A plurality of fault conditions are detected on a communication network onboard a vehicle. The detected fault conditions, a fault condition importance, environment conditions, and a vehicle operation mode are input to a neural network that outputs rankings for respective detected fault conditions. The neural network is trained by determining a loss function based on a maximum likelihood principle that determines a probability distribution that ranks the detected fault conditions. The vehicle is operated based on the rankings of the fault conditions.

Systems, methods, and non-transitory computer-readable media can receive a set of candidate training items for training an early stage model in a multi-stage recall optimization model, wherein the multi-stage recall optimization model comprises the early stage model and a target model. A random subset of the candidate training items is selected from the set of candidate training items. For each training item in the subset of candidate training items, a score is determined based on the target model. Each training item in the subset of candidate training items is labeled with a label based on a probability of the training item being a top-K of the set of candidate training items had the set of candidate training items been scored based on the target model.

The present disclosure provides radio-frequency (RF) systems that can detect the presence of RF signals received by the system, as well as determine characteristics such as the operating frequency of RF signals, the type of RF source that transmitted each RF signal, and/or the location of each RF source with high precision and sensitivity while using low cost, scalable electronics that are versatile enough for deployment in a variety of environments. Such systems can employ a network of RF sensors that can coordinate in response to communication with a computer to perform any such detection and/or determination using trained models executed onboard the RF sensors and/or the computer. RF signals may have unique characteristics when received at one or more RF sensors that may be detected using trained models described herein, even in high noise or non-line of sight (LOS) environments and with low cost, low resolution RF receiver hardware.

A method in an illustrative embodiment includes determining a first set of distilled samples from a first set of samples based on a characteristic distribution of the first set of samples, the first set of samples being associated with a first set of classifications. The method also includes acquiring a first set of characteristic representations associated with the first set of distilled samples. The method also includes adjusting the first set of characteristic representations so that a distance between characteristic representations associated with the same classification is less than a predetermined threshold. The method also includes determining, based on the adjusted first set of characteristic representations, a first set of classification characteristics of the first set of samples and associated with the first set of classifications, the classification characteristics being used to characterize a distribution of characteristic representations of samples having corresponding classifications in the first set of samples.

Described are methods and systems for selecting a highlight segment. The computer-implemented method comprises receiving a sequence of frames, and at least one user data; via a converting module, for each frame, selecting a local neighborhood around it. said neighborhood comprising at least one frame; and converting each neighborhood into a feature vector; via a high-lighting module, assigning a score to each of the feature vectors based on the user data; via a selection module, selecting at least one highlight segment based on the scoring of the feature vectors; and via an outputting module, outputting the highlight segment. The system comprises a receiving module configured to receive a sequence of frames, and at least one user data; a converting module configured to select a local neighborhood around each frame, said neighborhood comprising at least one frame, and convert each neighborhood into a feature vector, a highlighting module configured to assign a score to each of the feature vector based on the user data; a selection module configured to select at least one highlight segment based on the scoring of the feature vectors; and an output component configured to output the highlight segment.

Herein are machine learning techniques that adjust reconstruction loss of a reconstructive model such as an autoencoder based on importances of values of features. In an embodiment and before, during, or after training, the reconstructive model that more or less accurately reconstructs its input, a computer measures, for each distinct value of each feature, a respective importance that is not based on the reconstructive model. For example, importance may be based solely on a training corpus. For each feature during or after training, a respective original loss from the reconstructive model measures a difference between a value of the feature in an input and a reconstructed value of the feature generated by the reconstructive model. For each feature, the respective importance of the input value of the feature is applied to the respective original loss to generate a respective weighted loss. The weighted losses of the features of the input are collectively detected as anomalous or non-anomalous.

Computer-implemented methods, program products, and systems for provenance-based defense against poison attacks are disclosed. In one approach, a method includes: receiving observations and corresponding provenance data from data sources; determining whether the observations are poisoned based on the corresponding provenance data; and removing the poisoned observation(s) from a final training dataset used to train a final prediction model. Another implementation involves provenance-based defense against poison attacks in a fully untrusted data environment. Untrusted data points are grouped according to provenance signature, and the groups are used to train learning algorithms and generate complete and filtered prediction models. The results of applying the prediction models to an evaluation dataset are compared, and poisoned data points identified where the performance of the filtered prediction model exceeds the performance of the complete prediction model. Poisoned data points are removed from the set to generate a final prediction model.

A system for inspecting the reliability of an image. The system may include a processor in communication with a client device; and a storage medium. The storage medium may store instructions that, when executed, configure the processor to perform operations including: obtaining a plurality of images; categorizing the images into a plurality of image classes; calculating a plurality of probability outcomes; determining whether highest predicted probabilities of the images are less than a first threshold and whether an entropy of a predicted density of the probability outcomes exceeds a second threshold; indicating whether the image is associated with the image classes; ranking, the image amongst the plurality of images; filtering, a plurality of low reliability images according to a third threshold; providing, a likelihood of whether a user scanned a vehicle object associated with the image; and identifying a percentage of user scan failures.

Provided is an incorrect triple detection system including a triple selector configured to select a target triple (subject, type, object) in a knowledge base, a sampler configured to create a sentence model by connecting object triples sharing entities included in the target triple, a model builder configured to embed the sentence model into a vector space to create a training entity vector and build an embedding model, and an incorrect triple detector configured to detect an incorrect triple by inputting a test triple into the embedding model.

An example operation includes one or more of sensing from at least one sensor, a longitudinal acceleration and a lateral acceleration, receiving from the at least one sensor, a longitudinal acceleration signal based on the longitudinal acceleration and a lateral acceleration signal based on the lateral acceleration, filtering via at least one logic, the longitudinal acceleration signal and the lateral acceleration signal based on an interquartile range of the longitudinal acceleration signal and the lateral acceleration signal, yielding a plurality of filtered signals and determining via the at least one logic, a mobility index of a transport based on the filtered signals.