Google Patente

Devices, systems, and methods for isolating an insect pupa from an aqueous solution including a plurality of insect pupae. An isolation device may include a primary channel through which the aqueous solution and insect pupae travel, an outlet, and a secondary channel that intersects the primary channel. The secondary channel may transport a fluid to add it to the primary channel between an inlet and the outlet. When the sealable outlet is closed, an insect pupa located downstream of the intersection of the primary and secondary channels remains stationary while any insect pupae upstream of the intersection are pushed towards the inlet. The systems and methods describe herein may also provide for identifying and sorting insect pupae.

Estimating throughput for an input program executing on a processing architecture by calculating a plurality of cumulative distribution functions (CDFs) for each parameter of a plurality of parameters for describing processing architectures, the CDFs of the plurality of CDFs corresponding to respective ones of values for the parameter, and each CDF of the plurality of CDFs specifying a cumulative distribution of throughput calculations associated with the corresponding value and the input program, and the pluralities of CDFs generated for the parameters making up a set of pluralities of CDFs; and using the set of pluralities of CDFs and a machine learning model to estimate the throughput for the input program executing on the processing architecture based on a set of parameter values specifying the processing architecture.

Aspects of the disclosure are directed to enhancing RAG through a structured schema for RAG that captures a wider range of RAG techniques to serve as a foundation for improving performance and an optimization framework for efficiently serving the wider range of RAG techniques based on significant performance variability across RAG workloads. The optimization framework may achieve up to a 2x increase or more in query per second (QPS) per chip compared to alternative RAG frameworks built on LLM-system extensions.

A method in a radio access network, RAN, communicating with a user equipment, UE, the method comprising: sending to the UE a conditional configuration and a corresponding condition for a conditional PSCell addition or change procedure, wherein the corresponding condition is to be satisfied before the UE can communicate via a candidate cell of the RAN using the conditional configuration; after sending the conditional configuration and the corresponding condition, and before the corresponding condition is satisfied, receiving UE information from the UE, the UE information including a preferred configuration, in-device coexistence, IDC, assistance information, sidelink information, or multicast and/or broadcast service, MBS, interest information; and after the corresponding condition is satisfied, again receiving the UE information from the UE.

A computing device includes a temperature sensor for generating skin temperature data of a user of the computing device, a skin autofluorescence (SAF) sensor for generating SAF data of the user, and a controller configured to perform a plurality of operations. The plurality of operations includes receiving the skin temperature data and the SAF data over time, determining a correlation between the skin temperature data and the SAF data over time, applying a temperature correction factor to the SAF data to correct the SAF data for skin temperature of the user based on the correlation, determining at least one health parameter of the user using the corrected SAF data, and displaying, via a display of the computing device, the at least one health parameter to the user.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

A method for on-device AI during a virtual meeting includes receiving, at a first client device of a first participant of one or more participants of a virtual meeting, a request to perform a task pertaining to the virtual meeting from a second client device of a second participant of the plurality of participants; obtaining, at the first client device, first text data based on first audio data associated with a first audio stream produced by the second client device; performing, using an AI model of the first client device and using the first text data as input to the AI model, the task pertaining to the virtual meeting; and providing, by the first client device, information associated with the performed task pertaining to the virtual meeting to the second client device.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speaker diarization are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The actions further include determining that the audio data includes an utterance of a predefined hotword spoken by a first speaker. The actions further include identifying a first portion of the audio data that includes speech from the first speaker. The actions further include identifying a second portion of the audio data that includes speech from a second, different speaker. The actions further include transmitting the first portion of the audio data that includes speech from the first speaker and suppressing transmission of the second portion of the audio data that includes speech from the second, different speaker.

Provided are methods, systems, devices, and tangible non-transitory computer readable media for offline mapping and navigation. The disclosed technology determines a current region associated with a current location of a navigation device. Predicted travel destinations from the current region are determined based on aggregated travel data. The aggregated travel data can include previous travel destinations from within the current region. Based on the current region and the predicted travel destinations, additional regions associated with the current region are determined. Further, connectivity information is determined based on network data. The connectivity information is associated with network connectivity within the additional regions. The additional regions is prioritized based on the connectivity information. Furthermore, portions of offline map data associated with the respective additional regions are sent to the navigation device in an order partly based on the priority of the additional regions.

A wearable computing device has an outer covering, a housing, an electronic display screen arranged within the housing and viewable through the outer covering, a rotary crown positioned on a side of the electronic display screen, and at least one controller communicatively coupled to the rotary crown. The rotary crown is configured to receive a rotational input. The controller(s) is configured to apply a gain factor to the rotational input to generate a digital output for the electronic display screen. The gain factor is proportional to an angular velocity of the rotational input.

