A Novel User-Friendly Pipeline for Enhanced Natural Language Understanding in Human-Robot Interaction
This paper presents an innovative Natural Language Understanding (NLU) pipeline for humanrobot interactions (HRI), optimized for on-premises deployment in industrial settings. The proposed system integrates an end-to-end Automated Speech Recognition (ASR) system, a transformer-based model for intent and entity recognition, and a dynamic dialogue management system. These components operate on commodity hardware, ensuring real-time responsiveness without cloud dependency. The pipeline is uniquely extensible via an automated, offline training module that uses large language models like ChatGPT to generate datasets, reducing the need for specialized machine learning expertise.
Paper
The full text of this publication is not hosted on 44B due to licensing.
Read it at OpenAlex