Abstract: A method of analyzing a structural condition of a machine is described. The method includes determining a position of a portion of a machine, a rotating shaft, a fluid transfer system, or a reciprocating machine in operation; collecting an image of the portion with an image collector; synchronizing the determined position of the portion with the collected image of the portion; amplifying the synchronized image; and storing the amplified image to a memory. Related apparatuses, systems, storage media, techniques and articles are also described.

Abstract: Techniques for recommending media are described. A character preference function comprising a plurality of preference coefficients is accessed. A first character model comprises a first set of attribute values for the plurality of attributes of a first character. The first and second characters are associated with a first and second salience value, respectively. A second character model comprises a second set of attribute values for the plurality of attributes of a second character of the plurality of characters. A first character rating is calculated using the plurality of preference coefficients and the first set of attribute values. A second character rating of the second character is calculated using the plurality of preference coefficients with the second set of attribute values. A media rating is calculated based on the first and second salience values and the first and second character ratings. A media is recommended based on the media rating.

Abstract: A vertically-elongated member, which is preferably a support post used in a molten metal pump, includes a ceramic tube and tensioning structures to add a compressive load to the tube along its longitudinal axis. This makes the tube less prone to breakage. Another vertically-elongated member, such as a support post, includes one or more reinforcement members to help alleviate breakage. A device, such as a pump, used in a molten metal bath includes one or more of such vertical members.

Abstract: The techniques described herein relate to constructing and using seed audiences. In an embodiment, a method includes loading, by a processing device, a user event sequence, the user event sequence including a plurality of user events and a plurality of corresponding conversions; generating, by the processing device, a plurality of conversion neighborhoods based on the user event sequence, a given conversion neighborhood in the plurality of conversion neighborhood including at least one conversion rule and a set of user events from the plurality of user events; annotating, by the processing device, each conversion neighborhood in the plurality of conversion neighborhoods with categorical labels; and generating, by the processing device, seed audiences for each conversion neighborhood, a given seed audience including a ranked list of user events for each conversion rule associated with the conversion neighborhood.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to determine vehicle trailer weight. An example apparatus includes a first sensor disposed on a hitch coupled to a vehicle to measure a first rate of change of a weight of the vehicle. The example apparatus also includes a second sensor disposed on a chassis of the vehicle to measure a second rate of change of the weight of the vehicle. The example apparatus also includes a controller to monitor the first rate of change and the second rate of change to determine if the vehicle is improperly loaded.

Abstract: An end-to-end automatic speech recognition (ASR) system includes: a first encoder configured for close-talk input captured by a close-talk input mechanism; a second encoder configured for far-talk input captured by a far-talk input mechanism; and an encoder selection layer configured to select at least one of the first and second encoders for use in producing ASR output. The selection is made based on at least one of short-time Fourier transform (STFT), Mel-frequency Cepstral Coefficient (MFCC) and filter bank derived from at least one of the close-talk input and the far-talk input. If signals from both the close-talk input mechanism and the far-talk input mechanism are present for a speech segment, the encoder selection layer dynamically selects between the close-talk encoder and the far-talk encoder to select the encoder that better recognizes the speech segment. An encoder-decoder model is used to produce the ASR output.

Abstract: An automated program repair tool utilizes a neural transformer model with attention to predict the contents of a bug repair in the context of source code having a bug of an identified bug type. The neural transformer model is trained on a large unsupervised corpus of source code using a span-masking denoising optimization objective, and fine-tuned on a large supervised dataset of triplets containing a bug-type annotation, software bug, and repair. The bug-type annotation is derived from an interprocedural static code analyzer. A bug type edit centroid is computed for each bug type and used in the inference decoding phase to generate the bug repair.

Abstract: Disclosed are techniques for compressing data in a data storage system comprising searching a cluster of nearest neighbors, wherein the cluster has been created using a locality sensitive hashing algorithm, to determine if a data block can be compressed. In alternate embodiments, nearest neighbor clusters can be formed using unsupervised learning. Additionally, nearest neighbors can also be formed in alternate embodiments using one or more of the following algorithms: a k-means clustering algorithm, a k-medoids clustering algorithm, a mean shift algorithm, a generalized method of moment (GMM) algorithm, or a density based spatial clustering of applications with noise (DBSCAN) algorithm.

Abstract: A wireless device receives configuration parameters indicating a first transport block (TB) size threshold of a first random access (RA) type and indicating a second TB size threshold of a second RA type with the second TB size threshold being larger than the first TB size threshold. An RA type of an RA procedure for transmitting a data unit is selected with the selected RA type being the first RA type in response to a size of the data unit being smaller than or equal to the first TB size threshold and the selected RA type being the second RA type in response to the size of the data unit being: larger than the first TB size threshold and smaller than or equal to the second TB size threshold. The data unit is transmitted using the RA procedure based on the selected RA type.

