Baidu USA LLC

Baidu Online Network Technology (beijing) Co., Ltd.

Kyland Technology Co., Ltd.

Other

Beijing Baidu Netcom Science Technology Co., Ltd

Beijing Baidu Netcom Science and Technology Co., Ltd.

Beijing Zitiao Network Technology Co., Ltd.

Baidu.com Times Technology (beijing) Co. Ltd.

Shenzhen Grandsun Electronic Co., Ltd.

Beijing Sogou Technology Development Co., Ltd.

Tasly Pharmaceutical Group Co., Ltd.

Beijing Dream Ink Technologies Co., Ltd.

Saint-gobain Glass France

Huawei Digital Power Technologies Co., Ltd.

Array Biopharma Inc.

Ping Li

Rajinder Paul Singh

Mingming Sun

Miao Fan

Haifeng Wang

Dingcheng Li

Jingyuan Zhang

Hongliang Fei

Yue Feng

Xu Liang Li

Weiqun Xie

Xin Wang

Ying Li

Wei Sun

Haiquan Wu

Ruiwen Shi

Dong Li

Zhiwei Yan

Jizhou Huang

Xin Yu

Zhiyuan Zhang

Chao Feng

Shaogang Ren

Qiyun Jiang

Xijun Yan

Shuai Zhang

Peng Liu

Naifeng Wu

Shunnan Zhang

Ying Yan

Wensheng Zhang

Xiaohui Ma

Zhengliang Ye

Yanmei Li

Lihong Zhou

Shiwen Guo

Jianfeng Zhang

Shulong Tan

Ningbo Feng

Xiaohong Shen

Gang Qiao

Hai'ou Dong

Zhixin Zhou

Daya Verma

Lin Guo

Zhenlong Men

Ran Cui

Hanchen Lee

Chaochun Liu

Lin Du

Zehao Dou

Jiemo Li

Shijin Dong

Yanjun Hu

Jingyi Yang

Yingzhi Zhou

Yogita Krishnamachari

Xiaofeng Lin

Hua Yang

Lanlan Zhang

Chenming Li

Tao Liang

Sen Ye

Xueqiang Qiu

Xingpei Tang

Kefeng Meng

Xiaoyu Zhao

Yaliang Li

Xiaowei Sun

Minlong Peng

Mengjun Yang

Qiaojia Tu

Zeyu Dai

An Zhuo

Qianhua Piao

Xiao Lai Chen

Siamak Zamani Dadaneh

Qun Zhang

Xiao Huang

Qiaofei Wang

Yuting Liu

Jinxing Yu

Tan Yu

Dong Xin

Wutao Lin

Junheng Wang

Shiquan Hao

Fenyong Nan

Lianwen Qi

Fenglian Zhang

Minchao Qi

Zhexuan Yang

Jing Yu

Zhilong Xu

Jiayi Fei

Ren Liu

Laychoo Tan

An image segmentation method includes: obtaining an image to be segmented containing a target object; performing at least one image feature fusion processing on associated feature points based on the image to be segmented, and extracting global feature information during each image feature fusion processing, wherein the associated feature points are at least two feature points having a location association relation; and determining, based on the global feature information extracted, a segmentation mask for the target object.

Image segmentation method, device and medium

A multi-degree-of-freedom bionic dexterous hand includes a palm structure, a bionic thumb, a multifunctional bionic finger, and a dexterous bionic finger. The bionic thumb includes a first phalanx structure and a rotary disk, the rotary disk is configured to rotate on the palm structure, the first phalanx structure is hinged to the rotary disk, and the rotation axis of the rotary disk is not parallel to the rotation axis of the first phalanx structure. The multifunctional bionic finger includes a second phalanx structure, the second phalanx structure is universally hinged to the palm structure, and the second phalanx structure can perform the flexion-extension movement and the swinging movement. The dexterous bionic finger includes a third phalanx structure, the third phalanx structure is hinged to the palm structure, and the third phalanx structure can perform the flexion-extension movement.

