NLP(Natural Language Processing) has grown immensely in recent years and a lot of abbreviations are used in this field to name models that are used for different tasks like text summarization, sentiment analysis, question-answering, etc. These models are known as Language Models. Most of the Language models are named after muppets, yes you heard it right muppets. Elmo, BigBird, Bert, and Ernie are muppets of the famous American TV series Sesame Street.
These are the list of 15 most famous language models and algorithms used in the domain of NLP which are must know
1) LSTM - Long Short Term Memory
2) ELMo - Embeddings from Language Models
3) GloVe - Global Vectors for Word Representation
4) ULMFiT - Universal Language Model Fine-Tuning
5) Transformer-XL - Transformer Extra Long
6) BERT - Bidirectional Encoder Representation from Transformers
7) RoBERTa - Robustly Optimized BERT Pretraining Approach
8) ERNIE - Enhanced Representation through kNowledge IntEgration
9) GPT-2 - Generative Pretrained Transformer - 2
10) MT-DNN a.k.a BigBird - Multi-Task Deep Neural Network
11) PEGASUS - Pre-training with Extracted Gap-Sentences for Abstractive Summarization
12) ALBERT - A Lite BERT
13) ELECTRA - Efficiently Learning an Encoder that Classifies Token Replacements Accurately
14) T5 - Text-To-Text Transfer Transformer
15) DAAF - Data Augmentation and Auxiliary Feature
Top comments (1)
Has anyone watched the series Sesame Street which is a favourite of most Data Scientist dealing with NLP, as is obvious from the names of the best language models.