Shima Esfandyari

New DRIFT method enhances LLM reasoning for aspect sentiment prediction

Researchers have developed DRIFT, a novel method for Aspect Sentiment Quadruple Prediction (ASQP) that directly injects linearized dependency parse trees into LLM prompts. This approach allows models to reason over expl…