US 12,282,855 B2
Auto-formatting of a data table
Haoyu Dong, Beijing (CN); Zhouyu Fu, Redmond, WA (US); Shi Han, Beijing (CN); and Dongmei Zhang, Beijing (CN)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Appl. No. 17/291,818
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
PCT Filed Dec. 6, 2019, PCT No. PCT/US2019/064800
§ 371(c)(1), (2) Date May 6, 2021,
PCT Pub. No. WO2020/131409, PCT Pub. Date Jun. 25, 2020.
Claims priority of application No. 201811575065.3 (CN), filed on Dec. 21, 2018.
Prior Publication US 2022/0027740 A1, Jan. 27, 2022
Int. Cl. G06N 3/088 (2023.01); G06F 18/214 (2023.01); G06N 3/045 (2023.01); G06V 10/94 (2022.01); G06V 30/412 (2022.01)
CPC G06N 3/088 (2013.01) [G06F 18/214 (2023.01); G06N 3/045 (2023.01); G06V 10/95 (2022.01); G06V 30/412 (2022.01)] 15 Claims
OG exemplary drawing
 
1. A computer-implemented method of automatically reformatting a data table for improved visual presentation of the data table, comprising:
using one or more processors:
obtaining values of at least one attribute for a plurality of cells in the data table, the values of the at least one attribute indicating at least one of a semantic meaning of data filled in the plurality of cells or a structure of the data table, the plurality of cells being arranged in rows and columns in the data table;
extracting a feature representation of the values of the at least one attribute using a first learning network by:
mapping the values of the at least one attribute to a corresponding vectorized representation;
extracting a global feature representation from the vectorized representation;
extracting, within the first learning network and from the global feature representation, a row-wise or column-wise feature representation, the row-wise or column-wise feature representation capturing specific distribution patterns in row or column directions; and
obtaining a hybrid feature representation based upon a concatenation of the row-wise or column-wise feature representation to the global feature representation;
determining respective formats for the plurality of cells based on the hybrid feature representation using at least the first learning network;
adjusting the respective formats based upon a plurality of different target control indications to create a plurality of formats;
previewing, to a user, one or more of the plurality of formats in a graphical user interface (GUI) by displaying the data table adjusted according to one or more of the plurality of formats, the GUI allowing the user to select one of the plurality of formats;
receiving a selection of one of the plurality of formats; and
causing the data table to be displayed to the user in the selected format.