Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders

Documenti analoghi

• Multi-Rationale Explainable Object Recognition via Contrastive Conditional Inference
  di: Rasekh, Ali, et al.
  Pubblicazione: (2025)
• Harnessing Synthetic Preference Data for Enhancing Temporal Understanding of Video-LLMs
  di: Vani, Sameep, et al.
  Pubblicazione: (2025)
• SmartSight: Mitigating Hallucination in Video-LLMs Without Compromising Video Understanding via Temporal Attention Collapse
  di: Sun, Yiming, et al.
  Pubblicazione: (2025)
• TemporalVLM: Video LLMs for Temporal Reasoning in Long Videos
  di: Fateh, Fawad Javed, et al.
  Pubblicazione: (2024)
• MOOSE: Pay Attention to Temporal Dynamics for Video Understanding via Optical Flows
  di: Nguyen, Hong, et al.
  Pubblicazione: (2025)