Abstract: Methods described herein relate to culling training data for machine learning in a neural network. Methods may include: receiving sensor data from at least one image sensor, where the sensor data is representative of at least one image; processing the at least one image using a neural network; identifying a number of neurons of the neural network that are activated for each of the at least one image; identifying an image as an informative training image in response to the number of neurons activated for the respective image satisfying a predetermined value; and updating the neural network with the informative training image in response to at least one object of the informative training image being manually labeled. The at least one image may be a plurality of images, where the predetermined value may be established based on a distribution of a number of neurons fired for the plurality of images.

Abstract: Systems, methods, and storage media for automatically identifying skin conditions based on images and other data using machine learning methods and algorithms that encode expert knowledge of skin conditions. The systems, methods, and storage media may also leverage infrastructure that facilitates the collection, storage, labeling, dynamic organization, and expert review of existing and incoming data.

Abstract: A method, apparatus and computer program product are provided for enforcing a differential policy for autonomous vehicles along a road. In the context of a method, the may include: determining location information of a vehicle including a road segment and a direction of travel; identifying capabilities of sensors of the vehicle; determining an autonomous or semi-autonomous policy for the road segment specific to the vehicle in response to identifying the capabilities of the sensors of the vehicle; providing a first set of instructions for autonomous control in response to the capabilities of the sensors of the vehicle satisfying a first level of capability; and providing a second set of instructions for autonomous control in response to the capabilities of the sensors of the vehicle satisfying a second level of capability.

Abstract: A method, apparatus and computer program product are provided for iteratively establishing an optimized position of a feature from a plurality of images of the feature. Methods may include: receiving sensor data from at least one image sensor; processing the plurality of images to detect features within each of the plurality of images; for every combination of two images including the first feature, determining an estimated first feature position based on a triangulated position; establishing a spread of estimated first feature positions for the plurality of combinations of two images; determining a best-fit line passing through the spread of estimated first feature positions; iteratively minimizing a perpendicular error projection of each estimated first feature position relative to the best-fit line to obtain a reduced spread of optimized estimated first feature positions; and establishing from the reduced spread of optimized estimated feature positions an optimized estimated first feature position.

Abstract: Systems and methods are provided for distributed video storage and search over edge computing devices having a short-term memory and a long-term memory. The method may comprise caching a first portion of data on a first device. The method may further comprise determining, at a second device, whether the first device has the first portion of data. The determining may be based on whether the first piece of data satisfies a specified criterion. The method may further comprise sending the data, or a portion of the data, and/or a representation of the data from the first device to a third device.

Abstract: Methods described herein relate to using vision-based mapping to provide augmented reality of objects in the environment of the vehicle in response to objects being obscured. Methods may include: receiving sensor data from a vehicle traveling along a road segment; identifying a location of the vehicle on the road segment; retrieving map data from a map database of an environment of the road segment at the location of the vehicle on the road segment; identifying an object in the sensor data representing an obscured object in an environment of the vehicle; and providing for display of at least one object from the map data in an augmented reality of the environment, where the at least one object appears overlaid on the obscured object to a driver of the vehicle. The object in the sensor data may include a road sign having information relating to driving along the road segment.

Abstract: Aspects of the present disclosure relate to systems and methods that aid users with hearing and/or speech impediments to have a conversation with a remote phone number without human assistance. In one aspect, an application residing on a user's device, such as a smartphone, tablet computer, laptop, etc., may be used to initiate a phone call to a recipient. Upon initiating the phone call locally, a service residing on the server may receive a request to initiate a connection to the recipient. Once the recipient answers, the user may converse with the recipient by providing text input to their local app. The text input may be transmitted to the service. The service may use a text to speech converter to translate the received text to speech that can be delivered to the recipient.

Abstract: A method, apparatus and computer program product are provided for establishing the reliability of crowd sourced data based on the source of the data and the region in which the data is gathered. Methods may include: receiving map data for a network of roads in a geographic area; receiving location accuracy data from different regions within the geographic area, where the location accuracy data for each region may be indicative of an accuracy with which location can be established in the respective region; receiving a plurality of probe data points; determining, for each probe data point, the location accuracy data associated with the location information associated with the respective probe apparatus; determining, for each probe apparatus, a reliability of the one or more sensors; and updating a map database based on at least one probe data point.

Abstract: Methods and systems are provided to generate an uncorrupted version of an image given an observed image that is a corrupted version of the image. In some embodiments, a corruption mimicking (“CM”) system iteratively trains a corruption mimicking network (“CMN”) to generate corrupted images given modeled images, updates latent vectors based on differences between the corrupted images and observed images, and applies a generator to the latent vectors to generate modeled images. The training, updating, and applying are performed until modeled images that are input to the CMN result in corrupted images that approximate the observed images. Because the CMN is trained to mimic the corruption of the observed images, the final modeled images represented the uncorrupted version of the observed images.

