Automatic formulation of data science problem statements

Patent metadata

Field	Value
Publication number	US-11763084-B2
Application number	US-202016989882-A
Country	US
Kind code	B2
Filing date	Aug 10, 2020
Priority date	Aug 10, 2020
Publication date	Sep 19, 2023
Grant date	Sep 19, 2023

Abstract

A method comprises receiving a new data set; identifying at least one prior data set of a plurality of prior data sets that matches the new data set; generating a natural language data science problem statement for the new data set based on information associated with the at least prior one data set that matches the new data set; outputting the generated natural language data science problem statement for user verification; and in response to receiving user input verifying the natural language generated data science problem statement, generating one or more AutoAI configuration settings for the new data set based on one or more AutoAI configuration settings associated with the at least one prior data set that matches the new data set.

First claim

What is claimed is:

1. A method comprising:
receiving a new data set;
identifying at least one prior data set of a plurality of prior data sets that matches the new data set;
generating a natural language data science problem statement for the new data set based on information associated with the at least one prior data set that matches the new data set, wherein the natural language data science problem statement poses a question that is grounded in the information associated with the at least one prior data set that matches the new data set;
outputting the natural language data science problem statement to obtain user verification that the question posed by the natural language data science problem statement is applicable to the new data set;
receiving user input that verifies the natural language data science problem statement; and
generating one or more Automated Artificial Intelligence (AutoAI) configuration settings for the new data set based on one or more AutoAI configuration settings associated with the at least one prior data set that matches the new data set.

2. The method of claim 1 , wherein identifying the at least one prior data set is based on comparison of labels in the at least one prior data set and labels in the new data set.

3. The method of claim 1 , wherein identifying the at least one prior data set is based on comparison of data values in the at least one prior data set and data values in the new data set.

4. The method of claim 1 , wherein identifying the at least one prior data set comprises identifying a single prior data set with a closest match to the new data set.

5. The method of claim 1 , wherein identifying the at least one prior data set comprises identifying a plurality of prior data sets that match the new data set;
wherein generating the natural language data science problem statement for the new data set comprises generating a plurality of natural language data science problem statements, each of the plurality of natural language data science problem statements corresponding to one of the plurality of prior data sets that match the new data set; and
wherein outputting the natural language data science problem statement for user verification comprises outputting the plurality of natural language data science problem statements for user selection of one of the plurality of natural language data science problem statements.

6. The method of claim 1 , further comprising receiving user feedback regarding the natural language data science problem statement; and
updating a machine learning algorithm used in generating the natural language data science problem statement based on the user feedback.

7. The method of claim 1 , wherein generating the natural language data science problem statement comprises generating the natural language data science problem statement based on labels, data values and the one or more AutoAI configuration settings for the at least one prior data set that matches the new data set.

8. A system comprising:
an interface;
a memory; and
a processor communicatively coupled to the interface and to the memory, wherein the processor is configured to:
receive a new data set via the interface;
identify at least one prior data set of a plurality of prior data sets that matches the new data set;
generate a natural language data science problem statement for the new data set based on information associated with the at least one prior data set that matches the new data set, wherein the natural language data science problem statement poses a question that is grounded in the information associated with the at least one prior data set that matches the new data set;
output the natural language data science problem statement via the interface to obtain user verification that the question posed by the natural language data science problem statement is applicable to the new data set;
receive user input verifying the natural language data science problem statement; and
generate one or more Automated Artificial Intelligence (AutoAI) configuration settings for the new data set based on one or more AutoAI configuration settings associated with the at least one prior data set that matches the new data set.

9. The system of claim 8 , wherein the processor is configured to identify the at least one prior data set based on comparison of labels in the at least one prior data set and labels in the new data set.

10. The system of claim 8 , wherein the processor is configured to identify the at least one prior data set based on comparison of data values in the at least one prior data set and data values in the new data set.

11. The system of claim 8 , wherein the processor is configured to identify a single prior data set of with a closest match to the new data set.

12. The system of claim 8 , wherein the processor is configured to identify a plurality of prior data sets that match the new data set;
wherein the processor is configured to generate a plurality of natural language data science problem statements, each of the plurality of natural language data science problem statements corresponding to one of the plurality of prior data sets that match the new data set; and
wherein the processor is configured to output the plurality of natural language data science problem statements for user selection of one of the plurality of natural language data science problem statements.

13. The system of claim 8 , wherein the processor is further configured to:
receive user feedback regarding the natural language data science problem statement; and
update a machine learning algorithm used in generating the natural language data science problem statement based on the user feedback.

14. The system of claim 8 , wherein the processor is configured to generate the natural language data science problem statement based on labels, data values and the one or more AutoAI configuration settings for the at least one prior data set that matches the new data set.

15. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed by a processor, causes the processor to:
receive a new data set;
identify at least one prior data set of a plurality of prior data sets that matches the new data set;
generate a natural language data science problem statement for the new data set based on information associated with the at least one prior data set that matches the new data set, wherein the natural language data science problem statement poses a question that is grounded in the information associated with the at least one prior data set that matches the new data set;
output the natural language data science problem statement to obtain user verification that the question posed by the natural language data science problem statement is applicable to the new data set;
receive user input verifying the natural language data science problem statement; and
generate one or more Automated Artificial Intelligence (AutoAI) configuration settings for the new data set based on one or more AutoAI configuration settings associated with the at least one prior data set that matches the new data set.

16. The computer program product of claim 15 , wherein the computer readable program is further configured to cause the processor to identify the at least one prior data set based on comparison of labels in the at least one prior data set and labels in the new data set.

17. The computer program product of claim 15 , wherein the computer readable program is further conf

Assignees

• IBM

Inventors

• Wang Dakuo
• Chaudhary Arunima
• Gan Chuang
• Yu Mo
• Pan Qian
• Liu Sijia
• Weidele Daniel Karl I
• Valente Abel

Classifications

• G06F40/30 Primary

  Semantic analysis · CPC title

• G06F40/289 Primary

  Phrasal analysis, e.g. finite state techniques or chunking · CPC title

• G06F40/56

  Natural language generation · CPC title

• G06F40/279

  Recognition of textual entities · CPC title

• G06N20/00

  Machine learning · CPC title