[D] Are traditional NLP tasks such as text classification/NER/RE still important in the era of LLMs?

The large language models (LLMs) are comprehensive NLP task solvers. They can do many NLP tasks by the earliest program language, English, rather than the discriminative-based models that need complicated processes.

As a PhD student, I have been a little anxious about the future recently. My advisor supports a funded project on knowledge extraction and knowledge graph. But I think the future must be LLMs. My current plan is to do some RAG work after completing the NER and RE …

classification english future language language models large language large language models llms machinelearning ner nlp phd processes tasks text text classification