Abstract: A rail vehicle having rail wheels accommodated to guide the rail vehicle along a railway track. Each of the wheels is connected to the vehicle by an intermediate axle box that includes at least one accelerometer. A measurement system includes a receiver for receiving signals from the at least one accelerometer, and a vehicle-railway track interaction model that estimates the expected signals from the at least one accelerometer. The receiver and the vehicle-railway track interaction model connect to a comparator to compare the measured signals and the expected signals from the at least one accelerometer. The comparator connects to a tuning portion of the measurement system, which tuning portion is arranged to adjust parameters of the vehicle-railway track interaction model so as to provide a closer fit of the estimation of the expected signals from the at least one accelerometer with the measured signals from the at least one accelerometer.

Abstract: A system and a method are disclosed for a machine learned audio morpher that is trained such that the voice characteristics of a user spoken phrase are replaced with those of a target speaker, which removes and/or reduces the user identifiable information for the spoken phrase. Training can be performed by a user and a target speaker speaking the same or similar phrases and training the audio morpher to minimize the differences between the target speaker phrase and a morphed user phrase.

Abstract: Training and enhancement of neural network models, such as from private data, are described. A slave device receives a version of a neural network model from a master. The slave accesses a local and/or private data source and uses the data to perform optimization of the neural network model. This can be done such as by computing gradients or performing knowledge distillation to locally train an enhanced second version of the model. The slave sends the gradients or enhanced neural network model to a master. The master may use the gradient or second version of the model to improve a master model.

Abstract: Automated processing, enforcement and intelligent management of vehicle operation violations is disclosed. A method, that is also disclosed, includes obtaining at least one image within which is shown at least a portion of a vehicle. The method also includes receiving image data for the at least one image, and analyzing the image data to generate a violation score for each of one or more potential vehicle operation violations for the vehicle. When the violation score is in-between an upper violation threshold and a lower consideration threshold, user input is obtained that either affirms or rejects existence of a violation and then, once the user input is obtained, a violation notification is generated only when the user input affirms the existence of the violation. When the violation score is higher than the violation threshold, the violation notification is generated without the user input being obtained.

Abstract: A system for automated review of public safety incident reports include receiving structured incident data for an incident report from a submitting public safety officer including incident type information for the incident, receiving unstructured incident narrative text describing the incident, accessing an unstructured incident narrative feedback checking model applicable to incidents of the incident type, applying the model to the narrative text in light of supplemental information in the structured incident data or obtained from another source, identifying, by application of the model, matters in the narrative text likely to be flagged for correction by a human reviewer during a subsequent review, and providing feedback notifying the officer of the identified matters. The model may be retrained based on feedback or corrections provided by the officer in response to the notification of the identified matters or in response to requests for correction subsequently received from human reviewers.

Abstract: A device, system and method for providing indications of a discrepancy between a video and a textual description of an incident. A public-safety computing device determines a discrepancy between a video of an incident and a textual description of the incident, the textual description being a component of a public-safety incident report associated with the incident. The device, in response to determining the discrepancy, controls a notification device to provide an indication of the discrepancy.

Abstract: A system for updating metadata associated with historic data. The system includes an electronic computing device. The electronic computing device is configured to receive historic data and metadata associated with the historic data and, using a low capacity machine learning model, analyze the metadata associated with the historic data to determine a probability that the historic data includes the object of interest. The electronic computing device is also configured to compare the probability to a predetermined threshold. The electronic computing device is further configured to use a high capacity machine learning model to analyze the historic data to determine whether the historic data includes the object of interest and update the metadata of the historic data based on whether the historic data includes the object of interest to generate updated metadata, when the probability is greater than or equal to the predetermined threshold.

Abstract: A driver interface for use within an automobile provides responses to voice commands issued for example by a driver of the automobile. The interface includes a camera and microphone for capturing image data such as gestures and audio data from the automobile driver. The image data and audio data are processed to extract image and linguistic features from the image and audio data, which image and linguistic features are processed to interpret and infer a meaning of the voice command.

Abstract: Systems and methods for processing speech are described. In certain examples, image data is used to generate visual feature tensors and audio data is used to generate audio feature tensors. The visual feature tensors and the audio feature tensors are used by a linguistic model to determine linguistic features that are usable to parse an utterance of a user. The generation of the feature tensors may be jointly configured with the linguistic model. Systems may be provided in a client-server architecture.

Abstract: A method for managing intangible shopping transactions in a physical retail store is provided. The method may include monitoring a customer shopping within a physical retail store to detect occurrence of an intangible transaction. The method may also include determining that the intangible transaction has occurred. The method may further include storing a record of a transaction item associated with the intangible transaction, the record associated with the customer. The method may additionally include determining the customer is attempting to leave the physical retail store while not in possession of the transaction item. The method may also include providing an alert to indicate the customer is attempting to leave the physical retail store while not in possession of the transaction item.

