提前预判任务型对话系统的故障时刻，为智能客服与语音助手提供可靠护航。
arXiv:2607.03974v1 Announce Type: new Abstract: Task-oriented dialogue systems often fail before the final breakdown is obvious, but most evaluation o…
用输入重写替代监督澄清，让大模型在对话篇章解析上更稳，算法细节与实验并重。
arXiv:2607.01964v1 Announce Type: new Abstract: Rewriting inputs to improve frozen downstream models has become a common strategy in modern NLP pipeli…
聚焦LLM对话代理在数据库故障时的安全恢复策略，SIGDIAL 2026录用论文。
arXiv:2606.31307v1 Announce Type: new Abstract: Large language models used in task-oriented dialogue often produce fluent but unsafe responses when ba…