Identification of table partitions in documents with neural networks using global document context

1. A method, comprising:

obtaining a plurality of symbol sequences of a document, the document having one or more tables, wherein each of the plurality of symbol sequences of the document comprises at least one i) a table graphics element or ii) an alphanumeric sequence that includes one or more alphanumeric characters;

determining, by a processing device, a plurality of vectors comprising:

a first subset of one or more vectors, wherein each vector of the first subset is representative of at least one alphanumerical sequence, and

a second subset of one or more vectors, wherein each vector of the second subset is representative of at least one table graphics element;

generating, using one or more neural networks, a plurality of association hypotheses, each association hypothesis associating at least one vector from the first subset of one or more vectors with at least one vector from the second subset of one or more vectors; and

evaluating the plurality of association hypotheses to determine an association between a first alphanumeric sequence and a table partition of the one or more tables of the document.

2. The method of claim 1 , wherein generating the plurality of association hypotheses comprises:

obtaining, using the one or more neural networks, a plurality of recalculated vectors, wherein each of the plurality of recalculated vectors is based on at least a subset of the plurality of vectors; and

generating the plurality of association hypotheses using the plurality of recalculated vectors.

3. The method of claim 2 , wherein generating the plurality of association hypotheses using the plurality of recalculated vectors further comprises:

associating one of the plurality of recalculated vectors with the first alphanumeric sequence;

associating another one of the plurality of recalculated vectors with the table partition; and

determining that a first association hypothesis of the plurality of association hypotheses has a highest probability of occurrence, wherein the first association hypothesis comprises an association of a first recalculated vector of the plurality of recalculated vectors with a second recalculated vector of the plurality of recalculated vectors.

4. The method of claim 1 , wherein determining the plurality of vectors comprises:

determining, using the one or more neural networks, a word embedding for each one of the plurality of symbol sequences.

5. The method of claim 1 , further comprising:

determining, based on the association between the first alphanumeric sequence and the table partition, a location of the table partition; and

storing the location of the table partition.

6. The method of claim 5 , further comprising:

obtaining a subsequent document;

obtaining a symbol sequence of the subsequent document;

determining that the symbol sequence of the subsequent document has a location in the subsequent document that coincides, within a predetermined accuracy, with the location of the table partition; and

associating the symbol sequence of the subsequent document with the table partition.

7. The method of claim 1 , wherein the table graphics element is one of a horizontal line, a vertical line, a corner, a three-way line intersection, or a four-way line intersection.

8. The method of claim 1 , wherein obtaining the plurality of symbol sequences of the document comprises:

performing optical character recognition (OCR) of an image of the document to obtain an OCR text of the document and a plurality of table graphics elements; and

partitioning the OCR text and the plurality of table graphics elements into the plurality of symbol sequences.

9. A system comprising:

a memory; and

a processing device operatively coupled to the memory, the processing device configured to:

obtain a plurality of symbol sequences of a document, the document having one or more tables, wherein each of the plurality of symbol sequences of the document comprises at least one i) a table graphics element or ii) an alphanumeric sequence that includes one or more alphanumeric characters;

determine a plurality of vectors comprising:

a first subset of one or more vectors, wherein each vector of the first subset is representative of at least one alphanumerical sequence, and

a second subset of one or more vectors, wherein each vector of the second subset is representative of at least one table graphics element;

generate, using one or more neural networks, a plurality of association hypotheses, each association hypothesis associating at least one vector from the first subset of one or more vectors with at least one vector from the second subset of one or more vectors; and

evaluate the plurality of association hypotheses to determine an association between a first alphanumeric sequence and a table partition of the one or more tables of the document.

10. The system of claim 9 ; wherein to generate the plurality of association hypotheses, the processing device is configured to:

obtain, using the one or more neural networks, a plurality of recalculated vectors, wherein each of the plurality of recalculated vectors is based on at least a subset of the plurality of vectors; and

generate the plurality of association hypotheses using the plurality of recalculated vectors.

