approval is obtained from the designated approver. The explicit approval grants approval to perform the resource lifecycle event with regard to the cloud-based resource. As a result of an approval criterion being satisfied, performance of the resource lifecycle event with regard to the cloud-based resource is triggered. The approval criterion includes obtainment of the explicit approval.
According to implementations of the subject matter described herein, a solution for generation of three-dimensional avatar is provided. According to this solution, a trained diffusion model is obtained, which is trained based on a sample three-dimensional avatar of a sample object. A target feature representation is generated from a predetermined input using the diffusion model. The target feature representation comprises a set of target feature maps corresponding to a tri-plane, respectively, to characterize feature information of a target object in a three-dimensional space. A three-dimensional avatar of the target object is generated based on the target feature representation. In this way, high-quality three-dimensional avatars can be generated in an efficient way, with reduced memory and computing costs.
An example may, at a first device, input an entity embedding for an entity and an item embedding for a plurality of items to a scoring function. The entity embedding and the item embedding are pre-computed using a first language model. At the first device, a retrieval score for the entity and a first item of the plurality of items is computed. The retrieval score is used to identify an item subset of the plurality of items. A ranking prompt is input to a second language model. The second language model and the first language model have a common parameter value. The second language model generates a ranking score for the entity and a second item in response to the ranking prompt. An online system uses the ranking score to include or exclude the second item from a presentation of digital content to the entity via a device.
Examples are disclosed that relate to determining phase orders for phase data in a time-of-flight camera sensor, for use in unwrapping the phase data. One example provides a computing system comprising a depth sensor comprising a plurality of pixels, an illumination source, and a storage machine holding instructions executable by a logic machine to control the illumination source to output amplitude-modulated light at two or more modulation frequencies. The instructions are further executable to, for each pixel, make two or more phase measurements corresponding to different modulation frequencies, based at least on the two or more phase measurements, determine a series of phase order sets, determine a most likely phase order set by comparing the series of phase order sets to a line representing an evolution of phase with distance, and based on the most likely phase order set, determine a distance value associated with the pixel.
Aspects of the present disclosure include devices and methods for selecting relevant content in an enhanced view mode. In an example, a computer device may be configured to receive digital content to be displayed by a web-based application according to a first view mode. The computer device may be configured to determine an enhanced view mode is enabled, wherein the enhanced view mode is different from the first view mode. The computer device may be configured to identify one or more first portions of primary content of the digital content to be displayed according to the enhanced view mode based on user information. The computer device may be configured to cause display of the one or more first portions of the primary content in the web-based application according to the enhanced view mode, and one or more second portions of the primary content according to a second view mode different from the enhanced view mode.
A two-phase immersion cooling system may include a two-phase immersion cooling container defining a volume. The two-phase immersion cooling system may include an immersion fluid housed in the two-phase immersion cooling container filling at least a portion of the volume. The two-phase immersion cooling system may include at least one computing system that is at least partially submerged in the immersion fluid, such that the at least one computing system includes a printed circuit board, an integrated circuit, a ball grid array that electrically and mechanically couples the integrated circuit to the printed circuit board, and a sealant between the printed circuit board and the integrated circuit that substantially prevents fluid flow of the immersion fluid across at least a portion of a plurality of solder balls that are included in the ball grid array when the integrated circuit is operating at a temperature that boils the immersion fluid.
According to one implementation, a disclosed method includes transmitting one or more commands to instruct a heat rejection unit (HRU) of the cooling system to transmit a data signal along the liquid coolant loop by controllably altering pump power between a baseline power level and a second power level that imparts a predicable change in a flow characteristic measured at a subset of the devices coupled to the coolant loop. The method further provides for sampling values of a flow characteristic at one or more of the devices both before and after altering the pump power of the first HRU and transmitting a telemetry' stream including the first values to a processing system that confirms transmission of the data signal in response to detecting a pattern in the telemetry stream that matches an expected response pattern.
