0:000:00
<p>本期的 15 篇论文如下：</p><p>[00:24] 🧠 Qwen3-VL Technical Report（Qwen3-VL 技术报告）</p><p>[00:57] 🧠 PretrainZero: Reinforcement Active Pretraining（PretrainZero：强化主动预训练）</p><p>[01:36] 🎬 ViDiC: Video Difference Captioning（ViDiC：视频差异描述）</p><p>[02:24] 🧠 OneThinker: All-in-one Reasoning Model for Image and Video（OneThinker：面向图像与视频的全能推理模型）</p><p>[03:07] 🔄 Rethinking Prompt Design for Inference-time Scaling in Text-to-Visual Generation（重新思考文本到视觉生成中推理时扩展的提示设计）</p><p>[03:59] ⚙ Steering Vision-Language-Action Models as Anti-Exploration: A Test-Time Scaling Approach（引导视觉-语言-动作模型作为反探索：一种测试时缩放方法）</p><p>[04:46] 🤖 SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RL（SpaceTools：通过双重交互式强化学习实现工具增强的空间推理）</p><p>[05:22] 🔧 Thinking with Programming Vision: Towards a Unified View for Thinking with Images（以编程视觉思考：迈向图像思维的统一视角）</p><p>[06:01] 🔄 Flowing Backwards: Improving Normalizing Flows via Reverse Representation Alignment（逆向流动：通过反向表征对齐改进标准化流）</p><p>[06:51] 🎮 RELIC: Interactive Video World M...