May 2021 – Jun 2021
Worked on proposing and implementing a model/framework for Named Entity Recognition (NER) from the defence corpus.
The input embedding module includes Syntactic Textual Feature Embedding (Dependency Path, PoS tag, TF-IDF, Word Shape) and Semantic Textual Feature Embedding (Glove, Word2Vec, Fastext, BERT Embedding, ELMo Embedding, Domain Embedding). The NER model applied Bi-LSTM with self-attention, attached Conditional Random Fields (CRF).
Got a HD mark for this assignment.