Dialogue System
Dialogue systems aim to create natural and engaging conversations between humans and machines, primarily focusing on improving the accuracy, fluency, and contextual understanding of these interactions. Current research emphasizes enhancing memory capabilities, mitigating biases like hallucination and over-association, and improving robustness to noisy input such as from automatic speech recognition. This involves leveraging large language models (LLMs) and exploring novel architectures like mixture-of-experts and neuro-symbolic approaches, alongside the development of new evaluation benchmarks and datasets to better assess system performance. The advancements in this field have significant implications for various applications, including customer service, mental health support, and personalized education.
Papers
TK-KNN: A Balanced Distance-Based Pseudo Labeling Approach for Semi-Supervised Intent Classification
Nicholas Botzer, David Vasquez, Tim Weninger, Issam Laradji
Learning from Red Teaming: Gender Bias Provocation and Mitigation in Large Language Models
Hsuan Su, Cheng-Chu Cheng, Hua Farn, Shachi H Kumar, Saurav Sahay, Shang-Tse Chen, Hung-yi Lee
Are Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systems
Yixin Wan, Jieyu Zhao, Aman Chadha, Nanyun Peng, Kai-Wei Chang
From Data to Dialogue: Leveraging the Structure of Knowledge Graphs for Conversational Exploratory Search
Phillip Schneider, Nils Rehtanz, Kristiina Jokinen, Florian Matthes