The disclosed techniques provide a dynamic voltage and frequency scaling (DVFS) module and a domain residing in a core. The DVFS module includes a Voltage and Memory Assist Table (VMAT), including firmware-programmable registers that store target voltages that correspond to performance states and firmware-programmable registers that store memory assist values that correspond to the performance states. The domain includes memories. The DVFS module is configured to, during a current performance state: provide a core voltage to the domain, such that the core voltage is provided based on a target voltage among the target voltages that corresponds to the current performance state, and provide memory assist signaling to the memories, such that the memory assist signaling is provided based on the memory assist values that correspond to the current performance state. The DVFS module uses an efficient crawl to transition memory assist values from one performance state to another performance state.
Methods and systems for embedding a digital watermark in a stylus pen ink. A method (300) for embedding a digital watermark in a stylus pen ink by a digitizer comprises detecting (310) a touch from a stylus pen on a display of the digitizer, determining (312) a user identification, wherein the user identification is linked to the digital watermark, and based at least on the touch from the stylus pen on the display, displaying (322) the stylus pen ink on the display, the stylus pen ink embedding the digital watermark as a pattern of pixels.
G06F 21/16 - Program or content traceability, e.g. by watermarking
G06F 3/0354 - Pointing devices displaced or positioned by the userAccessories therefor with detection of 2D relative movements between the device, or an operating part thereof, and a plane or surface, e.g. 2D mice, trackballs, pens or pucks
G06F 21/32 - User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
G06F 21/64 - Protecting data integrity, e.g. using checksums, certificates or signatures
Privilege overreach is a security threat that occurs when users are granted excessive access to electronic applications without considering potential security risks. This threat can arise for several reasons such as lack of access control policies and tools, user role complexity, legacy systems with inherited user privileges, and gradual accumulation of unnecessary privileges. For example, when users change roles or take on additional responsibilities at a company, they may accumulate access rights to various electronic applications without shedding access rights for electronic applications associated with their previous roles, leading to a gradual increase in privileges that can result in privilege overreach. The present technology limits access privileges to provide secure group-based access to electronic applications by discovering active users, classifying groups, and prioritizing them. Applications from a directory and application sign-in data are read and analyzed, along with group membership data, to determine optimal groups allowed to access an application.
True random number generators including ring oscillator circuits leveraging frequency and phase collapse events are described. An example ring oscillator circuit (ROC) includes a first oscillating loop comprising a first set of inverters, where the first oscillating loop is to oscillate at a first frequency and phase. The ROC further includes a second oscillating loop comprising a second set of inverters, which differs from the first set of inverters either in terms of a type of inverters or a number of inverters. The second oscillating loop is to oscillate at a second frequency and phase, different from the first frequency and phase. The ROC further includes a third oscillating loop to, as a result of mode switching from a first mode of operation into a second mode of operation, oscillate at a third frequency and phase, different from the first frequency and phase and the second frequency and phase.
A system accesses metrics of features of workflows indicating an anomaly, the features represented as nodes in a graph and causal relationships as links, wherein a root node corresponds to the anomaly. The system, for each feature in each subgraph, computes an adjacent contribution score corresponding to each link for which the feature is a causing node, based on the metrics. The system determines a nonadjacent contribution score for each feature of the nodes in the graph to the anomaly by: assigning the adjacent contribution scores of features in a root subgraph of the multiple subgraphs that includes the root node to the corresponding features in the graph, for an adjacent subgraph of the multiple subgraphs that includes a shared feature that is also in the root subgraph, modifying the adjacent contribution scores of unique features of the adjacent subgraph based on the nonadjacent contribution score of the shared feature.
A computing device including a system-on-a-chip (SoC). The SoC includes a plurality of logic circuit blocks, including synchronization blocks and hardware accelerator blocks. A synchronization block is configured to receive a wait request from a first hardware accelerator block. The wait request includes one or more semaphores and one or more wait threshold values. The synchronization block is configured to store the wait request. The synchronization block is configured to receive, from a signal source block, a signal request that indicates a semaphore included among the one or more semaphores in the wait request. In response to receiving the signal request, the synchronization block is configured to update the semaphore. The synchronization block is configured to determine that the updated value of the semaphore has reached the wait threshold value and to transmit a wait completion response to at least the first hardware accelerator block.