Abstract: A device, system and method for providing indications of a discrepancy between a video and a textual description of an incident. A public-safety computing device determines a discrepancy between a video of an incident and a textual description of the incident, the textual description being a component of a public-safety incident report associated with the incident.

Abstract: An audio processing system is described. The audio processing system uses a convolutional neural network architecture to process audio data, a recurrent neural network architecture to process at least data derived from an output of the convolutional neural network architecture, and a feed-forward neural network architecture to process at least data derived from an output of the recurrent neural network architecture. The feed-forward neural network architecture is configured to output classification scores for a plurality of sound units associated with speech. The classification scores indicate a presence of one or more sound units in the audio data. The convolutional neural network architecture has a plurality of convolutional groups arranged in series, where a convolutional group includes a combination of two data mappings arranged in parallel.

Abstract: Automated processing, enforcement and intelligent management of vehicle operation violations is disclosed. A method, that is also disclosed, includes obtaining at least one image within which is shown at least a portion of a vehicle. The method also includes receiving image data for the at least one image, and analyzing the image data to generate a violation score for each of one or more potential vehicle operation violations for the vehicle. When the violation score is in-between an upper violation threshold and a lower consideration threshold, user input is obtained that either affirms or rejects existence of a violation and then, once the user input is obtained, a violation notification is generated only when the user input affirms the existence of the violation. When the violation score is higher than the violation threshold, the violation notification is generated without the user input being obtained.

Abstract: A system for automated review of public safety incident reports include receiving structured incident data for an incident report from a submitting public safety officer including incident type information for the incident, receiving unstructured incident narrative text describing the incident, accessing an unstructured incident narrative feedback checking model applicable to incidents of the incident type, applying the model to the narrative text in light of supplemental information in the structured incident data or obtained from another source, identifying, by application of the model, matters in the narrative text likely to be flagged for correction by a human reviewer during a subsequent review, and providing feedback notifying the officer of the identified matters. The model may be retrained based on feedback or corrections provided by the officer in response to the notification of the identified matters or in response to requests for correction subsequently received from human reviewers.

Abstract: Training and enhancement of neural network models, such as from private data, are described. A slave device receives a version of a neural network model from a master. The slave accesses a local and/or private data source and uses the data to perform optimization of the neural network model. This can be done such as by computing gradients or performing knowledge distillation to locally train an enhanced second version of the model. The slave sends the gradients or enhanced neural network model to a master. The master may use the gradient or second version of the model to improve a master model.

Abstract: Systems and methods for distributed training of a neural network model are described. Various embodiments include a master device and a slave device. The master device has a first version of the neural network model. The slave device is communicatively coupled to a first data source and the master device, and the first data source is inaccessible by the master device, in accordance with one embodiment. The slave device is remote from the master device. The master device is configured to output first configuration data for the neural network model based on the first version of the neural network model. The slave device is configured to use the first configuration data to instantiate a second version of the neural network model. The slave device is configured to train the second version of the neural network model using data from the first data source and to output second configuration data for the neural network model.

Abstract: Systems and methods for processing speech are described. In certain examples, image data is used to generate visual feature tensors and audio data is used to generate audio feature tensors. The visual feature tensors and the audio feature tensors are used by a linguistic model to determine linguistic features that are usable to parse an utterance of a user. The generation of the feature tensors may be jointly configured with the linguistic model. Systems may be provided in a client-server architecture.

Abstract: Rail vehicle (1) having rail wheels (3,4) accommodated to guide the rail vehicle along a railway track (2) and said vehicle comprising means for detection of a flaw or flaws in the railway track, wherein the rail vehicle is provided with a noncontact vibrometer (9,10) which is arranged to measure vibrational movement of the railway track surface.

Abstract: Rail vehicle (1) having rail wheels (3,4) accommodated to guide the rail vehicle along a railway track (2) and said vehicle comprising means for detection of a flaw or flaws in the railway track, wherein the rail vehicle is provided with a noncontact vibrometer (9,10) which is arranged to measure vibrational movement of the railway track surface.

Abstract: A SWIR hyperspectral imaging filter has serial stages along an optical signal path with angularly distributed birefringent retarders and polarizers. The retarders can include active retarders such as tunable liquid crystal birefringent elements, passive retarders such as fixed retarders, and/or combinations thereof. Distinctly different periodic transmission spectra are provided by different filter stages, each having multiple retarders, in particular with some stages having broad bandpass peaks at wide spectral spacing and other stages have very narrow closely spaced peaks. The respective spectra include at least one tunably selectable band at which the transmission spectra of the filter stages coincide, whereby the salutary narrow bandpass and wide spectral spacing ranges of different stages apply together, resulting in a high finesse wavelength filter suitable for spectral imaging.