Abstract: A mechanism to secure the base of a pole to the lid of a flatbox finisher is provided. The mechanism utilizes at least one spring grasping device on the lid of the flat box finisher to grasp and secure the base of the pole. The grasping mechanism is depressed and grasps a generally flat rectangular edge of the base of the pole and secures the base of the pole to the lid of the flatbox finisher by friction. In particular, as the base having the pole of the flatbox finisher is first slid over the grasping mechanism, the grasping mechanism is first pushed downward. After the base slides over the grasping mechanism, the grasping mechanism springs back upward and locks the base of the pole into place on the lid of the flatbox finisher. Wing nuts may then be used to further secure the base of the pole to the lid of the flatbox finisher.

Abstract: The present disclosure describes a device which is able to detect and deter wildlife. The present solution provides for a modular system which can have various modules received in a module housing and those modules can be replaced without the need for the device to be serviced by the manufacturer or a certified repair technician.

Abstract: A task definition is received. The task definition indicates at least a location from which one or more software image can be obtained and information usable to determine an amount of resources to allocate to one or more software containers for the one or more software image. A set of virtual machine instances in which to launch the one or more software containers is determined, the one or more software image is obtained from the location included in the task definition and is launched as the one or more of software containers within the set of virtual machine instances.

Abstract: A resource offloading system comprises a relational database configured to store relational queries including searching features, a server configured to offload the relational queries to non-relational queries, and a non-relational database configured to store the non-relational queries. The server is configured to generate access patterns for each relational query based on searching features included in the relational query, classify the access pattern based on a terminology defined by a searching feature into a classification, rank the classified access patterns, generate candidate access patterns from the ranked access patterns, create non-relational queries based on the candidate access patterns which are ranked beyond a threshold, and store the non-relational queries in the non-relational database.

Abstract: Systems, methods, and computer-readable media are disclosed for preventing duplicate content selection for digital presentation. In one embodiment, an example method may include determining a first bid request for a first content delivery slot at a first time, the first bid request having a first user identifier, determining a ranking of candidate product identifiers for which the first user identifier is eligible at the first time, and determining a second bid request for a second content delivery slot at a second time, the second bid request having the first user identifier, where the second time is within a predetermined time interval of the first time. Example methods may include determining an impression counter value for the first user identifier at the second time, selecting a product identifier from the candidate product identifiers using the impression counter value, and determining a first bid amount using the product identifier.

Abstract: A system for monitoring a vascular access is provided. The system includes a wearable vascular access monitor which can be a sleeve or other protective covering fitted with two or more sensors for obtaining physiological measurements at different locations from the vascular access. The sleeve or other protective covering is also fitted with an ultra-low power processor for relaying the physiological measurements to a patient's mobile phone. The patient's phone can evaluate the physiological measurements to determine a state of the vascular access, and if the physiological measurements fall out of nominal ranges, the patient's phone can alert a clinic, nurse, or physician. The system can be used to monitor fistulas or grafts used for hemodialysis or peritoneal dialysis.

Abstract: A synchronous packet-processing pipeline whose data paths are populated with data-plane stateful processing units (DSPUs) is provided. A DSPU is a programmable processor whose operations are synchronous with the dataflow of the packet-processing pipeline. A DSPU performs every computation with fixed latency. Each DSPU is capable of maintaining a set of states and perform its computations based on its maintained set of states. The programming of a DSPU determines how and when the DSPU updates one of its maintained states. Such programming may configure the DSPU to update the state based on its received packet data, or to change the state regardless of the received packet data.

Abstract: This disclosure describes how to identify objects in an augmented reality environment. More specifically, the various systems and methods described herein describe how an augmented reality device can recognize objects within a real world environment, determine where the object is located, and also identify the various surfaces of the object in real time or substantially real time.

Abstract: An endoscopic surgical clip applier (10) includes a handle assembly (100) configured to releasably engage at least two different endoscopic assemblies (200,300). The handle assembly (100) is configured to transition from a non-ratcheting use condition to a ratcheting use condition when an endoscopic assembly (200,300) configured for ratcheting use is engaged with the handle assembly (100). The handle assembly (100) is maintained in the non-ratcheting use condition when an endoscopic assembly (200,300) configured for non-ratcheting use is engaged with the handle assembly (100). Endoscopic assemblies (200,300) for use with the handle assembly are also provided.

Abstract: An endoscopic surgical clip applier (10) includes a handle assembly (100) configured to releasably engage at least two in different endoscopic assemblies (200, 300). The handle assembly (100) is configured to transition from a non-ratcheting use condition to a ratcheting use condition when an endoscopic assembly (200) configured for ratcheting use is engaged with the handle assembly (100). The handle assembly (100) is maintained in the non-ratcheting use condition when an endoscopic assembly (300) configured for non-ratcheting use is engaged with the handle assembly (100). Endoscopic assemblies (200, 300) for use with the handle assembly (100) are also provided.

Abstract: A memory stores a first and second cache. A processor stores a first query in a transaction log. The processor selects information from the transaction log, uses it to determine that a user will likely access a first portion of the database at a future date, and copies that portion into the first cache prior to the date. The processor further copies a second, frequently accessed portion of the database into the second cache. The processor additionally determines that a portion of data is in both caches and deletes it from the second cache. The processor receives a second query from the user, determines that it is directed at data in the first cache and accesses that cache. The processor also determines that the second query is not directed at data in the first cache but is directed at data in the second cache, and accesses that cache.