Examples solutions for evaluating network connectivity between computing resources within a computing environment include: generating a network packet for a proxy ping at a first host server that provides a first virtual machine (VM), the network packet being originally generated by the first host server and not generated by the first VM, the network packet being created to include a proxy ping flag within a header of the network packet; capturing a sending timestamp; receiving a response network packet generated by a second host server in response to interception of the network packet without the network packet being sent through to the second VM, the network packet having been identified for interception by the second host server based on identifying the proxy ping flag within the header of the network packet; capturing a response timestamp; and computing a latency value based on a difference between the sending and response timestamps.
Examples solutions for evaluating network connectivity between computing resources within a computing environment include: generating a network packet for a proxy ping at a first host server that provides a first virtual machine (VM), the network packet being originally generated by the first host server and not generated by the first VM, the network packet being created to include a proxy ping flag within a header of the network packet; capturing a sending timestamp; receiving a response network packet generated by a second host server in response to interception of the network packet without the network packet being sent through to the second VM, the network packet having been identified for interception by the second host server based on identifying the proxy ping flag within the header of the network packet; capturing a response timestamp; and computing a latency value based on a difference between the sending and response timestamps.
A computer-implemented method is presented for solving a many-body quantum system problem. The method comprises, at a classical computing device, receiving a many-body quantum system problem. A portfolio of approximate candidate methods for solving the many-body quantum system problem is received. Approximate solutions are generated to the many-body quantum system problem using the portfolio of approximate candidate methods. A quantum computing device is provided with the many-body quantum system problem and the approximate solutions. A quantum energy for each of the portfolio of approximate candidate methods is received from the quantum computing device. One or more approximate candidate methods are selected based on their respective quantum energy. For each of the selected approximate candidate methods, the many-body quantum system problem is solved using the one or more selected approximate candidate methods. Properties of a quantum state for the many-body quantum system are output.
A method comprising: receiving application data pertaining to a plurality of applications each having a respective known threat status; computing, using the application data, a feature vector for each application, the feature vector comprising values of features corresponding to a plurality of application-related entities for the plurality of applications; determining at least one neighbouring application for each application in the plurality of applications; training a threat detection model based on the known threat status of each application in the plurality of applications and the at least one feature vector of its at least one neighbouring application; computing, from the trained threat detection model, feature weights representative of relative importance of the features in the at least one feature vector of the at least one neighbouring application; determining a reputation score for a value of a feature using the feature weight of the feature and a reputation score of the application.
The description relates to hinged devices, such as hinged computing devices. One example can include a first portion associated with a first axial timing surface and a second portion associated with a second axial timing surface. The example can also include a clutch stack spanning between the first portion and the second portion and a timing shuttle configured to engage the first and second axial timing surfaces to synchronize rotation of the first and second portions through a range of rotation. The example can include orientation dependent cams that control compression of the clutch stack as the first and second portions rotate through the range of rotation.
The description relates to hinged devices, such as hinged computing devices. One example can include a first portion including a display and a second portion including an input device. This example can also include a hinge assembly rotationally securing the first and second portions through a range of angular orientations and a sensor positioned relative to the hinge assembly and configured to sense relative linear positions of the hinge assembly that correspond to angular orientations of the first and second portions.
H05K 5/02 - Casings, cabinets or drawers for electric apparatus Details
G01B 21/22 - Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for measuring angles or tapersMeasuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant for testing the alignment of axes
G06F 1/16 - Constructional details or arrangements
92.
TRAINING LANGUAGE MODELS FOR RETRIEVAL AND RANKING
An example may train a cross encoder embedding model using a ranking instruction, a combined input, a pseudo label, and a combined loss. The combined loss includes a ranking loss and a first retrieval loss. A first entity embedding of an entity and a first item embedding of an item may be obtained from the trained cross encoder embedding model. A first input including the first entity embedding obtained from the trained cross encoder embedding model, a second input including the first item embedding obtained from the trained cross encoder embedding model, and a second retrieval loss, may be used to train a dual encoder retrieval model to produce a trained dual encoder retrieval model. A system may use output of the trained dual encoder retrieval model to include or exclude items from a presentation of digital content items to the entity via a device.
