Zero-shot entity linking based on symbolic information
Methods, systems, and computer program products for zero-shot entity linking based on symbolic information are provided herein. A computer-implemented method includes obtaining a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph; training a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and mapping an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model.
1. A computer-implemented method, the method comprising:
obtaining a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph;
training a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and
mapping an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model;
wherein the method is carried out by at least one computing device.
2. The computer-implemented method of claim 1 , wherein the entity mention substring of the new sample of text is not seen during the training.
3. The computer-implemented method of claim 2 , wherein the mapping is performed without retraining the machine learning model.
4. The computer-implemented method of claim 1 , wherein the multi-task machine learning framework trains the machine learning model simultaneously on at least a primary task and an auxiliary task.
5. The computer-implemented method of claim 4 , wherein the primary task comprises an entity linking task and the auxiliary task comprises an entity type prediction task.
6. The computer-implemented method of claim 4 , wherein the training comprises:
generating predictions for the primary task and for secondary task; and
stitching the generated predictions for the primary task with the generated predictions of the secondary task to improve at least one performance metric corresponding to the primary task of the learning model.
7. The computer-implemented method of claim 6 , wherein the stitching comprises pruning a hierarchy of entity types at a specified level.
8. The computer-implemented method of claim 1 , wherein a given text sample in the training dataset comprises a mention of at least one entity in the subset of entities and text corresponding to a context of the mention.
9. The computer-implemented method of claim 1 , wherein the symbolic information comprises multiple entity types.
10. The computer-implemented method of claim 9 , wherein the multiple entity types are arranged in a logical hierarchy.
11. The computer-implemented method of claim 1 , wherein software is provided as a service in a cloud environment.
12. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing device to cause the computing device to:
obtain a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph;
train a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and
map an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model.
13. The computer program product of claim 12 , wherein the entity mention substring of the new sample of text is not seen during the training.
14. The computer program product of claim 13 , wherein the mapping is performed without retraining the machine learning model.
15. The computer program product of claim 12 , wherein the multi-task machine learning framework trains the machine learning model simultaneously on at least a primary task and an auxiliary task.
16. The computer program product of claim 15 , wherein the primary task comprises an entity linking task and the auxiliary task comprises an entity type prediction task.
17. The computer program product of claim 15 , wherein the training comprises:
generating predictions for the primary task and for secondary task; and
stitching the generated predictions for the primary task with the generated predictions of the secondary task to improve at least one performance metric corresponding to the primary task of the learning model.
18. The computer program product of claim 12 , wherein the symbolic information comprises multiple entity types.
19. The computer program product of claim 18 , wherein the multiple entity types are arranged in a logical hierarchy.
20. A system comprising:
a memory configured to store program instructions;
a processor operatively coupled to the memory to execute the program instructions to:
obtain a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph;
train a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and
map an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model.