Conversational System
Conversational systems aim to create natural and engaging interactions between humans and machines, focusing on improving response accuracy, efficiency, and user experience. Current research emphasizes enhancing models' understanding of context, including memory and multi-modal inputs (text, images, audio), using techniques like retrieval-augmented generation, transformer-based architectures, and contrastive learning to mitigate issues like hallucination and inconsistency. These advancements hold significant implications for various applications, from personalized research assistance and healthcare information access to improved customer service and more human-like interactions in social robots.
Papers
May 7, 2022
April 15, 2022
March 31, 2022
February 21, 2022
December 2, 2021