今日从 arXiv 订阅中筛选 8 篇论文。

Arxiv ID2607.10781幻觉翻译2607.10781

⚡ Is Energy Guidance All You Need? Training-Free Norm Injection for Driving World Models

](https://arxiv.org/abs/2607.10781) | **PDF**: [`2607.10781.pdf`](https://arxiv.org/pdf/2607.10781.pdf)

Is Energy Guidance All You Need? Training-Free Norm Injection for Driving World Models
Arxiv ID2607.10071幻觉翻译2607.10071

⚡ FlashBEV: Fast and Memory-Efficient Exact BEV Transformation with IO-Awareness

](https://arxiv.org/abs/2607.10071) | **PDF**: [`2607.10071.pdf`](https://arxiv.org/pdf/2607.10071.pdf)

FlashBEV: Fast and Memory-Efficient Exact BEV Transformation with IO-Awareness
Arxiv ID2607.11862幻觉翻译2607.11862

⚡ Evidence-Backed Video Question Answering

](https://arxiv.org/abs/2607.11862) | **PDF**: [`2607.11862.pdf`](https://arxiv.org/pdf/2607.11862.pdf)

Evidence-Backed Video Question Answering
Arxiv ID2607.09740幻觉翻译2607.09740

⚡ A Dynamic Scene Interaction Reasoning Framework for Scene-level Lane-Change Intention and Trajectory Prediction

](https://arxiv.org/abs/2607.09740) | **PDF**: [`2607.09740.pdf`](https://arxiv.org/pdf/2607.09740.pdf)

A Dynamic Scene Interaction Reasoning Framework for Scene-level Lane-Change Intention and Trajectory Prediction
Arxiv ID2607.11836幻觉翻译2607.11836

⚡ Cycle-World: Mitigating Error Accumulation in Long-term Video World Models via Reverse-Prediction Cycle Consistency

](https://arxiv.org/abs/2607.11836) | **PDF**: [`2607.11836.pdf`](https://arxiv.org/pdf/2607.11836.pdf)

Cycle-World: Mitigating Error Accumulation in Long-term Video World Models via Reverse-Prediction Cycle Consistency
Arxiv ID2607.10744幻觉翻译2607.10744

⚡ Traj-VLN: Learning Pixel-Space Interaction via Autoregressive Trajectory Generation

](https://arxiv.org/abs/2607.10744) | **PDF**: [`2607.10744.pdf`](https://arxiv.org/pdf/2607.10744.pdf)

Arxiv ID2607.10362幻觉翻译2607.10362

⚡ A Control Theory of Predictability in Latent World Models

](https://arxiv.org/abs/2607.10362) | **PDF**: [`2607.10362.pdf`](https://arxiv.org/pdf/2607.10362.pdf)

A Control Theory of Predictability in Latent World Models
Arxiv ID2607.11173幻觉翻译2607.11173

⚡ When Depth Is Better Told Than Shown: Depth-Ordinal Prompting for Vision-Language Spatial Reasoning

](https://arxiv.org/abs/2607.11173) | **PDF**: [`2607.11173.pdf`](https://arxiv.org/pdf/2607.11173.pdf)

When Depth Is Better Told Than Shown: Depth-Ordinal Prompting for Vision-Language Spatial Reasoning

自动生成于 2026-07-15 · 基于 arXiv Daily Digest