Implementations described herein relate to determining whether to modify segment(s) of responsive content, that is generated using a generative model (GM), and based on whether the segment(s) include opt-out content. The opt-out content can be associated with a given user or a given entity that has opted-out of: the GM being trained on data, that is associated with the given user or the given entity, since a last training cycle for the GM, or the GM being able to use the data, that is associated with the given user or the given entity, in generating the responsive content. If processor(s) of a system determine that a corresponding segment of the responsive content matches a corresponding segment of the opt-out content, then the processor(s) can modify the corresponding segment of the responsive content to generate modified responsive content, and cause the modified responsive content to be rendered at the client device.

Implementations relate to receiving natural language (NL) input associated with a client device; and generating response(s) to the NL input. Generating the response(s) includes: determining, based on the NL input, a base prompt; processing, using a first generative model (GM), first GM input to generate corresponding first GM output, the first GM input including the base prompt; determining, based on the first GM output, a plurality of extended prompts; obtaining, based on filtering the plurality of extended prompts, a subset of extended prompts; for each extended prompt of the subset: processing, using the first GM or a second GM, second GM input to generate corresponding second GM output, the second GM input including the respective extended prompt, and determining, based on the second GM output, a respective candidate response corresponding to the respective extended prompt; and obtaining, based on filtering the candidate response(s), the response(s) to the NL input.

Aspects of the disclosure are directed to a system for incorporating a CVBG (Chirped Volume Bragg Grating) into a prism to form a P-CVBG (Prism-Chirped Volume Bragg Grating) optical element for channel dispersion compensation. A prism region is configured to receive optical signals with multiple wavelengths and provide the wavelengths to the CVBG region by spatially separating the optical signals into various wavelengths. The CVBG region is configured to compensate for dispersion of the multiple wavelengths by reflecting the various wavelengths at different depths within the gratings of the CVBG region.

A method (400) for remote attestation includes establishing, using a cryptographic protocol (20), a communication session (22) between a first computing device (10a) and a second computing device (10b). The communication session includes communications encrypted by an ephemeral session key (24). The method includes receiving, at the first communication device via the communication session, from the second computing device, an attestation request (172) requesting the first computing device to provide an attestation report (162). The method includes generating, by the first computing device, the attestation report based on the ephemeral session key and sending, using the communication session, the attestation report to the second computing device.

One aspect is a method for recovering a rename map table for an out-of-order processor after a flush operation, the method includes executing, by an out-of-order processor, instructions including a flush operation resulting from a branch misprediction of a branch instruction. The out-of-order processor includes a reorder buffer tracking a program order for the instructions and a rename map table mapping multiple logical registers to multiple physical registers. The method also includes, in response to executing the flush operation, grouping the instructions, ordered from oldest to youngest according to the program order tracked in the reorder buffer, into multiple of segments, comparing the multiple segments to identify segments that contain a youngest superset of instructions that wrote to the multiple logical registers between a head of the reorder buffer and the branch instruction, and performing a recovery walk of the identified segments to recover the rename map table.

Implementations relate to depersonalizing and/or anonymizing generative model inputs and/or outputs to remove sensitive and/or private information, such as PII (personally identifiable information). In various implementations, a generative model input prompt may be retrieved from local memory of the edge computing device and further processed using one or more generative models to generate generative model output. The generative model input prompt may be submitted to an edge-based anonymization process including assembling, as an anonymization prompt and processing the anonymization prompt using one or more on-device generative models of the edge computing device to generate an anonymized version of the portion of the generative model input prompt that is stripped of PII. The anonymized version may further be inspected of any remaining PII and the fully anonymized version may be surfaced without concern of leaking PII.

A customer gains access to cloud networking resources through a customer device, such as a router providing a point of presence (POP) to the network. Due to limitations of remote equipment such as routers, complete routing information for data packets may not be available from the origin due to the dynamic nature of the network and security requirements. Proxy machines are placed at the network edge that serve as a proxy for routing network data packets to the virtual network Placing proxy machines at the network edge requires space, power and maintenance of the programmable processing machines serving as proxies. Techniques for bypassing proxy machines to reduce network congestion and latency are described. Data packets are selected to bypass proxies when additional routing information is not needed for the packet. Proxy machines may be implemented as virtual machines within a data center hosting the virtual network.

Zusammenfassung wird geladen …

Techniques are described herein for improving performance of machine learning model(s) and thresholds utilized in determining whether automated assistant function(s) are to be initiated. A method includes: receiving, via one or more microphones of a client device, audio data that captures a spoken utterance of a user; processing the audio data using a machine learning model to generate a predicted output that indicates a probability of one or more hotwords being present in the audio data; determining that the predicted output satisfies a secondary threshold that is less indicative of the one or more hotwords being present in the audio data than is a primary threshold; in response to determining that the predicted output satisfies the secondary threshold, prompting the user to indicate whether or not the spoken utterance includes a hotword; receiving, from the user, a response to the prompting; and adjusting the primary threshold based on the response.

