Get Started
Home
Topics
Search
Library
Research questionHow can noisy annotations in industrial robot dialogues be corrected for reliable dialogue state tracking?Industrial robot dialogue systems rely on accurate utterance and state annotations to track conversations. Annotation errors can make tracking failures reflect dataset noise rather than model capability.
AI
Evaluation & Benchmarks
Natural Language Processing
Robotics
Latest papersRecent research connected to this question, newest first.IRWOZ 2.0: A Large Language Model-driven Dialogue Dataset for Industrial Robot ConversationsThe source describes IRWOZ 2.0, a 390-dialogue dataset refined through manual corrections, automated typo removal, and LLM-enhanced generation. It reports dialogue-state-tracking benchmarks, including a GPT-2 BLEU-4 increase from 0.1651 to 0.5604 over the original IRWOZ, and publicly releases the dataset.research paper · Sep 3, 2026
Related questions
How can contradictions in dialogue-based human–robot interactions be formally represented across HRI and HAI domains?How can direct vision-language-action robot policies capture multi-timescale dynamics without learning undesirable behavior from mixed-quality deployment trajectories?How can full-duplex dialogue models learn natural acoustic turn-taking without degrading semantic responses?How can LLM orchestrators preserve continuous state when collaborating with non-language agents?