Abstract: Aqueous cathodic electrocoat compositions with improved edge protection and methods for preparing such electrocoat compositions are described herein. In an exemplary embodiment, an electrocoat composition includes water; a crosslinkable resin comprising a binder of an epoxy-amine adduct and a crosslinking agent; a neutralizing agent for neutralizing the epoxy amine adduct, wherein the neutralizing agent has a pKa value of greater than 0; and a pigment paste. The pigment paste may comprise silica, barium sulfate or silica and barium sulfate.

Abstract: One example method for improving the efficiency of robotic surgical procedures via surgical procedure data analysis. The method includes accessing surgical procedure data of a robotic surgical procedure. The surgical procedure data contains data or events associated with the robotic surgical system during the robotic surgical procedure. The method further includes accessing a procedure setup plan for the robotic surgical procedure corresponding to the surgical procedure data and determining an idle period of the surgical procedure based on the surgical procedure data. The method also includes analyzing the surgical procedure data to detect one or more events associated with the idle period to determine a cause of the idle period, and generating a recommendation for modifying the procedure setup plan for the robotic surgical procedure based on the determined cause of the idle period.

Abstract: Metric data may be received, where the metric data may include one or more time series, where each time series may include a set of metric datapoints corresponding to a set of time slices, and where each metric datapoint may be a measurement of a metric of a monitored object in an infrastructure. An infrastructure health score may be computed for each time slice in the set of time slices based on the metric data. A machine learning (ML) model may be trained based on metric datapoints corresponding to at least a current time slice and the infrastructure health score corresponding to a future time slice. The trained ML model may then be used to predict a future infrastructure health score based on at least current metric datapoints.

Abstract: An improved, efficient method for mapping world points from an environment (e.g., points generated by a LIDAR sensor of an autonomous vehicle) to locations (e.g., pixels) within rolling-shutter images taken of the environment is provided. This improved method allows for accurate localization of the world point in a rolling-shutter image via an iterative process that converges in very few iterations. The method poses the localization process as an iterative process for determining the time, within the rolling-shutter exposure period of the image, at which the world point was imaged by the camera. The method reduces the number of times the world point is projected into the normalized space of the camera image, often converging in three or fewer iterations.

Abstract: A process can be used to allow processing of thin layers of a workpiece including dies. The workpiece can include a base substrate and a plurality of layers overlying the base substrate. The process can include forming a polymer support layer over the plurality of layers; thinning or removing the base substrate within a component region of the workpiece, wherein the component region includes an electronic device; and singulating the workpiece into a plurality of dies after thinning or removing the base substrate. In another aspect, an electronic device can be formed using such process. In an embodiment, the workpiece may have a size corresponding to a semiconductor wafer to allow wafer-level, as opposed to die-level, processing.

Abstract: A method includes receiving sampled audio data corresponding to utterances and training a machine learning (ML) model, using the sampled audio data, to generate a high-fidelity audio stream from a low bitrate input bitstream. The training of the ML model includes de-emphasizing the influence of low-probability distortion events in the sampled audio data on the trained ML model, where the de-emphasizing of the distortion events is achieved by the inclusion of a term in an objective function of the ML model, which term encourages low-variance predictive distributions of a next sample in the sampled audio data, based on previous samples of the audio data.

Abstract: The present invention is directed to a method of cleaning and inspection of drilling mud carrying flow lines, each flow line having an inner surface. The method includes affixing a fitting to the flow line, the fitting having a main flow channel and a branch flow channel extending at an angle from said main flow channel. A first end portion of the fitting can have an attachment that enables connection of the fitting to the flow line at one flow line end portion. A second end portion of the fitting can have one or more doors that can be moved between opened and closed positions, the door or doors providing an opening that is not closed when the doors are in the closed position. A suction line can be connected to the branch flow channel. A cleaning tool can be guided first into the fitting and then into the flow line using a fluid carrying cable. The cleaning tool and jetting fluid can be used to clean drilling mud and debris from the inner surface of the flow line.

Abstract: A basketball training device that is usable trains basketball players to sweep the ball low when dribbling against and opposing player or protect the ball if they are being defended by an opposing player with high hands is disclosed. The training device includes a base with a friction surface to engage the gym floor. Upright fibers are able to bend upon contact with a ball and regain their original position.

Abstract: According to the invention, there is provided seed and plants of the corn variety designated CV505385. The invention thus relates to the plants, seeds and tissue cultures of the variety CV505385, and to methods for producing a corn plant produced by crossing a corn plant of variety CV505385 with itself or with another corn plant, such as a plant of another variety. The invention further relates to corn seeds and plants produced by crossing plants of variety CV505385 with plants of another variety, such as another inbred line. The invention further relates to the inbred and hybrid genetic complements of plants of variety CV505385.