Description
Zero-shot Named Entity Recognition (NER) enables the identification of entities in text with minimal effort. By leveraging pre-trained language models and contextual understanding, zero-shot NER extends entity recognition capabilities to new domains and languages. While the model card includes default labels as examples, it is important to highlight that users are not limited to these labels.
The model is designed to support any set of entity labels, allowing users to adapt it to their specific use cases. For best results, it is recommended to use labels that are conceptually similar to the provided defaults.
How to use
document_assembler = DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
sentence_detector = SentenceDetector()\
.setInputCols(["document"])\
.setOutputCol("sentence")
tokenizer = Tokenizer()\
.setInputCols(["sentence"])\
.setOutputCol("token")
labels = [ 'Admission_Discharge', 'Age', 'Alcohol', 'Body_Part', 'Clinical_Dept', 'Direction', 'Disease_Syndrome_Disorder', 'Dosage_Strength',
'Drug', 'Duration', 'Employment', 'Form', 'Frequency', 'Gender', 'Injury_or_Poisoning', 'Medical_Device', 'Modifier', 'Oncological', 'Procedure',
'Race_Ethnicity', 'Relationship_Status', 'Route', 'Section_Header', 'Smoking', 'Symptom', 'Test', 'Test_Result', 'Treatment', 'Vaccine']
pretrained_zero_shot_ner = PretrainedZeroShotNER().pretrained("zeroshot_ner_jsl_large", "en", "clinical/models")\
.setInputCols("sentence", "token")\
.setOutputCol("ner")\
.setPredictionThreshold(0.5)\
.setLabels(labels)
ner_converter = NerConverterInternal()\
.setInputCols("sentence", "token", "ner")\
.setOutputCol("ner_chunk")
pipeline = Pipeline().setStages([
document_assembler,
sentence_detector,
tokenizer,
pretrained_zero_shot_ner,
ner_converter
])
data = spark.createDataFrame([["""The patient is a 21-day-old Caucasian male here for 2 days of congestion - mom has been suctioning yellow discharge from the patient's nares, plus she has noticed some mild problems with his breathing while feeding (but negative for any perioral cyanosis or retractions). Additionally, there is no side effect observed after Influenza vaccine. One day ago, mom also noticed a tactile temperature and gave the patient Tylenol. Baby also has had some decreased p.o. intake. His normal breast-feeding is down from 20 minutes q.2h. to 5 to 10 minutes secondary to his respiratory congestion. He sleeps well, but has been more tired and has been fussy over the past 2 days. The parents noticed no improvement with albuterol treatments given in the ER. His urine output has also decreased; normally he has 8 to 10 wet and 5 dirty diapers per 24 hours, now he has down to 4 wet diapers per 24 hours. Mom denies any diarrhea. His bowel movements are yellow colored and soft in nature."""]]).toDF("text")
result = pipeline.fit(data).transform(data)
document_assembler = nlp.DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
sentence_detector = nlp.SentenceDetector()\
.setInputCols(["document"])\
.setOutputCol("sentence")
tokenizer = nlp.Tokenizer()\
.setInputCols(["sentence"])\
.setOutputCol("token")
labels = [ 'Admission_Discharge', 'Age', 'Alcohol', 'Body_Part', 'Clinical_Dept', 'Direction', 'Disease_Syndrome_Disorder', 'Dosage_Strength',
'Drug', 'Duration', 'Employment', 'Form', 'Frequency', 'Gender', 'Injury_or_Poisoning', 'Medical_Device', 'Modifier', 'Oncological', 'Procedure',
