Tencent Technology (shenzhen) Company Limited

Genemind Biosciences Company Limited

Lichun Liu

Jing Zhao

Yu Liao

Yabing Feng

Junbin Lai

Haixia Chai

Xuanliang Pan

Bin Zhou

Xiang Wang

Dan Li

Jianfeng Chen

Shenyuan Li

Huguang Jin

Siliang Huang

Jingtao Zhu

Juan Sun

Chunxia Qin

Li Wei Wang

Wei Xu

Chuan Chen

Meng Zhang

Chengtao Fan

Li Zhong

Jianchun Zhao

Chentao Fan

Xiangyong Yang

Zhibing Ai

Yan Feng

Quanchen Lin

Qunting Lin

Gailing Li

Guangli Gan

A data processing method and device are provided. The method includes: extracting a plurality of data sets from unlabeled data; and for each data set, creating a plurality of sample sets by assigning labels to data samples in the data set, respectively training, for each sample set created from the data set, a classifier by using the sample set and labeled data, obtaining a sample set that corresponds to a trained classifier with the highest performance, and adding the obtained sample set to a candidate training set. Each sample set includes the first preset number of data samples with respective labels, the labels of the data samples in each sample set constitutes a label combination, and label combinations corresponding to different sample sets are different from each other. The method also includes adding a second preset number of sample sets in the candidate training set to the labeled data.

Data processing method and device, classifier training method and system, and storage medium

A machine learning model training method includes: training a machine learning model using features of samples in a training set, where a sample in the training set corresponds to an initial first weight and an initial second weight. In one iteration, the method includes: determining a first sample set comprising one or more samples whose corresponding target variables are incorrectly predicted; determining an overall predicted loss of the first sample set based on the predicted losses and corresponding first weights of samples in the first sample set; updating the first weights and second weights of the samples in the first sample set based on the overall predicted loss of the first sample set; and inputting the second weights, the features, and the target variables of the samples in the training set to the machine learning model, and initiating a next iteration of training the machine learning model.

Machine learning model training method and apparatus, server, and storage medium

A machine learning model training method includes: training a machine learning model using features of each sample in a training set based on an initial first weight and an initial second weight. In one iteration, the method includes determining a first sample set in which a target variable is incorrectly predicted, and a second sample set in which a target variable is correctly predicted, based on a predicted loss of each sample; and determining overall predicted loss of the first sample set based on a predicted loss and a first weight of each sample in the first sample set. The method also includes updating the first weight and a second weight of each sample in the first sample set based on the overall predicted loss; and inputting the updated second weight, the features, and the target variable of each sample to the machine learning model, and initiating a next iteration.

A text deduplication method and apparatus, and a storage medium are provided. The method includes: obtaining a text set, the text set including a plurality of to-be-deduplicated texts; capturing, for each to-be-deduplicated text, a corresponding subtext string from the to-be-deduplicated text; and determining, in the text set, to-be-deduplicated texts having a same subtext string, to obtain text subsets. Each subtext string corresponds to a text subset, and each text subset includes one or more to-be-deduplicated texts that have the corresponding subtext string. The method also includes performing text deduplication processing on the text subset corresponding to each subtext string, to obtain a deduplicated text set corresponding to each subtext string; and obtaining, according to the deduplicated text set corresponding to each subtext string, a result text set of the text set after the deduplication.

Text deduplication method and apparatus, and storage medium

A method and a device for recommendation of media content are described. The method includes: calculating scores of media content labels in a computer media content library; in descending order of the scores, selecting a first threshold media content label as a candidate media content label; for the candidate media content label, finding out a media content corresponding to the candidate media content label from the media content library; for the media content corresponding to the candidate media content label, in descending order of the amount of page views, selecting a second threshold media content as a media content to be recommended corresponding to the candidate media content label; and recommending the media content to be recommended corresponding to the candidate media content label to a user. The method and device can recommend popular media contents to users, and reduce the style difference between the media contents.

Method and device for recommendation of media content

The disclosure provides a text information processing method. Training textual data is determined according to text information, and characters and strings are identified from the training textual data. For each of the identified characters, a respective independent probability of appearance among the training textual data is calculated. For each of the identified strings, a respective joint probability of appearance among the training textual data is calculated. Whether a particular string of the identified strings corresponds to a candidate neologism is determined according to independent probabilities of various characters of the particular string and the joint probability of the particular string. Moreover, the candidate neologism is determined as a neologism when the candidate neologism is not in a preset dictionary and a joint probability of the candidate neologism is greater than a preset threshold.

Text information processing method and apparatus

A method for grouping network service users includes acquiring attribute and/or behavior data of multiple users within a current period, and converting the attribute and/or behavior data into standardized data; determining multiple group central points according to the standardized data, and placing the standardized data in a group where a group central point having a shortest distance is located; determining group features of groups according to standardized data in the groups; and separately pushing corresponding service push information to users in the groups according to the group features of the groups. In addition, an apparatus for grouping network service users is further described.

Method and apparatus for grouping network service users

Growing community of inventors

Shenzhen, China

Lichun Liu