Efficient Intent-Based Filtering for Multi-Party Conversations Using Knowledge Distillation from LLMs
This work addresses efficiency for resource-constrained environments in conversational AI, but it is incremental as it builds on existing methods like knowledge distillation and fine-tuning.
The paper tackles the problem of high computational costs in processing multi-party conversations with large language models (LLMs) by proposing an intent-based filter that uses knowledge distillation from LLMs to select only relevant snippets for LLM processing, reducing operational costs as shown in experiments.
Large language models (LLMs) have showcased remarkable capabilities in conversational AI, enabling open-domain responses in chat-bots, as well as advanced processing of conversations like summarization, intent classification, and insights generation. However, these models are resource-intensive, demanding substantial memory and computational power. To address this, we propose a cost-effective solution that filters conversational snippets of interest for LLM processing, tailored to the target downstream application, rather than processing every snippet. In this work, we introduce an innovative approach that leverages knowledge distillation from LLMs to develop an intent-based filter for multi-party conversations, optimized for compute power constrained environments. Our method combines different strategies to create a diverse multi-party conversational dataset, that is annotated with the target intents and is then used to fine-tune the MobileBERT model for multi-label intent classification. This model achieves a balance between efficiency and performance, effectively filtering conversation snippets based on their intents. By passing only the relevant snippets to the LLM for further processing, our approach significantly reduces overall operational costs depending on the intents and the data distribution as demonstrated in our experiments.