Abstract: Provided is a target model training method, a multimodal data processing method, and devices therefor, relating to the field of artificial intelligence technology, and in particular to the fields of computer vision, deep learning, large model and other technologies. The target model training method includes: inputting sample data into a preset model to obtain initial multimodal features of the sample data; and using the initial multimodal features and a preset noise feature to perform model training on N diffusion networks in the preset model to obtain a target model when parameters of an image-text encoder of the preset model are fixed.

Abstract: The present invention discloses an InSAR-based measurement method and system for angle of critical deformation in a coal mining area, and relates to the technical field of geological disaster prevention and control of mountainous coal mines, comprising: acquiring a mining surface deformation field image using a small baseline subset interferometric synthetic aperture radar; and drawing profile lines along the strike direction and dip direction of the coal seam respectively in the mining surface deformation field image. The method provided by the present invention has the advantages of improving time resolution and reducing phase noise.

Abstract: Similarity-based deformable registration of medical images that contain tumor regions often undesirably leads to disproportionate volume changes to tumor regions. An image-registration framework for preserving tumor volumes while promoting similarity in non-tumor part is provided. The framework involves a two-stage process. In the first stage, similarity-based registration is used to identify potential tumor regions by their volume change, generating a soft tumor mask accordingly. In the second stage, a volume-preserving registration network is used. The network uses a novel adaptive volume-preserving loss that penalizes the change in size adaptively based on the masks calculated from the previous stage. The framework balances image similarity and volume preservation in different regions, i.e. non-tumor and tumor regions, by using soft tumor masks to adjust the imposition of volume-preserving loss on each one.

Abstract: A relay directional molecular communication method based on multiple attractants is applied in a field of molecular communication to address the inefficiency of existing technologies. The method for relay nodes is used by fixing three types of attractants, and the relay nodes only release one of the three types of attractants. By using relative priority between two attractants, nanomachines can quickly find a target. Compared with existing multi-attractant multi-relay algorithms, the method does not require many types of attractant resources due to the three types of attractants, saves a variety of the attractant resources, and still achieves an effect greater than three hop relays. It is realized that when a number of the relay nodes is certain, not only the type of attractants used can be reduced, but also the efficiency of finding the target can be maintained.

Abstract: A weakly-supervised framework for segmenting a three-dimensional (3D) medical image incorporates information of geometric prior and contrastive similarity in a loss-based fashion for enhancing distinguishability of an organ in presence of complex organ shape and image artifacts (e.g., low-contrast tissues) in the image. Training images with bounding-box annotations are used to train the framework. The geometric prior is advantageously built on point-cloud representations, which are much more fine-grained and flexible than volume representations, thus serving as better supervision than an inherent property of the bounding-box annotations (i.e. height and width). Furthermore, using contrastive similarity encourages organ pixels to gather around in a contrastive embedding space. It helps more-effectively distinguish low-contrast tissues from the background than using the conventionally-used gray space.

Abstract: Similarity-based deformable registration of medical images that contain tumor regions often undesirably leads to disproportionate volume changes to tumor regions. An image-registration framework for preserving tumor volumes while promoting similarity in non-tumor part is provided. The framework involves a two-stage process. In the first stage, similarity-based registration is used to identify potential tumor regions by their volume change, generating a soft tumor mask accordingly. In the second stage, a volume-preserving registration network is used. The network uses a novel adaptive volume-preserving loss that penalizes the change in size adaptively based on the masks calculated from the previous stage. The framework balances image similarity and volume preservation in different regions, i.e. non-tumor and tumor regions, by using soft tumor masks to adjust the imposition of volume-preserving loss on each one. It ensures that the tumor volumes are preserved during the registration process.

