Zero-AVSR: Zero-Shot Audio-Visual Speech Recognition with LLMs by Learning Language-Agnostic Speech Representations Paper • 2503.06273 • Published Mar 8, 2025 • 6
