This study marks a significant advancement by harnessing Large Language
Models (LLMs) for multi-intent spoken language understanding (SLU), proposing a
unique methodology that capitalizes on the generative power of LLMs within an
SLU context. Our innovative technique reconfigures entity slots specifically
for LLM application in multi-intent SLU environments and introduces the concept
of Sub-Intent Instruction (SII), enhancing the dissection and interpretation of
intricate, multi-intent communication within varied domains. The resultant
datasets, dubbed LM-MixATIS and LM-MixSNIPS, are crafted from pre-existing
benchmarks. Our research illustrates that LLMs can match and potentially excel
beyond the capabilities of current state-of-the-art multi-intent SLU models. It
further explores LLM efficacy across various intent configurations and dataset
proportions. Moreover, we introduce two pioneering metrics, Entity Slot
Accuracy (ESA) and Combined Semantic Accuracy (CSA), to provide an in-depth
analysis of LLM proficiency in this complex field.

本研究通过利用大型语言模型（LLMs）进行多意图口语语言理解（SLU）的研究，提出一种独特的方法，充分利用 LLMs 的生成能力，在 SLU 环境中重构了实体槽位，引入了子意图指令（SII）的概念，增强了对复杂多意图通信的解析和解释，从而形成了名为 LM-MixATIS 和 LM-MixSNIPS 的结果数据集。我们的研究表明，LLMs 可以匹配并潜在地超越当前最先进的多意图 SLU 模型的能力。此外，我们还介绍了两个开创性的度量标准，实体槽位准确度（ESA）和综合语义准确度（CSA），以深入分析 LLMs 在这一复杂领域的熟练程度。