The present invention relates to a method for training artificial neural network configured for 3D lane detection based on unlabelled image data from camera. The method includes generating a first set of 3D lane boundaries in first coordinate system based on first image, generating a second set of 3D lane boundaries in second coordinate system based on second image, transforming at least one of the second set of 3D lane boundaries and first set of 3D lane boundaries based on positional data associated with first image and second image, evaluating the first set of 3D lane boundaries against second set of 3D lane boundaries in common coordinate system in order to find matching lane pairs of first set of 3D lane boundaries and second set of 3D lane boundaries, and updating one or more model parameters of an artificial neural network based on a spatio-temporal consistency loss.

An information handling system includes storage and a processor. The processor identifies an occurrence of an inference model update event; in response to identifying the inference model update event: generates an inference model update package; provides the inference model update package to an entity that generated an inference model used by the information handling system; obtains, from the entity, a hybrid data adapted inference model that is based on the inference model, the inference model update package, and labeled data used to train the inference model; and obtains an inference, using the hybrid data adapted inference model, that indicates a feature is present in collected data.

Dictionary learning method and means for zero-shot recognition can establish the alignment between visual space and semantic space at category layer and image level, so as to realize high-precision zero-shot image recognition. The dictionary learning method includes the following steps: (1) training a cross domain dictionary of a category layer based on a cross domain dictionary learning method; (2) generating semantic attributes of an image based on the cross domain dictionary of the category layer learned in step (1); (3) training a cross domain dictionary of the image layer based on the image semantic attributes generated in step (2); (4) completing a recognition task of invisible category images based on the cross domain dictionary of the image layer learned in step (3).

An example system includes a processor to receive raw and unlabeled videos. The processor is to extract speech from the raw and unlabeled videos. The processor is to extract positive frames and negative frames from the raw and unlabeled videos based on the extracted speech for each object to be detected. The processor is to extract region proposals from the positive frames and negative frames. The processor is to extract features based on the extracted region proposals. The processor is to cluster the region proposals and assign a potential score to each cluster. The processor is to train a binary object detector to detect objects based on positive samples randomly selected based on the potential score.

A computing system can include an embedding model and a clustering model. The computing system input each of the plurality of inputs into the embedding model and receiving respective embeddings for the plurality of inputs as outputs of the embedding model. The computing system can input the respective embeddings for the plurality of inputs into the clustering model and receiving respective cluster assignments for the plurality of inputs as outputs of the clustering model. The computing system can evaluate a clustering loss function that evaluates a first average, across the plurality of inputs, of a respective first entropy of each respective probability distribution; and a second entropy of a second average of the probability distributions for the plurality of inputs. The computing system can modify parameter(s) of one or both of the clustering model and the embedding model based on the clustering loss function.

Systems and methods of the present disclosure provide an improved approach for open-set instance segmentation by identifying both known and unknown instances in an environment. For example, a method can include receiving sensor point cloud input data including a plurality of three-dimensional points. The method can include determining a feature embedding and at least one of an instance embedding, class embedding, and/or background embedding for each of the plurality of three-dimensional points. The method can include determining a first subset of points associated with one or more known instances within the environment based on the class embedding and the background embedding associated with each point in the plurality of points. The method can include determining a second subset of points associated with one or more unknown instances within the environment based on the first subset of points. The method can include segmenting the input data into known and unknown instances.

Embodiments of the present invention concern detecting Test-Time Evasion (TTE) attacks on neural network, particularly deep neural network (DNN), classifiers. The manner of detection is similar to that used to detect backdoors of a classifier whose training dataset was poisoned. Given knowledge of the classifier itself, the adversary subtly (even imperceptibly) perturbs their input to the classifier at test time in order to cause the class decision to change from a source class to a target class. For example, an image of a person who is unauthorized to access a resource can be modified slightly so that the classifier decides the image is that of an authorized person. The detector is based on employing a method (similar to that used to detect backdoors in DNNs) to discover different such minimal perturbations for each in a set of clean (correctly classified) samples, to change the sample's ground-truth (source) class to every other (target) class. For each (source, target) class pair, null distributions of the sizes of these perturbations are modeled. A test sample is similarly minimally perturbed by the detector from its decided-upon (target) class to every other (potential source) class. The p-values according to the corresponding null distributions of these test-sample perturbations are assessed using the corresponding nulls to decide whether the test sample is a TTE attack.

A system includes a computing platform having a hardware processor and a memory storing a software code and a neural network (NN) having multiple layers including a last activation layer and a loss layer. The hardware processor executes the software code to identify different combinations of layers for testing the NN, each combination including candidate function(s) for the last activation layer and candidate function(s) for the loss layer. For each different combination, the software code configures the NN based on the combination, inputs, into the configured NN, a training dataset including multiple data objects, receives, from the configured NN, a classification of the data objects, and generates a performance assessment for the combination based on the classification. The software code determines a preferred combination of layers for the NN including selected candidate functions for the last activation layer and the loss layer, based on a comparison of the performance assessments.

A computer vision system configured for detection and recognition of objects in video and still imagery in a live or historical setting uses a teacher-student object detector training approach to yield a merged student model capable of detecting all of the classes of objects any of the teacher models is trained to detect. Further, training is simplified by providing an iterative training process wherein a relatively small number of images is labeled manually as initial training data, after which an iterated model cooperates with a machine-assisted labeling process and an active learning process where detector model accuracy improves with each iteration, yielding improved computational efficiency. Further, synthetic data is generated by which an object of interest can be placed in a variety of setting sufficient to permit training of models. A user interface guides the operator in the construction of a custom model capable of detecting a new object.

A system and method of curating machine learning training data for improving a predictive accuracy of a machine learning model includes sourcing training data samples based on seeding instructions; returning a corpus of unlabeled training data samples based on a search of data repositories; assigning a distinct classification labels to each of the training data samples of the corpus; computing efficacy metrics for an in-scope corpus of labeled training data samples derived from a subset of training data samples of the corpus that have been assigned one of the plurality of distinct classification labels, wherein the efficacy metrics identify whether the in-scope corpus of labeled training data samples is suitable for training a target machine learning model; and routing the in-scope corpus of labeled training data samples based on the efficacy metrics.