Multi-degree-of-freedom bionic dexterous hand

Embodiments of the present disclosure disclose a method and apparatus for a scenario of editing multimedia resources, a device and a storage medium. The method includes: in response to an editing instruction triggered for a target editing template, displaying an import page of a material segment to be edited corresponding to the target editing template, wherein a setting entry for the material segment to be edited and a prompt entry for the material segment to be edited are presented on the import page; in response to a trigger for the prompt entry, displaying an original material segment corresponding to the material segment to be edited; in response to a trigger operation for the setting entry, determining the material segment to be edited; and editing the material segment to be edited into a target multimedia resource in a target editing mode indicated by the target editing template.

Method and apparatus for multimedia resource clipping scenario, device and storage medium

An air mattress relates to the field of air mattresses. The air mattress includes a mattress main body, an inflation and deflation device. The mattress main body is provided with an inflation chamber and a first airflow channel. The inflation chamber is communicated to the first airflow channel. The inflation and deflation device includes a driving device, a first air inlet and a first air outlet. When the first air outlet is communicated to the first airflow channel, the inflation and deflation device is in an inflation state, and the driving device is configured to drive external air to pass through the first air inlet and the first air outlet in sequence to form an inflation airflow discharged from the first air outlet; and the inflation airflow discharged from the first air outlet fills the inflation chamber through the first airflow channel to inflate the mattress main body.

Air mattress

An information search method is provided. The method includes: searching for webpages related to a search request through a search engine; extracting respective texts related to the search request from respective webpages and splicing the texts to obtain a spliced text; obtaining a text segment from the spliced text; and sending the obtained text segment to the search engine, to display the obtained text segment in an information search result through the search engine. The present disclosure can bring great advantages to the search engine in terms of user experience and interaction, and can satisfy user requirements for a function of an intelligent question and answer. Through present disclosure, it is beneficial to directly presenting a short text with higher relevance to the search request to the user, thereby saving time in screening information for the user.

Information search method, device, apparatus and computer-readable medium

Inflatable bed

The present disclosure provides a method for selecting an annotated sample. The method includes: determining a first attribute and a second attribute of a sample characteristic; in which the first attribute is a characteristic attribute of the sample characteristic in a source field sample set, and the second attribute is a characteristic attribute of the sample characteristic in a target field sample set; and determining a target annotated sample from a plurality of candidate annotated samples of the source field sample set according to the first attribute and the second attribute; in which the target annotated sample is configured to train a classification model, the classification model includes a model for determining an emotion polarity by analyzing an input sample to be classified.

Method for selecting annotated sample, apparatus, electronic device and storage medium

Learning disentangled representations is an important topic in machine learning for a wide range of applications. Disentangled latent variables represent interpretable semantic information and reflect separate factors of variation in data. Although generative models may learn latent representations and generate data samples as well, existing models may ignore the structural information among latent representations. Described in the present disclosure are embodiments to learn disentangled latent structural representations from data using decomposable variational auto-encoders, which simultaneously learn component representations and encode component relationships. Embodiments of a novel structural prior for latent representations are disclosed to capture interactions among different data components. Embodiments are applied to data segmentation and latent relation discovery among different data components. Experiments on several datasets demonstrate the utility of the present model embodiments.

Learning latent structural relations with segmentation variational autoencoders

The thriving of deep models and generative models provides approaches to model high dimensional distributions. Generative adversarial networks (GANs) can approximate data distributions and generate data samples from the learned data manifolds as well. Presented herein are embodiments to estimate the implicit likelihoods of GAN models. In one or more embodiments, a stable inverse function of the generator is learned with the help of a variance network of the generator. The local variance of the sample distribution may be approximated by the normalized distance in the latent space. Simulation studies and likelihood testing on data sets validate embodiments, which outperformed several baseline methods in these tasks. An embodiment was also applied to anomaly detection. Experiments show that the embodiments herein can achieve state-of-the-art anomaly detection performance.

