Extraction of genealogy data from obituaries
Systems, methods, and other techniques for extracting data from obituaries are provided. In some embodiments, an obituary containing a plurality of words is received. Using a machine learning model, an entity tag from a set of entity tags may be assigned to each of one or more words of the plurality of words. Each particular tag from the set of entity tags may include a relationship component and a category component. The relationship component may indicate a relationship between a particular word and the deceased individual. The category component may indicate a categorization of the particular word to a particular category from a set of categories. The extracted data may be stored in a genealogical database.
1. A computer-implemented method for extracting data from articles, the computer-implemented method comprising:
determining that an image contains an article comprising a plurality of words and corresponding to an individual;
assigning, using an entity-tagging machine learning (ML) model, an entity tag from a set of entity tags to one or more words of the plurality of words, the one or more words defining an entity;
wherein each entity tag from the set of entity tags comprises a relationship component and a category component;
wherein the relationship component indicates a relationship between a particular word of the plurality of words to which the entity tag is assigned and the individual;
wherein the category component indicates a categorization of the particular word of the plurality of words to a particular category from a set of categories;
assigning, using a gender-prediction ML model, a gender tag to each word of the plurality of words having a category component of the entity tag corresponding to a person category; and
generating synthesized data based on the entity tag;
wherein generating synthesized data comprises performing inferences for missing information based on the image.
2. The computer-implemented method of claim 1 , further comprising storing the synthesized data in a genealogical database.
3. The computer-implemented method of claim 1 , further comprising:
assigning, using a name-assignment ML model, a name-part tag to a word of the plurality of words having a category component of the entity tag corresponding to a person category, the name-part tag identifying a type of name of the word.
4. The computer-implemented method of claim 1 , further comprising:
determining that the article is an obituary and the individual is a deceased person; and
selecting the set of entity tags from a plurality of sets of entity tags, the set of entity tags corresponding to obituaries.
5. The computer-implemented method of claim 1 , further comprising:
determining that the article is a wedding announcement; and
selecting the set of entity tags from a plurality of sets of entity tags, the set of entity tags corresponding to wedding announcements.
6. The computer-implemented method of claim 1 , further comprising:
generating the synthesized data based on the entity tag, the gender tag, and a name-part tag.
7. A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
determining that an image contains an article containing a plurality of words and corresponding to an individual;
assigning, using an entity-tagging machine learning (ML) model, an entity tag from a set of entity tags to one or more words of the plurality of words, the one or more words defining an entity;
wherein each entity tag from the set of entity tags comprises a relationship component and a category component;
wherein the relationship component indicates a relationship between a particular word of the plurality of words to which the entity tag is assigned and the individual; and
wherein the category component indicates a categorization of the particular word of the plurality of words to a particular category from a set of categories;
assigning, using a gender-prediction ML model, a gender tag to each word of the plurality of words having a category component of the entity tag corresponding to a person category; and
generating synthesized data based on the entity tag;
wherein generating synthesized data comprises performing inferences for missing information based on the image.
8. The non-transitory computer-readable medium of claim 7 , wherein the operations further comprise:
assigning, using a name-assignment ML model, a name-part tag to a word of the plurality of words having a category component of the entity tag corresponding to a person category, the name-part tag identifying a type of name of the word.
9. The non-transitory computer-readable medium of claim 8 , wherein the operations further comprise:
generating the synthesized data based on the entity tag, the gender tag, and the name-part tag; and
storing the synthesized data in a genealogical database.
10. The non-transitory computer-readable medium of claim 9 , wherein the synthesized data are stored on the genealogical database with the image.
11. The non-transitory computer-readable medium of claim 7 , wherein the operations further comprise:
determining that the article is an obituary and the individual is a deceased person; and
selecting the set of entity tags from a plurality of sets of entity tags, the set of entity tags corresponding to obituaries.
12. The non-transitory computer-readable medium of claim 7 , wherein the operations further comprise:
determining that the article is a wedding announcement; and
selecting the set of entity tags from a plurality of sets of entity tags, the set of entity tags corresponding to wedding announcements.
13. The non-transitory computer-readable medium of claim 7 , wherein determining the relationship component comprises:
classifying, using a relationship-classification ML model, a relationship between a pair of entities identified from the article.
14. The non-transitory computer-readable medium of claim 13 , wherein the relationship between the pair of entities is selected from a set of relationships.
15. The non-transitory computer-readable medium of claim 13 , wherein the relationship-classification ML model is a convolutional neural network or a recurrent neural network.
16. A system for extracting data from articles comprising:
one or more processors; and
a non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
determining that an image contains an article containing a plurality of words and corresponding to an individual;
assigning, using an entity-tagging machine learning (ML) model, an entity tag from a set of entity tags to one or more words of the plurality of words, the one or more words defining an entity;
wherein each entity tag from the set of entity tags comprises a relationship component and a category component;
wherein the relationship component indicates a relationship between a particular word of the plurality of words to which the entity tag is assigned and the individual; and
wherein the category component indicates a categorization of the particular word of the plurality of words to a particular category from a set of categories;
assigning, using a gender-prediction ML model, a gender tag to each word of the plurality of words having a category component of the entity tag corresponding to a person category;
assigning, using a name-assignment ML model, a name-part tag to a word of the plurality of words having a category component of the entity tag corresponding to a person category, the name-part tag identifying a type of name of the word; and
generating synthesized data based on the entity tag;
wherein generating synthesized data comprises performing inferences for missing information based on the image.
17. The system of claim 16 , wherein the operations further comprise:
generating synthesized data based on the entity tag;
wherein generating synthesized data comprises performing inferences for missing information based on the image.
18. The system of claim 16 , wherein determining the relationship component comprises:
classifying, using a relationship-classification ML model, a relationship between a pair of entities identified from the article.
19. The system of claim 18 , wherein the relationship between the pair of entities is selected from a set of relationships.
20. The system of claim 18 , wherein the relationship-classification ML model is a convolutional neural network or a recurrent neural network.