Abstract: A method of determining a position for a virtual object is described. A location of a client device is determined, and, based on the determined location a set of map segments is retrieved. A virtual object is determined to be displayed on the client device. Relation vectors between the virtual object and each map segment of the retrieved set of map segments are obtained. Each relation vector is weighted based on object parameters of the virtual object. A position to display the virtual object is determined based on the weighted relation vectors. The virtual object is provided for display on the client device as the determined position.

Abstract: A method, apparatus, non-transitory computer readable medium, and system for document retrieval include obtaining a query and a document. A prompt generator generates a prompt for a reasoning model based on the query and the document. The reasoning model generates a reasoning result based on the prompt. In some cases, the reasoning result indicates that the document answers the query. A machine learning model provides the document in response to the query based on the reasoning result.

Abstract: Techniques are disclosed for sparse layer-wise training of neural networks. An example system includes at least one processing device including a processor coupled to a memory. The at least one processing device can be configured to implement the following steps: obtaining class predictions while saving activations for only a number ‘k’ layers of a neural network, using the class predictions to calculate a layer shallowness measure for the neural network, using the layer shallowness measure to determine a number ‘u’ of layers to update in the neural network, and partially updating the neural network by training only the number ‘u’ layers of the neural network.

Abstract: Cutting quantum circuits is disclosed. Probability distributions for quantum circuit parameters are generated from historical quantum circuit data. When cutting the quantum circuit, the probability distributions are sampled to obtain a set of initial circuit parameters that can function as constraints in the cutting operation.

Abstract: Implementations utilize a hybrid use of a smaller LLM and a larger LLM to generate and refine content responsive to a user query/request for content generation. In various implementations, the smaller LLM is utilized to process the user query for content generation, to generate initial content responsive to the user query for content generation. The user query for content generation and the initial content can be utilized to generate a text prompt, where the text prompt can be configured to further include a request for focused edit(s). Such a text prompt can be processed using the larger LLM, to generate focused edit(s) to the initial content that refine the initiated content, so that revised content (with improved accuracy) responsive to the user query for content generation is acquired.

Abstract: An edge device includes a microphone and is configured to detect a static position of the edge device for a first time period and to record audio received by the microphone while the static position remains constant. The device determines whether a target sound is within the audio received by the microphone by analyzing audio and sends an alert when the target sound is determined to be within the audio. The target sound may include a firearm discharge. Additionally, or alternatively, the edge device analyzes the audio for types of decibel levels and echoes inside the audio that track to the firearm discharge. The alert may comprise a package of information that includes a location.

Abstract: Enabling gesture input includes obtaining hand tracking data based on one or more camera frames, detecting a contact event between a first finger and a second finger based on the hand tracking data, and determining a first contact location on the first finger and a second contact location on the second finger. In accordance with a determination that the first contact location and the second contact location are within a first predefined gesture zone for a first gesture, an input action is enabled corresponding to the first gesture.

Abstract: Systems and methods for creating a three-dimensional digital model of a scene from a two-dimensional input image of the scene are described. One aspect includes receiving the two-dimensional input image of the scene, and predicting a latent tensor from the input image. The latent tensor comprises three-dimensional geometrical information of the received two-dimensional input image and information about one or more surfaces occluded in the received two-dimensional input image. The latent tensor and the received two-dimensional image may be inputs to the reconstruction neural network that predicts a three-dimensional model. A paired dataset of input images and corresponding latent tensors may be obtained by a joint training of the reconstruction neural network and an encoding neural network that outputs latent tensors from multiview inputs. The process that predicts the latent tensor from the single input image may then be trained on the paired dataset.

Abstract: Power management of a hybrid co-located power plant, based on optimization algorithms oriented towards determining firm and predictable power scheduled, considering the minimization of the storage usage and the minimization of the energy curtailment, including: generating a power generation schedule; determining an energy storage power schedule; determining a power injection schedule; determining a state of charge target schedule of the energy storage system during the predefined time range on the basis of the forecasts parameters; wherein in the predefined time range the method comprises: supplying and/or absorbing the power from the energy storage system to/from the second point of connection according to the energy storage power schedule; controlling the power supply of the power plant from the renewable energy generation equipment to the first point of connection and/or to the energy storage system on the basis of the power injection schedule and the state of charge target.

Abstract: Micro-ferrules and pre-terminated micro-cable assemblies are disclosed. The micro-ferrules are configured for making and using cable assemblies in micro-ducts. The pre-terminated micro-cable assembly includes a main cable and one or more subunits that include a pre-selected number of optical fibers arranged within the main cable. Each of the one or more subunits includes at least a first end terminated with a micro-ferrule.

Abstract: A power management circuit for computer systems includes a power converter circuit that generates different voltage levels at different time periods. Multiple voltage regulator circuits are coupled to the output of the power converter circuit and to respective local power supply nodes. Switch devices are used to bypass the voltage regulator circuits during corresponding ones of the different time periods.

