Data labeling method, apparatus and system
AssigneeBeijing Baidu Netcom Science Technology Co., Ltd.
InventorsGuanchao Wang, Yuqian Jiang, Shuhao Zhang, Tao Jiang, Siqi Wang
Life of this patent
Estimating the term…
Abstract
A data labeling method, apparatus and system are provided. The method includes: sampling a data source according to an evaluation task for the data source to obtain sampled data; generating a labeling task from the sampled data; sending the labeling task to a labeling device; and receiving a labeled result of the labeling task from the labeling device. As such, an automatic evaluation of data can be implemented by using the evaluation task, and evaluation efficiency is improved.
Claims25
A data labeling method, comprising:
- sampling a data source according to an evaluation task for the data source, to obtain sampled data;
- generating a labeling task from the sampled data;
- sending the labeling task to a labeling device;
- and
- receiving a labeled result of the labeling task from the labeling device,
- wherein the method further comprises:
- generating the evaluation task for the data source, wherein the evaluation task comprises an evaluation index, and the evaluation index has a corresponding sampling rule.
A data labeling apparatus, comprising:
- one or more processors;
- and
- a storage device configured to store one or more programs, wherein
- the one or more programs, when executed by the one or more processors, cause the one or more processors to:
- sample a data source according to an evaluation task for the data source, to obtain sampled data;
- generate a labeling task from the sampled data;
- send the labeling task to a labeling device;
- and
- receive a labeled result of the labeling task from the labeling device,
- wherein the one or more programs, when executed by the one or more processors, cause the one or more processors further to:
- generate an evaluation task for the data source, wherein the evaluation task comprises an evaluation index, and the evaluation index has a corresponding sampling rule.
A non-transitory computer-readable storage medium, comprising computer executable instructions stored thereon, wherein the executable instructions, when executed by a computer, cause the computer to:
- sample a data source according to an evaluation task for the data source, to obtain sampled data;
- generate a labeling task from the sampled data;
- send the labeling task to a labeling device;
- and
- receive a labeled result of the labeling task from the labeling device,
- wherein the executable instructions, when executed by a computer, further cause the computer to:
- generate the evaluation task for the data source, wherein the evaluation task comprises an evaluation index, and the evaluation index has a corresponding sampling rule.
A data labeling method, comprising:
- sampling a data source according to an evaluation task for the data source, to obtain sampled data;
- generating a labeling task from the sampled data;
- sending the labeling task to a labeling device;
- and
- receiving a labeled result of the labeling task from the labeling device,
- wherein the method further comprises:
- receiving an inspection result and address information of a first data source which has passed a first inspection;
- and
- performing a second inspection on the first data source to use the first data source which has passed the second inspection as the data source;
- and generating the evaluation task for the data source.
A data labeling apparatus, comprising:
- one or more processors;
- and
- a storage device configured to store one or more programs, wherein
- the one or more programs, when executed by the one or more processors, cause the one or more processors to:
- sample a data source according to an evaluation task for the data source, to obtain sampled data;
- generate a labeling task from the sampled data;
- send the labeling task to a labeling device;
- and
- receive a labeled result of the labeling task from the labeling device,
- wherein the one or more programs, when executed by the one or more processors, cause the one or more processors further to:
- receive an inspection result and address information of a first data source which has passed a first inspection;
- and
- perform a second inspection on the first data source to use the first data source which has passed the second inspection as the data source, and generate the evaluation task for the data source.
A non-transitory computer-readable storage medium, comprising computer executable instructions stored thereon, wherein the executable instructions, when executed by a computer, cause the computer to:
- sample a data source according to an evaluation task for the data source, to obtain sampled data;
- generate a labeling task from the sampled data;
- send the labeling task to a labeling device;
- and
- receive a labeled result of the labeling task from the labeling device,
- wherein the executable instructions, when executed by a computer, further cause the computer to:
- receive an inspection result and address information of a first data source which has passed a first inspection;
- and
- perform a second inspection on the first data source to use the first data source which has passed the second inspection as the data source;
- and generate the evaluation task for the data source.
Dependencies are inferred from the claim text.