Concepts
At a glance
The Concepts task extracts general DBPedia concepts that are directly referenced or alluded to (but not directly referenced) in the input text. For example, in the text IBM announces new advances in quantum computing.,
a concept extracted may be http://dbpedia.org/resource/IBM, http://dbpedia.org/resource/Quantum_computing and http://dbpedia.org/resource/Qubit.
| Class definition |
|---|
watson_nlp.workflows.concepts.concepts_workflow.ConceptsWorkflow |
The Concepts model was trained to assign concepts from DBPedia (2016 version)
Pretrained models
Model names are listed below. For language support, see Supported languages.
| Model ID | Container Image |
|---|---|
| concepts_alchemy-workflow_lang_de_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_de_stock:1.4.1 |
| concepts_alchemy-workflow_lang_en_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_en_stock:1.4.1 |
| concepts_alchemy-workflow_lang_es_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_es_stock:1.4.1 |
| concepts_alchemy-workflow_lang_fr_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_fr_stock:1.4.1 |
| concepts_alchemy-workflow_lang_it_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_it_stock:1.4.1 |
| concepts_alchemy-workflow_lang_ja_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_ja_stock:1.4.1 |
| concepts_alchemy-workflow_lang_ko_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_ko_stock:1.4.1 |
| concepts_alchemy-workflow_lang_pt_stock | cp.icr.io/cp/ai/watson-nlp_concepts_alchemy-workflow_lang_pt_stock:1.4.1 |
For details of the Concepts type system, see Understanding model type systems.
Running models
The Concepts model request accepts the following fields:
| Field | Type | Required Optional Repeated |
Description |
|---|---|---|---|
raw_document |
watson_core_data_model.nlp.RawDocument |
required | The input document on which to perform Concepts predictions |
limit |
int32 |
optional | The maximum number of concepts to return. Default is 50 |
Example requests
REST API
curl -s \
"http://localhost:8080/v1/watson.runtime.nlp.v1/NlpService/ConceptsPredict" \
-H "accept: application/json" \
-H "content-type: application/json" \
-H "Grpc-Metadata-mm-model-id: concepts_alchemy-workflow_lang_en_stock" \
-d '{ "raw_document": { "text": "Einstein postulated that light propagates through empty space with a definite speed 'c' independent of the speed of the observer (or source)." }, "limit": 2 }'
Response
{"concepts":[
{"text":"Special_relativity", "relevance":0.7799297, "dbpediaResource":"http://dbpedia.org/resource/Special_relativity"},
{"text":"Vacuum", "relevance":0.6394, "dbpediaResource":"http://dbpedia.org/resource/Vacuum"}
],
"producerId":{
"name":"Concepts Workflow",
"version":"0.0.1"
}
}
Python
import grpc
from watson_nlp_runtime_client import (
common_service_pb2,
common_service_pb2_grpc,
syntax_types_pb2,
)
channel = grpc.insecure_channel("localhost:8085")
stub = common_service_pb2_grpc.NlpServiceStub(channel)
request = common_service_pb2.ConceptsRequest(
raw_document=syntax_types_pb2.RawDocument(text="Einstein postulated that light propagates through empty space with a definite speed 'c' independent of the speed of the observer (or source)."),
limit=2
)
response = stub.ConceptsPredict(
request, metadata=[("mm-model-id", "concepts_alchemy-workflow_lang_en_stock")]
)
print(response)
Response
concepts {
text: "Special_relativity"
relevance: 0.78003943
dbpedia_resource: "http://dbpedia.org/resource/Special_relativity"
}
concepts {
text: "Vacuum"
relevance: 0.639851689
dbpedia_resource: "http://dbpedia.org/resource/Vacuum"
}
producer_id {
name: "Concepts Workflow"
version: "0.0.1"
}