Venner Shipley LLP

In some examples, systems and methods for multiple-sensor object tracking are provided. For example, a method includes: receiving one or more first image transformations generated based on a plurality of previous images from a plurality of sensor feeds respectively; receiving a plurality of images from the plurality of sensor feeds; generating a transformation update based on the one or more image transformations and the plurality of images using an image registration model; generating a second image transformation based on the transformation update and the one or more first image transformations; applying the second image transformation to the plurality of images to generate a set of transformed images; and identifying a plurality of objects using a multiple object tracking model based on the set of transformed images.

Implementations enable scalable generation of high-quality, diverse training data for vision-language model(s) (VLM(s)). Processor(s) of a system can configure a dialog between at least a first VLM and a second VLM, cause the dialog to be conducted, and generate training instance(s) based on the dialog. In configuring the dialog, the first VLM is provided by a target image and a first set of instructions for the dialog, and the second VLM is provided with an ordered set of candidate images (e.g., the target image and additional image(s)) and a second set of instructions for the dialog. In causing the dialog to be conducted, the second VLM generates question(s) (e.g., using the second set of instructions) to ask the first VLM in furtherance of identifying the target image, in the ordered set of candidate images, and the first VLM generates response(s) (e.g., using the first set of instructions) to the question(s).

In some examples, systems and methods for multiple-sensor object tracking are provided. For example, a method includes: receiving a plurality of tracks generated based on a plurality of images from a plurality of sensor feeds, the plurality of sensor feeds including a first sensor feed and a second sensor feed, the plurality of tracks including a first track detected in a first image from the first sensor feed and a second track detected in a second image from the second sensor feed; determining a candidate association between the first track and the second track; generating an association confidence score corresponding to the candidate association based at least in part on a track history; evaluating whether the candidate association meets one or more association criteria; and in response to the candidate association meeting the one or more association criteria, merging the first track and the second track to generate a merged track; and updating the plurality of tracks to include the merged track replacing the first track and the second track.

In some examples, systems and methods for data mapping are provided. An example method includes a method for data mapping is provided. In some examples, the method includes receiving the unstructured data including text data; receiving one or more extracted data parts from the GenAI model; accessing a hierarchical data model, the hierarchical data model including one or more sub-hierarchies; generating a search prompt using the unstructured data and the one or more highest-level structure elements; providing the search prompt to the GenAI model; receiving one or more identified sub-hierarchies; for each data part of the one or more extracted data parts, traversing each identified sub-hierarchy of the one or more identified sub-hierarchies to determine a list of relevant structure elements in the hierarchical data model based on a relevance scoring; and generating a mapping prompt based on the list of relevant structure elements and the unstructured data.

According to one aspect, a magnetic flux concentrator includes: a ferromagnetic material having a loop shape interrupted by an air gap region and forming a central opening, wherein the air gap region includes a first side having at least two protrusions and a second side having at least one protrusion interleaved with, and not in mechanical contact with, the protrusions of the first side. According to another aspect, a sensor utilizes such a magnetic flux concentrator to detect a difference in current between two conductors.

An irrigation system for medical use, such as a rectal irrigation system, and a method for evaluation of such a system, are disclosed. The system comprises a catheter, an irrigation liquid reservoir and an electrically operated diaphragm pump for pumping air into the irrigation liquid reservoir and/or into the catheter. A pressure sensor is associated with an output end of the diaphragm pump and a controller evaluates the operation of the diaphragm pump by: measuring the pressure at two opposite end positions for a diaphragm of the diaphragm pump in at least two different pump cycles; estimating a current effective output volume per pumping cycle based on the measured pressures; comparing this to a previously determined, primary effective output volume; and provide an evaluation of the operation of the diaphragm pump.

In some embodiments, a method includes extracting one or more request features from request data for a transaction, generating one or more derived features based on the request features, encoding the one or more request features and the one or more derived features into a format for a security decision model, and generating, using the security decision model, a security protocol decision based on model features including the one or more request features and the one or more derived features.

Implementations enable efficient identification of model vulnerabilities and subsequent model improvement without extensive real-world testing, leading to more robust and reliable models. Some implementations are directed to causing a training image to be processed to generate a synthetic image that includes one or more variations to the training image. The training image is one based on which a model has been trained. The variations in the synthetic image can include various changes, such as adding a new object, changing lighting, modifying the background, or altering the position of an object. The synthetic image is processed, using the model, to generate output. This output is then processed to determine whether the synthetic image is an outlier for the model. In response to determining that the synthetic image is an outlier for the model, remediating action(s) are performed, such as causing the model to be further trained based on the synthetic image.

A computing system can facilitate discovery and one-on-one communication among multiple software applications of a multi-application execution environment. The system can enable software applications to broadcast and/or detect advertisement messages indicative of discoverable software applications and supported capabilities, such as for targeted messaging, sharing state, and/or the like. The system can enable software applications to detect and invoke capabilities of other software applications from outside said software applications. The system can enable software applications to share their states such that states are synchronized among multiple software applications. In some embodiments, the system can cause generation and display of interactive graphical user interfaces (GUIs) and/or of various GUI elements configured to indicate and/or facilitate discovery and/or one-on-one communication among multiple software applications, such as for detecting and invoking functionalities of software applications from outside said software applications, and/or such as for sharing states among multiple software applications.

A hybrid electro-optic modulator is described. The modulator comprises an electro-optic layer configured to support a hybrid mode propagating in a first direction, a pair of electrodes spaced apart in a second direction perpendicular to the first direction, the electrodes being positioned on the same side of the electro-optic layer and between the propagation path of the hybrid mode, wherein each electrode has a cross-sectional shape in a plane defined by the second direction and a third direction perpendicular to the first and second directions, the cross-sectional shape including a lower region joined to an elevated region via an angled region. The lower region is closer to the electro-optic layer than the elevated region in the third direction, the angled region extends towards the propagation path of the hybrid mode in the second direction, and the pair of elevated regions are closer than the pair of lower regions in the second direction.