Abstract: A weakly-supervised framework for segmenting a three-dimensional (3D) medical image incorporates information of geometric prior and contrastive similarity in a loss-based fashion for enhancing distinguishability of an organ in presence of complex organ shape and image artifacts (e.g., low-contrast tissues) in the image. Training images with bounding-box annotations are used to train the framework. The geometric prior is advantageously built on point-cloud representations, which are much more fine-grained and flexible than volume representations, thus serving as better supervision than an inherent property of the bounding-box annotations (i.e. height and width). Furthermore, using contrastive similarity encourages organ pixels to gather around in a contrastive embedding space. It helps more-effectively distinguish low-contrast tissues from the background than using the conventionally-used gray space.

Abstract: A relay directional molecular communication method based on multiple attractants is applied in a field of molecular communication to address the inefficiency of existing technologies. The method for relay nodes is used by fixing three types of attractants, and the relay nodes only release one of the three types of attractants. By using relative priority between two attractants, nanomachines can quickly find a target. Compared with existing multi-attractant multi-relay algorithms, the method does not require many types of attractant resources due to the three types of attractants, saves a variety of the attractant resources, and still achieves an effect greater than three hop relays. It is realized that when a number of the relay nodes is certain, not only the type of attractants used can be reduced, but also the efficiency of finding the target can be maintained.

Abstract: This application discloses a CSI feedback method, a related device, and a readable storage medium, and pertains to the field of communication technologies. The method of this application may include: obtaining, by a terminal, N CSI report configurations, where the N CSI report configurations are used for determining M pieces of first information; and sending, by the terminal, a first CSI report, where the first CSI report is calculated based on the M pieces of first information, the first information is any one of the following: first reference signal resource, and first parameter corresponding to the first reference signal resource; N is a positive integer; and M is an integer greater than 1.

Abstract: An active tracking system includes an imager configured to image the temperature of a biological tissue and a heating laser configured to heat regions of the biological tissue. The imager locates high-temperature regions of the biological tissue and the heating laser is controlled to point toward target regions of the biological tissue based on the located high-temperature regions. The active tracking system can be used to control a heating laser to continuously heat a target region of a biological tissue even when the target region moves relative to the heating laser. The active tracking system could allow one or more target regions of a biological tissue to be ‘tagged’ with heat by the heating laser and to be tracked even when the one or more target regions move relative to the heating laser. Devices and methods for operating such active tracking systems are also provided.

Abstract: The invention relates to a multiple-tar-cabin electronic cigarette, comprising an atomizer and a battery assembly. The battery assembly supplies electricity to the atomizer. The atomizer comprises tar cabins and heaters that are arranged in a housing. The tar cabins supply tobacco tar to the heaters. After the heaters atomize the tobacco tar, the atomized tobacco tar is brought away by air that enters via an air inlet of the housing, and is smoked via a cigarette holder. The tar cabins include a first tar cabin and a second tar cabin that are mutually isolated. The heaters include a first heater that receives the tobacco tar from the first tar cabin and a second heater that receives the tobacco tar from the second tar cabin. The two tar cabins in the invention can be injected with tobacco tar of different tastes. In this way, users can enjoy cigarettes of different tastes.

Abstract: A method of hybrid data labeling for machine learning, including receiving multiple unlabeled objects forming an unlabeled data set, pre-labeling the unlabeled data set by a machine learning system to output a pending label data pool, bifurcating the pending label data pool by the machine learning system into high and low confidence sets, dispatching the high confidence set to a machine labeler, dispatching the low confidence set to a human labeler, merging the label sets to return a pre-review label data pool, determining a difference between the pending label data pool and the pre-review label data pool, review labeling the data objects, if the determined difference of the data objects is greater than a predefined error threshold and storing the data objects to a reviewed pool if the determined difference of the data objects is less than and equal to the predefined error threshold.