Estimating the implicit likelihoods of generative adversarial networks

The present disclosure provides a liquid metal conductive paste and an electronic device, and relates to a technical field of new materials. The liquid metal conductive paste provided by the present disclosure includes: 1%-50% by weight of a liquid metal microcapsule, 30%-80% by weight of a conductive powder, 1%-25% by weight of a base polymer and 10%-40% by weight of a solvent. A capsule wall of the liquid metal microcapsule is made of a coating polymer, and a capsule core is made of a liquid metal. Melting point of the liquid metal satisfies: the liquid metal is in a liquid state at least when the wire made of the liquid metal conductive paste is deformed. The present disclosure can achieve a better flexible wire.

Liquid metal conductive paste and electronic device

Embodiments of the present disclosure provide a method and a device for evaluating a comment quality, an electronic device, and a computer readable storage medium. The method includes: selecting a metadata key associated with a comment of an object from metadata of the object, the metadata including a plurality of key-value pairs; determining a value corresponding to the metadata key based on the metadata; and evaluating the comment quality based on the comment and the value corresponding to the metadata key.

Method and device for evaluating comment quality, and computer readable storage medium

Described herein are embodiments for a deep level-wise extreme multi-label learning and classification (XMLC) framework to facilitate the semantic indexing of literatures. In one or more embodiments, the Deep Level-wise XMLC framework comprises two sequential modules, a deep level-wise multi-label learning module and a hierarchical pointer generation module. In one or more embodiments, the first module decomposes terms of domain ontology into multiple levels and builds a special convolutional neural network for each level with category-dependent dynamic max-pooling and macro F-measure based weights tuning. In one or more embodiments, the second module merges the level-wise outputs into a final summarized semantic indexing. The effectiveness of Deep Level-wise XMLC framework embodiments is demonstrated by comparing it with several state-of-the-art methods of automatic labeling on various datasets.

Systems and methods for large scale semantic indexing with deep level-wise extreme multi-label learning

Described herein are embodiments of a framework named as total correlation variational autoencoder (TC_VAE) to disentangle syntax and semantics by making use of total correlation penalties of KL divergences. One or more Kullback-Leibler (KL) divergence terms in a loss for a variational autoencoder are discomposed so that generated hidden variables may be separated. Embodiments of the TC_VAE framework were examined on semantic similarity tasks and syntactic similarity tasks. Experimental results show that better disentanglement between syntactic and semantic representations have been achieved compared with state-of-the-art (SOTA) results on the same data sets in similar settings.

Total correlation variational autoencoder strengthened with attentions for segmenting syntax and semantics

Described herein are embodiments for question answering over knowledge graph using a Knowledge Embedding based Question Answering (KEQA) framework. Instead of inferring an input questions' head entity and predicate directly, KEQA embodiments target jointly recovering the question's head entity, predicate, and tail entity representations in the KG embedding spaces. In embodiments, a joint distance metric incorporating various loss terms is used to measure distances of a predicated fact to all candidate facts. In embodiments, the fact with the minimum distance is returned as the answer. Embodiments of a joint training strategy are also disclosed for better performance. Performance evaluation on various datasets demonstrates the effectiveness of the disclosed systems and methods using the KEQA framework.

Knowledge-graph-embedding-based question answering

Presented herein are embodiments of an unsupervised cross-lingual sentiment classification model (which may be referred to as multi-view encoder-classifier (MVEC)) that leverages an unsupervised machine translation (UMT) system and a language discriminator. Unlike previous language model (LM)-based fine-tuning approaches that adjust parameters solely based on the classification error on training data, embodiments employ an encoder-decoder framework of an UMT as a regularization component on the shared network parameters. In one or more embodiments, the cross-lingual encoder of embodiments learns a shared representation, which is effective for both reconstructing input sentences of two languages and generating more representative views from the input for classification. Experiments on five language pairs verify that an MVEC embodiment significantly outperforms other models for 8/11 sentiment classification tasks.

Cross-lingual unsupervised classification with multi-view transfer learning

Growing community of inventors

Beijing, China

Ping Li