LibriSQA: A Novel Dataset and Framework for Efficient Spoken Question Answering with Large Language Models

Large Language Models can effectively align and comprehend speech information, enabling the development of universal multimodal models for efficient Spoken Question Answering.