Seed1.8把Agent做进基座，语言训练让视觉退化

也值得关注

• Seed1.8 Model Card: Towards Generalized Real-World Agency score 4
  机构: ByteDance；关键词(6): deployment, latency, agentic, tool use, code generation
• High-Quality and Efficient Turbulence Mitigation with Events score 4
  关键词(1): latency；顶会接收: CVPR
• SATTC: Structure-Aware Label-Free Test-Time Calibration for Cross-Subject EEG-to-Image Retrieval score 4
  关键词(1): scaling；顶会接收: CVPR
• Mamba Learns in Context: Structure-Aware Domain Generalization for Multi-Task Point Cloud Understanding score 4
  关键词(3): lightweight, mamba, reasoning；顶会接收: CVPR
• CTCal: Rethinking Text-to-Image Diffusion Models via Cross-Timestep Self-Calibration score 4
  关键词(1): text-to-image；顶会接收: CVPR
• Memory-Efficient Fine-Tuning Diffusion Transformers via Dynamic Patch Sampling and Block Skipping score 4
  关键词(3): deployment, fine-tuning, text-to-image；顶会接收: CVPR
• MEMO: Human-like Crisp Edge Detection Using Masked Edge Prediction score 4
  关键词(3): lightweight, edge, fine-tuning；顶会接收: CVPR
• Predictive Regularization Against Visual Representation Degradation in Multimodal Large Language Models score 4
  关键词(2): reasoning, vision-language；顶会接收: CVPR
• PlanaReLoc: Camera Relocalization in 3D Planar Primitives via Region-Based Structure Matching score 4
  关键词(1): lightweight；顶会接收: CVPR
• Cross-modal Fuzzy Alignment Network for Text-Aerial Person Retrieval and A Large-scale Benchmark score 3
  顶会接收: CVPR