Iteratively expanding concepts

1. A computer-implemented method comprising:

determining a set of target concepts;

parsing an ontology to build a cache of a set of expressions for each target concept of the set of target concepts, wherein each expression from the set of expressions maps a paired particular concept and particular modifier to a specific medical concept;

receiving a document comprising a text segment that includes unstructured data;

performing, by a concept matching engine comprising a neural network, a first traversal of the text segment of the document to identify first surface forms, the first surface forms comprising a target object term and a verb term that modifies the target object term, wherein the neural network comprises a plurality of resistive switching devices which each encode a weight of an element of a natural language process (NLP) model of the neural network;

iteratively performing, by the concept matching engine, one or more additional traversals of the text segment to identify one or more second surface forms that modify the identified first surface forms, wherein at least one of the one or more second surface forms comprises a body part term that modifies the target object term of the verb term, and wherein the body part term is absent from the first surface form;

identifying a first target concept using the set of expressions based upon the target object term and the verb term;

identifying at least one second target concept using the set of expressions based upon the target object term and the body part term; and

annotating the text segment of the document with an annotation by providing to the neural network the first target concept as the particular concept and the body part term as the particular modifier to search the cache to identify the specific medical concept, wherein the neural network is trained on structured text, unstructured text, and annotations, and wherein the annotation to the text segment comprises the identified specific medical concept.

2. The computer-implemented method of claim 1 , wherein performing the one or more additional traversals comprises analyzing the document via a natural language processing technique.

3. The computer-implemented method of claim 1 , further comprising mapping the set of target concepts to expressions found in the ontology.

4. The computer-implemented method of claim 1 , wherein each respective surface form of the first surface forms is a synonym, antonym, or variant of a target concept.

5. The computer-implemented method of claim 1 , wherein the ontology comprises a dictionary.

6. The computer-implemented method of claim 1 , wherein the document comprises unstructured text.

7. A system comprising:

a memory having computer readable instructions; and

one or more processors for executing the computer readable instructions, the computer readable instructions controlling the one or more processors to perform operations comprising:

determining a set of target concepts;

parsing an ontology to build a cache of a set of expressions for each target concept of the set of target concepts, wherein each expression from the set of expressions maps a paired particular concept and particular modifier to a specific medical concept;

receiving a document comprising a text segment that includes unstructured data;

performing, by a concept matching engine comprising a neural network, a first traversal of the text segment of the document to identify first surface forms, the first surface forms comprising a target object term and a verb term that modifies the target object term, wherein the neural network comprises a plurality of resistive switching devices which each encode a weight of an element of a natural language process (NLP) model of the neural network;

iteratively performing, by the concept matching engine, one or more additional traversals of the text segment to identify one or more second surface forms that modify the identified first surface forms, wherein at least one of the one or more second surface forms comprises a body part term that modifies the target object term of the verb term, and wherein the body part term is absent from the first surface form;

identifying a first target concept using the set of expressions based upon the target object term and the verb term;

identifying at least one second target concept using the set of expressions based upon the target object term and the body part term; and

annotating the text segment of the document with an annotation by providing to the neural network the first target concept as the particular concept and the body part term as the particular modifier to search the cache to identify the specific medical concept, wherein the neural network is trained on structured text, unstructured text, and annotations, and wherein the annotation to the text segment comprises the identified specific medical concept.

8. The system of claim 7 , wherein performing the one or more additional traversals comprises analyzing the document via a natural language processing technique.

9. The system of claim 7 , further comprising mapping the set of target concepts to expressions found in the ontology.

10. The system of claim 7 , wherein each respective surface form of the first surface forms is a synonym, antonym, or variant of a target concept.

11. The system of claim 7 , wherein the ontology comprises a dictionary.

12. The system of claim 7 , wherein the document comprises unstructured text.

13. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform operations comprising:

determining a set of target concepts;

parsing an ontology to build a cache of a set of expressions for each target concept of the set of target concepts, wherein each expression from the set of expressions maps a paired particular concept and particular modifier to a specific medical concept;

receiving a document comprising a text segment that includes unstructured data;

performing, by a concept matching engine comprising a neural network, a first traversal of the text segment of the document to identify first surface forms, the first surface forms comprising a target object term and a verb term that modifies the target object term, wherein the neural network comprises a plurality of resistive switching devices which each encode a weight of an element of a natural language process (NLP) model of the neural network;

iteratively performing, by the concept matching engine, one or more additional traversals of the text segment to identify one or more second surface forms that modify the identified first surface forms, wherein at least one of the one or more second surface forms comprises a body part term that modifies the target object term of the verb term, and wherein the body part term is absent from the first surface form;

identifying a first target concept using the set of expressions based upon the target object term and the verb term;

identifying at least one second target concept using the set of expressions based upon the target object term and the body part term; and

annotating the text segment of the document with an annotation by providing to the neural network the first target concept as the particular concept and the body part term as the particular modifier to search the cache to identify the specific medical concept, wherein the neural network is trained on structured text, unstructured text, and annotations, and wherein the annotation to the text segment comprises the identified specific medical concept.

14. The computer program product of claim 13 , wherein performing the one or more additional traversals comprises analyzing the document via a natural language processing technique.

15. The computer program product of claim 13 , further comprising mapping the set of target concepts to expressions found in the ontology.

16. The computer program product of claim 13 , wherein each respective surface form of the first surface forms is a synonym, antonym, or variant of a target concept.

17. The computer program product of claim 13 , wherein the ontology comprises a dictionary.