Abstract: The disclosure describes an improved drop-on-demand, controlled volume technique for dispensing resist onto a substrate, which is then imprinted to create a patterned optical device suitable for use in optical applications such as augmented reality and/or mixed reality systems. The technique enables the dispensation of drops of resist at precise locations on the substrate, with precisely controlled drop volume corresponding to an imprint template having different zones associated with different total resist volumes. Controlled drop size and placement also provides for substantially less variation in residual layer thickness across the surface of the substrate after imprinting, compared to previously available techniques. The technique employs resist having a refractive index closer to that of the substrate index, reducing optical artifacts in the device.

Abstract: Dynamic orchestration of quantum circuit execution. A runtime prediction is performed on a quantum circuit to determine or estimate an amount of resources and execution time that are needed to execute the quantum circuit. The runtime prediction may also estimate the resources, time, and other factors associated with cutting the quantum circuit. When cutting the circuit is beneficial, the quantum subcircuits generated by cutting the quantum circuit are subject to runtime prediction. This information generated by the runtime prediction, along with telemetry data from computing and quantum resources and service level objectives, is used to generate an execution plan for performing or executing the quantum circuit or the quantum subcircuits when the quantum circuit is cut.

Abstract: One example method includes receiving a user job that includes a quantum computing workload, evaluating the quantum computing workload, based on the evaluating, solving an orchestration optimization problem by identifying a computing resource for execution of the quantum computing workload, and the computing resource is selected from a defined group of computing resources of a computing infrastructure, and orchestrating the quantum computing workload to the computing resource for execution.

Abstract: Systems and methods for displaying a three-dimensional video at a different frame rate from which the video was stored or acquired are described. In one aspect, a frame N corresponding to a frame number N in the video is received at a video reader and decoder. A 3D mesh including a plurality of vertices, a texture, and one or more offset vectors associated with each vertex of the 3D mesh in the frame N are retrieved and transmitted to a rendering device. At least one vertex of the 3D mesh is adjusted according to the corresponding offset vectors and a playback time moment. The playback time moment is between N and N+1, where N+1 is a frame number corresponding to a frame N+1 immediately following the frame N. The adjusted 3D mesh and the texture are rendered according to one or more camera parameters associated with a display device.

Abstract: A method is disclosed for determining an eye position and/or eye gesture, in particular for use in controlling a function of an application system, in particular a wearable. The method includes (i) recording a profile of at least one EEG signal over time by way of a contact electrode in order to obtain a signal time series of the at least one EEG signal, (ii) providing input data sets from a signal block which arises from the signal time series, for successive time windows, and (iii) evaluating the input data sets in a data-based eye movement model in order to obtain the eye position or eye gesture, the eye movement model being formed using a neural network with one or more recurrent layers, in particular a respective LSTM layer or GRU layer or a transformer layer, and one or more neuron layers.

Abstract: Systems, devices, and methods for an unmanned aerial vehicle (UAV); a trace gas sensor disposed on the UAV, where the gas sensor is configured to measure a gas point concentration; a wind sensor, where the wind sensor is configured to determine a discrete wind vector corresponding to the gas point concentration measurement; and where the discrete wind vector and gas point concentration measurement are acquired substantially concurrently and co-locally.

Abstract: An eyepiece waveguide for augmented reality applications includes a substrate and a set of incoupling diffractive optical elements coupled to the substrate. A first subset of the set of incoupling diffractive optical elements is operable to diffract light into the substrate along a first range of propagation angles and a second subset of the set of incoupling diffractive optical elements is operable to diffract light into the substrate along a second range of propagation angles. The eyepiece waveguide also includes a combined pupil expander diffractive optical element coupled to the substrate.

Abstract: Apparatus and methods are described including a left-ventricular assist device configured to assist left-ventricular functioning of a subject. The left-ventricular assist device includes a frame, a rigid axial shaft extending from a proximal end of the frame to a distal end of the frame, and an impeller disposed on the rigid axial shaft. A motor is disposed outside the subject's body, and a flexible drive cable extends from outside the subject's body to the axial shaft. The drive cable imparts rotational motion from the motor to the rigid axial shaft, to thereby rotate the impeller. Other applications are also described.

Abstract: Systems and methods for providing a mobile memory module in an automatic transfer switch (ATS) allow the memory module to be easily removed from a controller of the ATS. The mobile memory module is a discrete, nonvolatile memory module that stores operational settings for the ATS and can be mounted and removed via a removable fastener. When thus mounted, the mobile memory module is connected to a signal path that allows data communication with a microprocessor on the controller. The microprocessor can retrieve operational settings from the mobile memory module as needed and store updated or new operational settings on the mobile memory module as required. In some embodiments, the mobile memory module has a protective outer housing that is hardened to withstand catastrophic transfer switch failure while still retaining all operational settings stored on the module.