- I have been reading on the capabilities of LLM based conversational agents and have been wondering if there is even possibility for any further enhancement with the addition of NER to such system.
- If so, in which case could a conversational agents powered by an LLM like say Dolly 2.0 be enhanced by NER?
It's probably best to look into some research but here is some information which might help in the meantime based on some general thoughts;
Adding NER to a LLM can enhance the models ability to understand and respond to user input. NER can help the system identify and extract important contextual information such as names, locations, and dates from the user's input, which can then be used to generate more accurate and relevant responses. For example, if a user asks for information about a specific event, the system can use NER to identify the date and location of the event and provide more detailed information.
You may want to also look into the following methods or use cases;
Topic Extraction: NER can help identify relevant keywords and topics from user input, which can be used to personalize the conversation and provide more relevant responses.
Sentiment Analysis: By identifying named entities related to emotions or sentiment, NER can help classify the overall tone of the conversation and adjust responses accordingly.
Filtering Out Irrelevant Content: NER can be used to filter out irrelevant content from user input, such as stop words, prepositions, and conjunctions, which can improve the accuracy of the conversational agent's responses.
Entity Linking: NER can help link named entities to external sources of information, such as Wikipedia or Google Knowledge Graph, and provide more informative responses to user queries.
A combination of these with prompt pre-processing and filtering may provide better responses.
I hope this helps in some way!