Intel Corporation

Beijing Youzhuju Network Technology Co., Ltd.

Tao Kong

Anbang Yao

Yurong Chen

Hao Zhao

Yun Ren

Ya Jing

Hongtao Wu

Lei Li

Hang Li

Jie Xu

Ming Lu

Yiwen Guo

Hanbo Zhang

Minghuan Liu

Chilam Cheang

Xinghang Li

Taozheng Yang

Minzhao Zhu

Kuankuan Sima

Qie Sima

Embodiments of the disclosure relate to a method and apparatus for controlling a robot, a device, a robot, and a medium. The method for controlling a robot according to the embodiments of the disclosure includes determining, based on a real-time image captured by the robot at a first moment, a reference motion parameter and a reference control force corresponding to the real-time image. The method further includes determining a target pose and a target control force of the robot at a second moment after the first moment according to the reference motion parameter and the reference control force. The method further includes determining a target action of the robot at the second moment according to the target pose and the target control force.

Method and apparatus for robot control, device, robot, and medium

The present disclosure relates to a localization method and apparatus, an electronic device, and a storage medium. The localization method comprises: acquiring a target local semantic point cloud map, a global semantic grid map, and a laser point cloud map; performing pose identification on the basis of the target local semantic point cloud map and the global semantic grid map to obtain a set of candidate poses; and determining a target pose according to the laser point cloud map and the set of candidate poses.

Localization method and apparatus, electronic device, and storage medium

The present application discloses an information processing method, a task execution method, an apparatus, a device and a medium. The method includes: processing, through a target visual encoding model in a target analysis model, obtained image information to be analyzed, to obtain a corresponding target sequence; fusing, through a target feature fusion model in the target analysis model, the target sequence and obtained text information to be analyzed, to obtain a target fusion result; processing the target fusion result through a target task analysis model in the target analysis model to obtain target task information; and controlling the action execution apparatus to perform an action corresponding to the target task information. The target analysis model is obtained by training an initial analysis model and the initial analysis model comprises an initial visual encoding model and an initial feature fusion model.

Information processing method, task execution method, apparatus, device and medium

An example apparatus for mining multi-scale hard examples includes a convolutional neural network to receive a mini-batch of sample candidates and generate basic feature maps. The apparatus also includes a feature extractor and combiner to generate concatenated feature maps based on the basic feature maps and extract the concatenated feature maps for each of a plurality of received candidate boxes. The apparatus further includes a sample scorer and miner to score the candidate samples with multi-task loss scores and select candidate samples with multi-task loss scores exceeding a threshold score.

Joint training of neural networks using multi-scale hard example mining

Region proposal is described for image regions that include objects of interest. Feature maps from multiple layers of a convolutional neural network model are used. In one example a digital image is received and buffered. Layers of convolution are performed on the image to generate feature maps. The feature maps are reshaped to a single size. The reshaped feature maps are grouped by sequential concatenation to form a combined feature map. Region proposals are generated using the combined feature map by scoring bounding box regions of the image. Objects are detected and classified objects in the proposed regions using the feature maps.

Region proposal for image regions that include objects of interest using feature maps from multiple layers of a convolutional neural network model

A convolutional neural network framework is described that uses reverse connection and obviousness priors for object detection. A method includes performing a plurality of layers of convolutions and reverse connections on a received image to generate a plurality of feature maps, determining an objectness confidence for candidate bounding boxes based on outputs of an objectness prior, determining a joint loss function for each candidate bounding box by combining an objectness loss, a bounding box regression loss and a classification loss, calculating network gradients over positive boxes and negative boxes, updating network parameters within candidate bounding boxes using the joint loss function, repeating performing the convolutions through to updating network parameters until the training converges, and outputting network parameters for object detection based on the training images.

Convolutional neural network framework using reverse connections and objectness priors for object detection

Joint training of neural networks using multi scale hard example mining

Growing community of inventors

Beijing, China

Tao Kong