Task UI layout representing semantical relations

Patent metadata

Field	Value
Publication number	US-10613841-B2
Application number	US-201815970200-A
Country	US
Kind code	B2
Filing date	May 3, 2018
Priority date	May 3, 2018
Publication date	Apr 7, 2020
Grant date	Apr 7, 2020

Abstract

A method and system including at least one data set including one or more properties in a task; a task UI module; a semantic grouping module including a neural network and a property cluster module; a display; and a processor in communication with the task UI module and the semantic grouping module and operative to execute processor-executable process steps to cause the system to: receive the data set at the semantic grouping module; calculate a property vector for each property in the data set, wherein the property vector includes a location of the property vector in a vector space; determine one or more property clusters, via the property cluster module, for all of the property vectors; and automatically generate a section in a user interface for each of the one or more property clusters via the task user interface module. Numerous other aspects are provided.

First claim

What is claimed is:

1. A system comprising:
at least one data set including one or more properties in a task;
a task UI module;
a semantic grouping module including a neural network and a property cluster module;
a display; and
a processor in communication with the task UI module and the semantic grouping module and operative to execute processor-executable process steps to cause the system to:
receive the data set at the semantic grouping module;
calculate a property vector for each property in the data set, wherein the property vector includes a location of the property vector in a vector space;
determine one or more property clusters, via the property cluster module, for all of the property vectors; and
automatically generate a section in a user interface for each of the one or more property clusters via the task user interface module.

2. The system of claim 1 , wherein the neural network is trained with existing user interface layouts.

3. The system of claim 2 , wherein training the neural network further comprises processor-executable process steps to cause the system to:
extract each property from the existing user interface layouts;
compare the extracted properties to properties of previously trained user interface layouts in the neural network; and
adjust the neural network based on the comparison to calculate locations for each extracted property as a property vector.

4. The system of claim 3 , wherein the location of each calculated property vector is relative to the other property vectors.

5. The system of claim 3 , wherein calculating the location of each property vector further comprises processor-executable steps to cause the system to:
determine a semantic similarity between the extracted properties; and
locate a position for the extracted property vectors having a high semantic similarity closer together in the vector space than extracted property vectors having a low semantic similarity.

6. The system of claim 5 , wherein determining two extracted properties have a high semantic similarity is based on processor-executable steps to cause the system to:
determine the two extracted properties have corresponding properties in the existing user interface layout that are positioned in a same section of the existing user interface layout.

7. The system of claim 3 , wherein determining a first and a second extracted property have a high semantic similarity is based on processor-executable steps to cause the system to:
determine a property corresponding to the first extracted property is positioned in the existing user interface layout near a property corresponding to a third extracted property and a property corresponding to the second extracted property in the existing user interface layout is positioned near the third extracted property.

8. The system of claim 1 , wherein automatically generating the section in the user interface for each of the one or more property clusters further comprises processor-executable steps to cause the system to:
receive, at a user interface renderer, the generated one or more property clusters; and
generate the user interface based on the received generated one or more property clusters, wherein the property vectors in a same property cluster are positioned on the user interface in a same section.

9. A computer-implemented method comprising:
training a neural network with existing user interface layouts for a first task;
receiving at least one data set at a semantic grouping module, the data set including one or more properties in a second task;
calculating a property vector for each property in the data set, wherein the property vector includes a location of the property vector in a vector space;
determining one or more property clusters for all of the positioned property vectors; and
automatically generating a section in a user interface for each of the one or more property clusters corresponding to one or more properties of the one or more property vectors.

10. The method of claim 9 , wherein the first task and the second task are the same.

11. The method of claim 9 , wherein training the neural network further comprises:
extracting each property from the existing user interface layouts;
comparing the extracted properties to properties of previously trained user interface layouts in the neural network; and
adjusting the neural network based on the comparison to calculate locations for each extracted property as a property vector.

12. The method of claim 11 , wherein the location of each calculated property vector is relative to the other property vectors.

13. The method of claim 11 , wherein calculating the location of each property vector further comprises:
determining a semantic similarity between the extracted properties; and
locating a position for the extracted property vectors having a high semantic similarity closer together in the vector space than extracted property vectors having a low semantic similarity.

14. The method of claim 13 , wherein determining two extracted properties have a high semantic similarity is based on determining the two extracted properties have corresponding properties in the existing user interface layout that are positioned in a same section of the existing user interface layout.

15. The method of claim 11 , wherein determining a first and a second extracted property have a high semantic similarity is based on determining a property corresponding to the first extracted property is positioned in the existing user interface layout near a property corresponding to a third extracted property, and a property corresponding to the second extracted property in the existing user interface layout is positioned near the third extracted property.

16. The method of claim 9 , wherein automatically generating the section in the user interface for each of the one or more property clusters further comprises:
receiving, at a user interface renderer, the generated one or more property clusters of property vectors; and
generating the user interface based on the received generated one or more property clusters of property vectors, wherein the property vectors in a same property cluster are positioned on the user interface in a same section.

17. A non-transitory computer-readable medium storing program code, the program code executable by a computer system to cause the computer system to:
receive the data set at a semantic grouping module;
calculate a property vector for each property in the data set, wherein the property vector includes a location of the property vector in a vector space;
determine one or more property clusters for all of the positioned property vectors via a property cluster module; and
automatically generate a section in a user interface for each of the one or more property clusters, via a task user interface module.

18. The medium of claim 17 , wherein the semantic grouping module includes a neural network that is trained with existing user interface layouts.

19. The medium of claim 18 , wherein training the neural network further comprises program code executable by the computer system to cause the system to:
extract each property from the existing user interface layouts;
compare the extracted properties to properties of previously trained user interface layouts in the neural network; and
adjust the neural network based on the comparison to calculate locations for each extracted property as a property vector.

20. The medium of claim 17 wherein automatically generating the section

Assignees

• Sap Se

Inventors

• Meyer Joachim
• Nicolay Fabian

Classifications

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06N20/00

  Machine learning · CPC title

• G06N3/08

  Learning methods · CPC title

• G06F8/38 Primary

  for implementing user interfaces · CPC title

• G06K9/6218

  Physics · mapped topic