'Race_Ethnicity', 'Relationship_Status', 'Route', 'Section_Header', 'Smoking', 'Symptom', 'Test', 'Test_Result', 'Treatment', 'Vaccine']
pretrained_zero_shot_ner = medical.PretrainedZeroShotNER().pretrained("zeroshot_ner_jsl_large", "en", "clinical/models")\
.setInputCols("sentence", "token")\
.setOutputCol("ner")\
.setPredictionThreshold(0.5)\
.setLabels(labels)
ner_converter = medical.NerConverterInternal()\
.setInputCols("sentence", "token", "ner")\
.setOutputCol("ner_chunk")
pipeline = nlp.Pipeline().setStages([
document_assembler,
sentence_detector,
tokenizer,
pretrained_zero_shot_ner,
ner_converter
])
data = spark.createDataFrame([["""The patient is a 21-day-old Caucasian male here for 2 days of congestion - mom has been suctioning yellow discharge from the patient's nares, plus she has noticed some mild problems with his breathing while feeding (but negative for any perioral cyanosis or retractions). Additionally, there is no side effect observed after Influenza vaccine. One day ago, mom also noticed a tactile temperature and gave the patient Tylenol. Baby also has had some decreased p.o. intake. His normal breast-feeding is down from 20 minutes q.2h. to 5 to 10 minutes secondary to his respiratory congestion. He sleeps well, but has been more tired and has been fussy over the past 2 days. The parents noticed no improvement with albuterol treatments given in the ER. His urine output has also decreased; normally he has 8 to 10 wet and 5 dirty diapers per 24 hours, now he has down to 4 wet diapers per 24 hours. Mom denies any diarrhea. His bowel movements are yellow colored and soft in nature."""]]).toDF("text")
result = pipeline.fit(data).transform(data)
val document_assembler = new DocumentAssembler()
.setInputCol("text")
.setOutputCol("document")
val sentence_detector = new SentenceDetector()
.setInputCols("document")
.setOutputCol("sentence")
val tokenizer = new Tokenizer()
.setInputCols("sentence")
.setOutputCol("token")
labels = Array("Admission_Discharge", "Age", "Alcohol", "Body_Part", "Clinical_Dept", "Direction", "Disease_Syndrome_Disorder", "Dosage_Strength",
"Drug", "Duration", "Employment", "Form", "Frequency", "Gender", "Injury_or_Poisoning", "Medical_Device", "Modifier", "Oncological", "Procedure",
"Race_Ethnicity", "Relationship_Status", "Route", "Section_Header", "Smoking", "Symptom", "Test", "Test_Result", "Treatment", "Vaccine")
val pretrained_zero_shot_ner = PretrainedZeroShotNER().pretrained("zeroshot_ner_jsl_large", "en", "clinical/models")
.setInputCols(Array("sentence", "token"))
.setOutputCol("ner")
.setPredictionThreshold(0.5)
.setLabels(labels)
val ner_converter = new NerConverterInternal()
.setInputCols(Array("sentence", "token", "ner"))
.setOutputCol("ner_chunk")
val pipeline = new Pipeline().setStages(Array(
document_assembler,
sentence_detector,
tokenizer,
pretrained_zero_shot_ner,
ner_converter
))
val data = Seq([["""The patient is a 21-day-old Caucasian male here for 2 days of congestion - mom has been suctioning yellow discharge from the patient's nares, plus she has noticed some mild problems with his breathing while feeding (but negative for any perioral cyanosis or retractions). Additionally, there is no side effect observed after Influenza vaccine. One day ago, mom also noticed a tactile temperature and gave the patient Tylenol. Baby also has had some decreased p.o. intake. His normal breast-feeding is down from 20 minutes q.2h. to 5 to 10 minutes secondary to his respiratory congestion. He sleeps well, but has been more tired and has been fussy over the past 2 days. The parents noticed no improvement with albuterol treatments given in the ER. His urine output has also decreased; normally he has 8 to 10 wet and 5 dirty diapers per 24 hours, now he has down to 4 wet diapers per 24 hours. Mom denies any diarrhea. His bowel movements are yellow colored and soft in nature."""]]).toDF("text")
