UniverSLU: Universal Spoken Language Understanding for Diverse Tasks with Natural Language Instructions

arXiv:2310.02973v2 Announce Type: replace
Abstract: Recent studies leverage large language models with multi-tasking capabilities, using natural language prompts to guide the model's behavior and surpassing performance of task-specific models. Motivated by this, we ask: can we build a single model that jointly performs various spoken language understanding (SLU) tasks? We start by adapting a pre-trained automatic speech recognition model to additional tasks using single-token task specifiers. We enhance this approach through instruction tuning, i.e., finetuning by describing the task using …

abstract arxiv behavior build capabilities cs.cl cs.sd diverse eess.as guide language language models language understanding large language large language models natural natural language natural language prompts performance prompts slu spoken spoken language understanding studies tasks type understanding universal