Abstract: A method of hybrid data labeling for machine learning, including receiving multiple unlabeled objects forming an unlabeled data set, pre-labeling the unlabeled data set by a machine learning system to output a pending label data pool, bifurcating the pending label data pool by the machine learning system into high and low confidence sets, dispatching the high confidence set to a machine labeler, dispatching the low confidence set to a human labeler, merging the label sets to return a pre-review label data pool, determining a difference between the pending label data pool and the pre-review label data pool, review labeling the data objects, if the determined difference of the data objects is greater than a predefined error threshold and storing the data objects to a reviewed pool if the determined difference of the data objects is less than and equal to the predefined error threshold.

Abstract: An active tracking system includes an imager configured to image the temperature of a biological tissue and a heating laser configured to heat regions of the biological tissue. The imager locates high-temperature regions of the biological tissue and the heating laser is controlled to point toward target regions of the biological tissue based on the located high-temperature regions. The active tracking system can be used to control a heating laser to continuously heat a target region of a biological tissue even when the target region moves relative to the heating laser. The active tracking system could allow one or more target regions of a biological tissue to be ‘tagged’ with heat by the heating laser and to be tracked even when the one or more target regions move relative to the heating laser. Devices and methods for operating such active tracking systems are also provided.

Abstract: The invention relates to a multiple-tar-cabin electronic cigarette, comprising an atomizer and a battery assembly. The battery assembly supplies electricity to the atomizer. The atomizer comprises tar cabins and heaters that are arranged in a housing. The tar cabins supply tobacco tar to the heaters. After the heaters atomize the tobacco tar, the atomized tobacco tar is brought away by air that enters via an air inlet of the housing, and is smoked via a cigarette holder. The tar cabins include a first tar cabin and a second tar cabin that are mutually isolated. The heaters include a first heater that receives the tobacco tar from the first tar cabin and a second heater that receives the tobacco tar from the second tar cabin. The two tar cabins in the invention can be injected with tobacco tar of different tastes. In this way, users can enjoy cigarettes of different tastes.

Abstract: An active tracking system includes an imager configured to image the temperature of a biological tissue and a heating laser configured to heat regions of the biological tissue. The imager locates high-temperature regions of the biological tissue and the heating laser is controlled to point toward target regions of the biological tissue based on the located high-temperature regions. The active tracking system can be used to control a heating laser to continuously heat a target region of a biological tissue even when the target region moves relative to the heating laser. The active tracking system could allow one or more target regions of a biological tissue to be ‘tagged’ with heat by the heating laser and to be tracked even when the one or more target regions move relative to the heating laser. Devices and methods for operating such active tracking systems are also provided.

Abstract: A method of data communication rate control is provided. The method includes an application service provider receiving, from a mobile service provider, an indication of a data rate for transmission of data from a server associated with the application service provider to a mobile device. The method also includes transmitting data to the mobile device from the server associated with the application service provider based on the received data rate indication.

Abstract: Systems, devices, and methods for a transformable aerial vehicle are provided. In one aspect, a transformable aerial vehicle includes: a central body and at least two transformable frames assemblies respectively disposed on the central body, each of the at least two transformable frame assemblies having a proximal portion pivotally coupled to the central body and a distal portion; an actuation assembly mounted on the central body and configured to pivot the at least two frame assemblies to a plurality of different vertical angles relative to the central body; and a plurality of propulsion units mounted on the at least two transformable frame assemblies and operable to move the transformable aerial vehicle.

Abstract: A tool discharges one battery pack and a charger simultaneously charges another battery pack to provide for continuous and circular operation of the tool. Tool system includes a tool and a lithium based battery pack having a nominal voltage of at least 56 volts for supplying power to the tool. A heat absorbing device is associated with the battery pack and includes a phase-change material.

Abstract: Systems and methods for using radio layer information to enhance network transport protocols are provided. Channel characteristics are obtained from a radio layer in a mobile device. The channel characteristics indicate the quality of a connection between the mobile device and a base station. Based on the channel characteristics, a bandwidth of the connection between the mobile device and the base station is calculated. A server is instructed to transmit data to the mobile device at the data rate determined based on the determined bandwidth.