Implementations set forth herein relate to an automated assistant that can initialize execution of an assistant command associated with an interpretation that is predicted to be responsive to a user input, while simultaneously providing suggestions for alternative assistant command(s) associated with alternative interpretation(s) that is/are also predicted to be responsive to the user input. The alternative assistant command(s) that are suggested can be selectable such that, when selected, the automated assistant can pivot from executing the assistant command to initializing execution of the selected alternative assistant command(s). Further, the alternative assistant command(s) that are suggested can be partially fulfilled prior to any user selection thereof. Accordingly, implementations set forth herein can enable the automated assistant to quickly and efficiently pivot between assistant commands that are predicted to be responsive to the user input.

A method includes identifying, by a processor, a media stream comprising a sequence of encoded video frames and estimating, in the sequence of video frames, a share of video frames that are fragmented into respective pluralities of network packets, wherein each plurality of network packets comprises a network packet having a size below a predefined packet threshold size. Whether the share exceeds a predefined threshold value is determined and responsive to determining that the share exceeds a predefined threshold value, a target bitrate of the video stream is reduced.

A wearable computing device includes one or more sensors for detecting low motion of a user of the wearable computing device and a processor communicatively coupled with the sensor(s). The processor is configured to perform a plurality of operations, including, in response to detecting the low motion of the user, monitoring an oxygen saturation level of the user for a drop from above an upper threshold to consistently below a lower threshold, upon detecting the oxygen saturation level dropping from above the upper threshold to consistently below the lower threshold, determining an oxygen saturation drop above a certain magnitude, in response to the oxygen saturation drop being above the certain magnitude, requesting an input from the user via the wearable computing device, and implementing, via the wearable computing device, an emergency response if the input is not received.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training machine learning models using data of multiple entities without compromising the precise joining/alignment of the data. In one aspect, a method includes obtaining two or more datasets, wherein each dataset is received from a different entity that maintains the dataset. The two or more datasets are joined using a set of keys. A loss function for training a machine learning model is generated based on inputs comprising the joined two or more datasets, the machine learning model including multiple model parameters. Noise is injected into one or more derivatives computed from the loss function, the inputs, or both. The model parameters are updated using the noised one or more derivatives.

Aspects of the disclosure are directed to an error detection approach performed in a processing cycle with improved accuracy and reduced latency. The approach involves determining whether the size of a received input value is greater than, equal to, or less than the size of an accumulator value and performing error detection accordingly using parallel lookup tables corresponding to portions of the input value. To avoid bit reversals in the error detection approach, results stored in the lookup tables may be stored in a bit reversed order. For input values that are less than the size of an accumulator value, the accumulator value is split into two portions to increase accuracy of the error detection result.

Zusammenfassung wird geladen …

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a diffusion neural network using a region-aware fine-tuning process. After training, the diffusion neural network can be used to generate an image conditioned on a conditioning input.

Implementations relate to determination and visual display of spoken menus for calls. In some implementations, a computer-implemented method includes receiving audio data output in a call between a call device and a device associated with a target entity. The audio data includes speech indicating one or more selection options for a user of the call device to navigate through a call menu provided by the target entity in the call. Text is determined by programmatically analyzing the audio data, the text representing the speech. The selection options are determined based on programmatically analyzing at least one of the text or the audio data. At least a portion of the text is displayed by the call device during the call, as one or more visual options that correspond to the selection options. The visual options are each selectable via user input to cause corresponding navigation through the call menu.

Implementations described herein relate to receiving user input directed to an automated assistant, processing the user input to determine whether data from a server and/or third-party application is needed to perform certain fulfillment of an assistant command included in the user input, and generating a prompt that requests a user consent to transmitting of a request to the server and/or the third-party application to obtain the data needed to perform the certain fulfillment. In implementations where the user consents, the data can be obtained and utilized to perform the certain fulfillment. In implementations where the user does not consent, client data can be generated locally at a client device and utilized to perform alternate fulfillment of the assistant command. In various implementations, the request transmitted to the server and/or third-party application can be modified based on ambient noise captured when the user input is received.

Methods, systems, and apparatuses, including computer programs encoded on computer storage media, for training a target generative neural network over a plurality of training iterations. At each iteration, a first data item is generated by processing a conditioning input using the target generative neural network. An improvement generative neural network then processes the first data item and the conditioning input to generate a second, preferred data item. A training example is generated that includes the first and second data items and indicates that the second data item is preferred over the first. The target generative neural network is then trained on this training example. By using this iterative process to dynamically generate preference data, the described techniques improve the performance of the generative neural network beyond the limitations of static, offline datasets without requiring computationally expensive reward models or external human annotation.

