Why Your Tokenizer Fails in Information Fusion: A Timing-Aware Pre-Quantization Fusion for Video-Enhanced Audio Tokenization

Documenti analoghi

• Why Pre-trained Models Fail: Feature Entanglement in Multi-modal Depression Detection
  di: Zhang, Xiangyu, et al.
  Pubblicazione: (2025)
• Zero-Shot Cognitive Impairment Detection from Speech Using AudioLLM
  di: Shahin, Mostafa, et al.
  Pubblicazione: (2025)
• Beyond Video-to-SFX: Video to Audio Synthesis with Environmentally Aware Speech
  di: Niu, Xinlei, et al.
  Pubblicazione: (2025)
• Mamba in Speech: Towards an Alternative to Self-Attention
  di: Zhang, Xiangyu, et al.
  Pubblicazione: (2024)
• When LLMs Meets Acoustic Landmarks: An Efficient Approach to Integrate Speech into Large Language Models for Depression Detection
  di: Zhang, Xiangyu, et al.
  Pubblicazione: (2024)