Abstract: A language processing method includes following steps. An initial dataset including initial phrases and initial intent labels about the initial phrases is obtained. A first intent classifier is trained with the initial dataset. Augmented phrases are produced corresponding to the initial phrases by sentence augmentation. First predicted intent labels about the augmented phrases and first confidence levels of the first predicted intent labels are generated by the first intent classifier. The augmented phrases are classified into augmentation subsets according to comparisons between the first predicted intent labels and the initial intent labels and according to the first confidence levels. A second intent classifier is trained according to a part of the augmentation subsets by curriculum learning. The second intent classifier is configured to distinguish an intent of an input phrase within a dialogue.

Abstract: A data classification method, for classifying unlabeled images into an inlier data set or an outlier data set, include following steps. The unlabeled images are obtained. An assigned inlier image is selected among the unlabeled images. A similarity matrix is computed and the similarity matrix includes first similarity scores of the unlabeled images relative to the assigned inlier image. Each of the unlabeled images is classified into an inlier data set or an outlier data set according to the similarity matrix, so as to generate inlier-outlier predictions of the unlabeled images.

Abstract: A multi-label classification method for generating labels annotated on medical images. An initial dataset including medical images and partial input labels is obtained. The partial input labels annotate a labeled part of abnormal features on the medical images. A first multi-label classification model is trained with the initial dataset. Difficulty levels of the medical images in the initial dataset are estimated based on predictions generated by the first multi-label classification model. The initial dataset is divided based on the difficulty levels of the medical images into different subsets. A second multi-label classification model is trained based on subsets with gradually increasing difficulty levels during different curriculum learning rounds. Predicted labels annotated on the medical images are generated about each of the abnormal features based on the second multi-label classification model.

Abstract: A medical image processing method includes following steps. A first medical image about a first patient under a first examination condition is obtained. A second medical image about the first patient under a second examination condition is obtained. A first label corresponding to the first medical image is collected. The first label marks a lesion within the first medical image. A transformation function between the first medical image and the second medical image is calculated by aligning the first medical image with the second medical image. The transformation function is applied to convert the first label into a second label corresponding to the second medical image.

Abstract: A medical image detection system includes a memory and a processor. The processor is configured to execute the neural network model stored in the memory. The neural network model includes a feature extractor, a feature pyramid network, a first output head and a second output head. The feature extractor is configured for extracting intermediate tensors from a medical image. The feature pyramid network is associated with the feature extractor. The feature pyramid network is configured for generating multi-resolution feature maps according to the intermediate tensors. The first output head is configured for generating a global prediction according to the multi-resolution feature maps. The second output head is configured for generating local predictions according to the multi-resolution feature maps. The processor is configured to generate output information based on the medical image, the global prediction and the local predictions.

Abstract: A control method for a reinforcement learning system includes following operations. The reinforcement learning system obtains training data relating to an interaction system. The interaction system interacts with a reinforcement learning agent. A neural network model is utilized by the reinforcement learning agent for selecting sequential actions from a set of candidate actions. The neural network model is trained to maximize cumulative rewards collected by the reinforcement learning agent in response to the sequential actions. During training of the neural network model, auxiliary rewards of the cumulative rewards are provided to the reinforcement learning agent according to a comparison between symptom inquiry actions of the sequential actions and diagnosed symptoms in the training data.

Abstract: A medical system is able to provide a symptom query interpretation and/or a disease diagnosis interpretation. The medical system includes an interface and a processor. The interface is configured for receiving an input state. The processor is coupled with the interface. The processor is configured to execute a symptom checker to select a current action, from a plurality of candidate symptom queries and a plurality of candidate disease predictions, according to the input state. In response to the current action is a first symptom query, the processor is configured to execute an interpretable module interacted with the symptom checker to generate a diagnostic tree for simulating possible diagnosis paths, and generate a symptom query interpretation about the first symptom query according to the diagnostic tree.

Abstract: A machine learning method includes steps of: obtaining, by a processor, a model parameter from a memory, and performing, by a processor, a classification model according to the model parameter, wherein the classification model comprises a plurality of neural network structural layers; calculating, by the processor, a first loss and a second loss according to a plurality of training samples, wherein the first loss corresponds to an output layer of the plurality of neural network structural layers, and the second loss corresponds to one, which is before the output layer, of the plurality of neural network structural layers; and performing, by the processor, a plurality of updating operations for the model parameter according to the first loss and the second loss to train the classification model.

Abstract: A training method suitable for a reinforcement learning system with a reward function to train a reinforcement learning model and including: defining at least one reward condition of the reward function; determining at least one reward value range corresponding to the at least one reward condition; searching for at least one reward value from the at least one reward value range by a hyperparameter tuning algorithm; and training the reinforcement learning model according to the at least one reward value.

Abstract: A method for controlling a medical system includes the following operations. The medical system receives an initial symptom. A neural network model is utilized to select at least one symptom inquiry action. The medical system receives at least one symptom answer to the at least one symptom inquiry action. A neural network model is utilized to select at least one medical test action from candidate test actions according to the initial symptom and the at least one symptom answer. The medical system receives at least one test result of the at least one medical test action. A neural network model is utilized to select a result prediction action from candidate prediction actions according to the initial symptom, the at least one symptom answer and the at least one test result.

Abstract: A control method for a reinforcement learning system includes following operations. The reinforcement learning system obtains training data relating to an interaction system. The interaction system interacts with a reinforcement learning agent. A neural network model is utilized by the reinforcement learning agent for selecting sequential actions from a set of candidate actions. The neural network model is trained to maximize cumulative rewards collected by the reinforcement learning agent in response to the sequential actions. During training of the neural network model, auxiliary rewards of the cumulative rewards are provided to the reinforcement learning agent according to a comparison between symptom inquiry actions of the sequential actions and diagnosed symptoms in the training data.