Generally disclosed herein is an approach to mitigating hardware degradation of server machines caused by frequent chip temperature fluctuations based on controlling the power consumption level, changes in xPU temperature of server machines, and the job start latency for the server machines altogether. According to some examples, a power and temperature optimization system may monitor xPU temperature fluctuations caused by inter-job fluctuations related to the xPU's deep idle state. The xPU's deep idle state may refer to a state where the xPU turns off or reduces the voltage of the xPU components to save power when a job or a unit of work assigned to the xPU stops. The xPU's deep idle state may continue until the next job or unit of work starts.

A method and system for controlling performance of a workload partitioned among a plurality of accelerator chips (301-304) of a multi-chip system (300). One or more processors (110) may receive performance speed data for each of the accelerator chips (610), obtain a model of the partitioned workload (620), determine a portion of the workload that is either overworked or underworked based on the model of the partitioned workload and the performance speed data for each of the plurality of accelerator chips (630), and adjust a performance speed of an accelerator chip that performs the portion of the partitioned workload that is either overworked or underworked (640).

A media application receives an initial image and a textual request to change the initial image, the initial image including a subject with a face. The media application generates, from the initial image, a preserving mask that corresponds to the face of the subject. The media application provides the textual request, the initial image, and the preserving mask as input to a diffusion model. The diffusion model outputs a denoised initial image based on the initial image; performs text conditioning of the textual request and forward diffusion to generate a noisy translated image that satisfies the textual request; and outputs based on the noisy translated image, extracted features, and self-attention maps, a denoised translated image. The media application blends the denoised initial image, the preserving mask, and the denoised translated image to form an output image, wherein the preserving mask prevents modification to the face from the initial image.

The technology is generally directed to error detection and correction in memory devices using dynamic random-access memory (DRAM) elements. Error correction codes (ECC) provide the ability to correct errors in memory elements. The ability to extend ECC capabilities to correct errors in defective elements is achievable when the location of the error(s) is known. Error correction metadata is calculated in the memory device when a write operation is performed. The metadata is stored in available space on the memory device. To locate uncorrectable errors, a pair of elements is selected and ECC is performed to recreate the data in the pair of elements. The metadata is recalculated on the recreated data and compared to the originally stored metadata. A match in metadata indicates the chosen pair contained the error. A mismatch indicates that the recreated data does not contain the error. Locating errors extends ECC correction capability.

The technology is directed to error detection and correction in a memory device containing dynamic random-access memory (DRAM) elements. The DRAM elements are configured to store data in a first number of elements, store error correction code (ECC) information in a second number of elements and storing an error correction code in a third number of elements. The amount of memory allocated to ECC is reduced without adversely affecting the ECC capability by storing an error correction code, such as a cyclic redundancy code (CRC) in the freed-up memory space. A location of an uncorrectable error is identified assuming one of the elements contains an error then recalculating the selected data using the ECC, recalculating the CRC and comparing it to the original CRC. A match indicates the faulty value has been found and corrected. A mismatch indicates that the selected element is not the location of the error.

Methods, systems, and apparatus, including computer-readable storage media for hardware-accelerated fine-grained sparse computation. The accelerator provides for improved performance for structured fine-grained sparse AI workloads, for example by accelerating sparse matrix multiplication required to execute or train AI models. Sparse data is compressed to remove zero-valued elements before being streamed into a matrix multiplication unit (MXU) of the accelerator. The accelerator stores a gains matrix, which can be the matrix for multiplying with the received input matrix. The accelerator uses an index array mapping locations of elements in the compressed matrix with locations of elements in the matrix's precompressed form, to generate a multiplier matrix from the gains matrix. Aspects of the disclosure also provide for accelerated gains matrix loading in a hardware accelerator or other type of processor. The accelerator can load the gains matrix more efficiently in a compressed form, and then un-compress the matrix once loaded.

Methods, systems, and apparatus, including computer-readable storage media and computer program products for determining thresholds for guaranteeing the availability of partitions of different quantities of computing devices and using those thresholds to more accurately provision computing resources to users of a computing platform. A partition (also called a "slice") of computing devices can be provisioned to a user of a computing platform. Slice-size thresholds can indicate how many computing devices on the platform can be provisioned as part of a single slice. Determining slice-size thresholds before provisioning computing resources reduces or eliminates the chance that requests to provision a slice of computing devices are approved when the platform is not able to fulfill those requests with the largest slice-size possible for a given request.

Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for increasing sparsity to improve neural network efficiency. In some implementations, a system stores parameter values of parameter matrices of one or more layers of a neural network. The parameter values of the parameter matrices include (i) weight values of the one or more layers of the neural network, and (ii) predictor values that have been trained to predict levels of importance of items processed by the neural network. The system generates an output, including: determining a value for each of multiple items using the predictor values, selecting a proper subset of the items based on the values in the vector based on a threshold, and generating output of the one or more layers limiting computation based on the selected proper subset.