11. The system of claim 10 , wherein to generate the plurality of association hypotheses using the plurality of recalculated vectors, the processing device is further configured to:

associate one of the plurality of recalculated vectors with the first alphanumeric sequence;

associate another one of the plurality of recalculated vectors with the table partition; and

determine that a first association hypothesis of the plurality of association hypotheses has a highest probability of occurrence, wherein the first association hypothesis comprises an association of a first recalculated vector of the plurality of recalculated vectors with a second recalculated vector of the plurality of recalculated vectors.

12. The system of claim 9 , wherein to determine the plurality of vectors, the processing device is configured to:

determine, using the one or more neural networks, a word embedding for each one of the plurality of symbol sequences.

13. The system of claim 9 , wherein the processing device is further configured to:

determine, based on the association between the first alphanumeric sequence and the table partition, a location of the table partition; and

store the location of the table partition.

14. The system of claim 13 , wherein the processing device is further configured to:

obtain a subsequent document;

obtain a symbol sequence of the subsequent document;

determine that the symbol sequence of the subsequent document has a location in the subsequent document that coincides, within a predetermined accuracy, with the location of the table partition; and

associate the symbol sequence of the subsequent document with the table partition.

15. The system of claim 9 , wherein the table graphics element is one of a horizontal line, a vertical line, a corner, a three-way line intersection, or a four-way line intersection.

16. The system of claim 9 , wherein to obtain the plurality of symbol sequences of the document, the processing device is configured to:

performing optical character recognition (OCR) of an image of the document to obtain an OCR text of the document and a plurality of table graphics elements; and

partition the OCR text and the plurality of table graphics elements into the plurality of symbol sequences.

17. A non-transitory machine-readable storage medium including instructions that, when accessed by a processing device, cause the processing device to:

obtain a plurality of symbol sequences of a document, the document having one or more tables, wherein each of the plurality of symbol sequences of the document comprises at least one i) a table graphics element or ii) an alphanumeric sequence that includes one or more alphanumeric characters;

determine a plurality of vectors comprising:

a first subset of one or more vectors, wherein each vector of the first subset is representative of at least one alphanumerical sequence, and

a second subset of one or more vectors, wherein each vector of the second subset is representative of at least one table graphics element;

generate, using: one or more neural networks, a plurality of association hypotheses, each association hypothesis associating at least one vector from the first subset of one or more vectors with at least one vector from the second subset of one or more vectors; and

evaluate the plurality of association hypotheses to determine an association between a first alphanumeric sequence and a table partition of the one or more tables of the document.

18. The non-transitory machine-readable storage medium of claim 17 , wherein to cause the processing device to generate the plurality of association hypotheses, the instructions are to cause the processing device to:

obtain, using the one or more neural networks, a plurality of recalculated vectors, wherein each of the plurality of recalculated vectors is based on the plurality of vectors; and

generate the plurality of association hypotheses using the plurality of recalculated vectors.

19. The non-transitory machine-readable storage medium of claim 18 , wherein to cause the processing device to generate the plurality of association hypotheses using the plurality of recalculated vectors, the instructions are further to cause the processing device to:

associate one of the plurality of recalculated vectors with the first alphanumeric sequence;

associate another one of the plurality of recalculated vectors with the table partition; and

determine that a first association hypothesis of the plurality of association hypotheses has a highest probability of occurrence, wherein the first association hypothesis comprises an association of a first recalculated vector of the plurality of recalculated vectors with a second recalculated vector of the plurality of recalculated vectors.

20. The non-transitory machine-readable storage medium of claim 17 , wherein the instructions are further to cause the processing device to:

determine, based on the association between the first alphanumeric sequence and the table partition, a location of the table partition;

store the location of the table partition;

obtain a subsequent document;

obtain a symbol sequence of the subsequent document;

determine that the symbol sequence of the subsequent document has a location in the subsequent document that coincides, within a predetermined accuracy, with the location of the table partition; and

associate the symbol sequence of the subsequent document with the table partition.