The technology disclosed implements Human-in-the-loop (HITL) active learning with a feedback look via a user interface that is expressly designed for the suggested images to admit multiple fast feedbacks, including selection, dismissal, and annotation. Then, the downstream selection policy for subsequent sampling iterations is based on the available data interpreted in the context of the previous selections, dismissals, and annotations.

A medical image processing device includes an image acquisition unit that acquires a medical video image, a brightness analysis unit that analyzes brightness information of each of a plurality of specific medical images within a specific time range among a plurality of medical images constituting the medical video image to output brightness analysis information, and an image selection unit that selects a training medical image to be used for machine learning from among the plurality of specific medical images using the brightness analysis information.

A method for a system and method for morphometric detection of malignancy associated change (MAC) is disclosed including the acts of obtaining a sample; imaging cells to produce 3D cell images for each cell; measuring a plurality of different structural biosignatures for each cell from its 3D cell image to produce feature data; analyzing the feature data by first using cancer case status as ground truth to supervise development of a classifier to test the degree to which the features discriminate between cells from normal or cancer patients; using the analyzed feature data to develop classifiers including, a first classifier to discriminate normal squamous cells from normal and cancer patients, a second classifier to discriminate normal macrophages from normal and cancer patients, and a third classifier to discriminate normal bronchial columnar cells from normal and cancer patients.

A method of updating a trained cardinality estimation model includes receiving a cardinality estimation model with cardinality labels and detecting a drift in underlying data or predicates of the cardinality estimation model. The type of the detected drift is determined and new test queries that mimic test queries for the detected drift are synthesized. A portion of the synthesized test queries is selected to reduce annotation cost and used to update the cardinality estimation model.

Embodiments relate to a system, program product, and method for information extraction and annotation of a data set. Neural models are utilized to automatically attach machine annotations to data elements within an unlabeled data set. The attached machine annotations are evaluated and a score is attached to the annotations. The score reflects a confidence of correctness of the annotations. A labeled data set is iteratively expanded with selectively evaluated annotations based on the attached score. The labeled data set is applied to an unexplored corpus to identify matching corpus data to populated instances of the labeled data set.

An apparatus for text-to-image generation which is a self-supervised based on one-stage generative adversarial network and uses a discriminator network that extracts an image feature may comprise: a text encoder that extracts a sentence vector from input text; a discriminator that determines whether or not an image matches the text from the sentence vector and the image input from a generator; and a decoder that is connected to an encoder inside the discriminator, wherein the decoder and the encoder form an autoencoder structure inside the discriminator.

The present disclosure provides a region of interest (ROI) detection system. The system may be configured to acquire a target image and an ROI detection model, and perform ROI detection on the target image by applying the ROI detection model to the target image. The ROI detection model may be a trained cascaded neural network including a plurality of sequentially connected trained models. The plurality of trained models may include a trained first model and at least one trained second model downstream to the trained first model in the trained cascaded neural network. The plurality of trained models may be sequentially trained. Each of the trained second model may be trained using a plurality of training samples determined based on one or more trained models of the plurality of trained models generated before the generation of the trained second model.

ADAS includes a processing circuit and a memory which stores instructions executable by the processing circuit. The processing circuit executes the instructions to cause the ADAS to receive, from a vehicle that is in motion, a video sequence, generate a position image including at least one object included in the stereo image, generate a second position information associated with the at least one object based on reflected signals received from the vehicle, determine regions each including at least a portion of the at least one object as candidate bounding boxes based on the stereo image and the position image, and selectively adjusting class scores of respective ones of the candidate bounding boxes associated with the at least one object based on whether a respective first position information of the respective ones of the candidate bounding boxes matches the second position information.

A fault diagnosis method for power electronic circuits based on optimizing a deep belief network, including steps. (1) Use RT-LAB hardware-in-the-loop simulator to set up fault experiments and collect DC-link output voltage signals in different fault types. (2) Use empirical mode decomposition to extract the intrinsic function components of the output voltage signal and its envelope spectrum and calculate various statistical features to construct the original fault feature data set. (3) Based on the feature selection method of extreme learning machine, remove the redundancy and interference features, as fault sensitive feature data set. (4) Divide the fault sensitive feature set into training samples and test samples, and primitively determine the structure of the deep belief network. (5) Use the crow search algorithm to optimize the deep belief network. (6) Obtain the fault diagnosis result.

Methods and systems for performing machine learning include selecting first and second training data from one or more training sets in one or more databases. Mixed training data is formed by subtracting a value of each data element in the second training data from a value of a corresponding data element in the first training data. A machine learning process is trained using the mixed training to augment data used by the machine learning process.