ju-resplande / intro_ner Goto Github PK

Materiais e dicas de introdução a NER.

Só escrever uma issue que acrescento mais materiais aqui. 😉

Named-Entitity Recognition (NER) também conhecido como extração de entidades é uma tarefa de processamento de linguagem natural (NLP) voltada para a classificação de sequências (de tokens).

• https://towardsdatascience.com/named-entity-recognition-with-nltk-and-spacy-8c4a7d88e7da
• https://medium.com/elinttech/uma-vis%C3%A3o-geral-sobre-named-entity-recognition-ner-4dc4e3b5e37a

• https://arxiv.org/abs/1603.01360
• https://arxiv.org/abs/1910.11470

• https://cursosextensao.usp.br/course/view.php?id=2721
• @lucasbrandao4770: https://colab.research.google.com/drive/1dhEJ2VuphulLVwg9GKf0L9wEaVePN5cB
• https://github.com/thunlp/NLP-THU
• https://github.com/NielsRogge/Transformers-Tutorials
• https://github.com/EricFillion/happy-transformer
• https://github.com/keon/awesome-nlp
• https://github.com/ivan-bilan/The-NLP-Pandect
• https://github.com/dair-ai/Transformers-Recipe

• https://www.sciencedirect.com/science/article/pii/S1110866520301596
• https://en.wikipedia.org/wiki/Inside%E2%80%93outside%E2%80%93beginning_(tagging)

• Simple-transformers: https://github.com/ThilinaRajapakse/simpletransformers/blob/master/examples/named_entity_recognition/named_entity_recognition.py
• Flair: https://github.com/flairNLP/flair/tree/master/examples/ner
• FARM: https://github.com/deepset-ai/FARM/blob/master/examples/ner.py
• Transformers: https://github.com/huggingface/notebooks/blob/main/examples/token_classification.ipynb

• https://huggingface.co/datasets?task_categories=task_categories:token-classification
• @piegu: https://forum.ailab.unb.br/t/datasets-em-portugues/251
• https://github.com/davidsbatista/NER-datasets

• https://github.com/gdutramartins/po-ner-2-portuguese-ner

• Por @perkhad rota de estudo:
  1. Estudar os formatos (esquemas) padrões (ex. Conll)
  2. Ler o artigo sobre bi-lstm para ner
  3. Testar frameworks alto nível do transformers (ex: Flairnlp, Simple-transformers FARM)
  4. Ir para implementação do Transformers