The description relates to connectors for connecting components of a system. One example can include a rectangular housing that has connectors positioned along at least one side. A single axis deployment mechanism can be configured to constrain movement of the connectors to a single axis. Upon receiving input from an angle that is perpendicular or oblique to the axis, the single axis deployment mechanism can be configured to cause the connectors to move sequentially from either of a stored position or a deployed position to the other of the stored position or the deployed position.
This document relates to the compressing and/or decompressing of spatial data. A block of spatial data can be compressed by iteratively performing compression iterations on portions of the block and splitting the portions into further portions until compression is completed. The compressed data can include first encoded values indicating whether matches were obtained for comparisons to test values during the compression iterations. The compressed data can also include second encoded values reflecting results of one or more modifications performed on the test values during the compression iterations.
A computing system is disclosed that includes a processor and memory. The memory stores instructions that, when executed by the processor, cause the processor to perform several acts. The acts include generating a prompt that is to be input to a generative language model. The prompt includes conversational input set forth by a user. The acts further comprise providing the prompt as input to the generative language model, and receiving conversational output from the generative language model, where the generative language model generated the conversational output based upon the prompt. Additionally, the acts comprise receiving an indication that the user has performed an interface mode change action and updating a search engine results page (SERP) to provide information related to the conversational output generated by the generative language model. The acts further comprise presenting the updated SERP to the user on a client computing device.
The techniques describe effective detection of latency-related issues for a cloud service operating in a distributed computing environment. To detect the latency-related issues, a system first determines baseline latency behavior at the tenant level (e.g., on a tenant-by-tenant basis) and compares a tenant's current latency behavior to the baseline latency behavior. If the comparison yields that the current latency behavior for the tenant is following the baseline latency behavior, the tenant is deemed healthy. However, if the comparison yields that the current latency behavior for the tenant is not closely following the baseline latency behavior, the tenant is deemed unhealthy. Once the system has made binary health determinations for various tenants on a tenant-by-tenant basis, the system is configured to aggregate the unhealthy determinations across a group of tenants to determine whether the cloud service is experiencing latency-related issues.
An example monitors an event stream of a multi-agent application system. A first event of the event stream is determined to correspond to a state of a first task agent of the multi-agent application system. The state is established using historical data relating to a first task performed by the first task agent. The first event is routed to the first task agent. Output of a second task performed by the first task agent is received from the first task agent. The second task is performed by the first task agent in response to the first event.
G06Q 10/0637 - Strategic management or analysis, e.g. setting a goal or target of an organisationPlanning actions based on goalsAnalysis or evaluation of effectiveness of goals
A system includes a cooling unit (130) and a thermally conductive structure (132). The cooling unit (130) and thermally conductive structure (132) define a gap between the cooling unit (130) and the thermally conductive structure (132). The thermally conductive structure (132) includes an appendage (152) extending from the thermally conductive structure (132) across the gap to the cooling unit (130) for thermal conduction with the cooling unit (130). The gap is sized to receive a printed circuit board (128) and attached components.
An example maps input to a first task executable by a first task agent of a multi-agent application system. A search is formulated using the input and the first task. The search is executed on a second layer of a multi-layer memory accessible to the orchestrator agent. First cross-agent context data related to the input and the first task is extracted from results of the search. The first task and the first cross-agent context data are routed to the first task agent. Output is received from the first task agent. The output is generated via execution of the first task by the first task agent using the first cross-agent context data. A response to the input is provided. The response includes the output generated by the first task agent using the first cross-agent context data generated by the second task agent.
G06Q 50/00 - Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
100.
SECURE EXECUTION OF AN AI MODEL ON A NEURAL PROCESSING UNIT OF A CLIENT DEVICE
Techniques are described herein that are capable of securely executing an Al model on a neural processing unit (NPU) of a client device. The NPU runs the Al model. The NPU encrypts data, which includes an Al prompt, using a cryptographic key. The NPU provides the encrypted data to a cloud-based security service via a utility in an operating system that executes on the computing system. The NPU receives a response indicator from the cloud-based security service via the utility. The response indicator results from an analysis of a decrypted representation of the encrypted data. The response indicator indicates that an alternative response is to be provided in lieu of an Al response, which is received from the Al model as a result of the Al prompt, as a response to the Al prompt. The NPU provides the alternative response as the response to the Al prompt.
patents
