Legal Multilabel Classifier on Covid-19 Exceptions (Italian)

Description

This is the Multi-Label Text Classification model that can be used to identify up to 5 classes to facilitate analysis, discovery, and comparison of legal texts in Italian related to COVID-19 exception measures. The classes are as follows:

  • Closures/lockdown
  • Government_oversight
  • Restrictions_of_daily_liberties
  • Restrictions_of_fundamental_rights_and_civil_liberties
  • State_of_Emergency

Predicted Entities

Closures/lockdown, Government_oversight, Restrictions_of_daily_liberties, Restrictions_of_fundamental_rights_and_civil_liberties, State_of_Emergency

Download Copy S3 URI

How to use

document_assembler = nlp.DocumentAssembler() \
    .setInputCol("text")\
    .setOutputCol("document")

tokenizer = nlp.Tokenizer()\
    .setInputCols(["document"]) \
    .setOutputCol("token")

embeddings = nlp.BertEmbeddings.pretrained("bert_embeddings_bert_base_italian_xxl_cased", "it") \
    .setInputCols(["document", "token"])\
    .setOutputCol("embeddings")

embeddingsSentence = nlp.SentenceEmbeddings() \
    .setInputCols(["document", "embeddings"])\
    .setOutputCol("sentence_embeddings")\
    .setPoolingStrategy("AVERAGE")
    
multilabelClfModel = nlp.MultiClassifierDLModel.pretrained('legmulticlf_covid19_exceptions_italian', 'it', "legal/models") \
    .setInputCols(["sentence_embeddings"])\
    .setOutputCol("class")

clf_pipeline = nlp.Pipeline(
    stages=[document_assembler, 
            tokenizer,
            embeddings, 
            embeddingsSentence,
            multilabelClfModel])

df = spark.createDataFrame([["Al di fuori di tale ultima ipotesi, secondo le raccomandazioni impartite dal Ministero della salute, occorre provvedere ad assicurare la corretta applicazione di misure preventive quali lavare frequentemente le mani con acqua e detergenti comuni."]]).toDF("text")

model = clf_pipeline.fit(df)
result = model.transform(df)

result.select("text", "class.result").show(truncate=False)

Results

+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+---------------------------------+
|text                                                                                                                                                                                                                                                  |result                           |
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+---------------------------------+
|Al di fuori di tale ultima ipotesi, secondo le raccomandazioni impartite dal Ministero della salute, occorre provvedere ad assicurare la corretta applicazione di misure preventive quali lavare frequentemente le mani con acqua e detergenti comuni.|[Restrictions_of_daily_liberties]|
+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+---------------------------------+

Model Information

Model Name: legmulticlf_covid19_exceptions_italian
Compatibility: Legal NLP 1.0.0+
License: Licensed
Edition: Official
Input Labels: [sentence_embeddings]
Output Labels: [class]
Language: it
Size: 13.9 MB

References

Train dataset available here

Benchmarking

label                                                   precision  recall  f1-score  support 
Closures/lockdown                                       0.88       0.94    0.91      47      
Government_oversight                                    1.00       0.50    0.67      4       
Restrictions_of_daily_liberties                         0.88       0.79    0.83      28      
Restrictions_of_fundamental_rights_and_civil_liberties  0.62       0.62    0.62      16      
State_of_Emergency                                      0.67       1.00    0.80      6       
micro-avg                                               0.82       0.83    0.83      101     
macro-avg                                               0.81       0.77    0.77      101     
weighted-avg                                            0.83       0.83    0.83      101     
samples-avg                                             0.81       0.84    0.81      101