val result = pipeline.fit(data).transform(data)
Results
+----------------------+-----+---+--------------+----------+
|chunk |begin|end|ner_label |confidence|
+----------------------+-----+---+--------------+----------+
|21-day-old |18 |27 |Age |0.99855894|
|Caucasian |29 |37 |Race_Ethnicity|0.9965018 |
|male |39 |42 |Gender |0.9979837 |
|for 2 days |49 |58 |Frequency |0.9547805 |
|congestion |63 |72 |Symptom |0.9777157 |
|mom |76 |78 |Gender |0.99721634|
|nares |136 |140|Body_Part |0.8162966 |
|she |148 |150|Gender |0.99808997|
|mild |169 |172|Modifier |0.9933716 |
|his |188 |190|Gender |0.9682989 |
|retractions |259 |269|Symptom |0.93004584|
|Influenza vaccine |326 |342|Vaccine |0.9917237 |
|mom |358 |360|Gender |0.99286497|
|Tylenol |418 |424|Drug |0.7294259 |
|Baby |427 |430|Age |0.95155627|
|His |473 |475|Gender |0.99755836|
|5 to 10 minutes |532 |546|Frequency |0.5138932 |
|his |561 |563|Gender |0.99703395|
|respiratory congestion|565 |586|Symptom |0.78011936|
|He |589 |590|Gender |0.997498 |
|tired |623 |627|Symptom |0.8767565 |
|fussy |642 |646|Symptom |0.95738137|
|albuterol treatments |710 |729|Treatment |0.53156906|
|ER |744 |745|Clinical_Dept |0.9652154 |
|His |748 |750|Gender |0.9987986 |
|urine output |752 |763|Symptom |0.5371252 |
|he |794 |795|Gender |0.99845266|
|per 24 hours |833 |844|Frequency |0.8483188 |
|he |851 |852|Gender |0.99878913|
|per 24 hours |880 |891|Frequency |0.7822129 |
|Mom |894 |896|Gender |0.98790914|
|diarrhea |909 |916|Symptom |0.96422964|
|His |919 |921|Gender |0.9979843 |
+----------------------+-----+---+--------------+----------+
Model Information
| Model Name: | zeroshot_ner_jsl_large |
| Compatibility: | Healthcare NLP 5.5.1+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 1.6 GB |
Benchmarking
label precision recall f1-score support
Admission_Discharge 0.6976 0.9814 0.8155 322
Age 0.8040 0.9495 0.8707 812
Alcohol 0.7731 0.9684 0.8598 95
Body_Part 0.7486 0.8022 0.7744 10493
Clinical_Dept 0.7665 0.9536 0.8499 1704
Direction 0.7585 0.9451 0.8416 4316
Disease_Syndrome_Disorder 0.8701 0.6791 0.7628 5267
Dosage_Strength 0.7500 0.8009 0.7746 1266
Drug 0.8876 0.8645 0.8759 2723
Duration 0.4256 0.9129 0.5805 918
Employment 0.6494 0.8693 0.7434 375
Form 0.4963 0.7683 0.6030 259
Frequency 0.5561 0.7341 0.6328 1019
Gender 0.9779 0.9920 0.9849 5612
Injury_or_Poisoning 0.5710 0.7056 0.6312 992
Medical_Device 0.8075 0.9034 0.8528 5610
Modifier 0.4157 0.9181 0.5723 2929
Oncological 0.7931 0.8135 0.8032 740
Procedure 0.7274 0.7884 0.7567 6509
Race_Ethnicity 0.9916 1.0000 0.9958 118
Relationship_Status 0.4623 0.9608 0.6242 51
Route 0.7375 0.7963 0.7658 967
Section_Header 0.9047 0.9774 0.9396 10262
Smoking 0.9720 0.9630 0.9674 108
Symptom 0.7663 0.7455 0.7557 11590
Test 0.7322 0.6542 0.6910 6117
Test_Result 0.3766 0.7947 0.5110 1398
Treatment 0.3548 0.6404 0.4566 456
Vaccine 0.6522 0.7143 0.6818 21
accuracy - - 0.8770 242063
macro avg 0.7129 0.8499 0.7635 242063
weighted avg 0.8955 0.8770 0.8826 242063