Embodied AI 具身智能 Vision-Language-Action / VLAVLA × Reinforcement LearningVLA × World ModelRobot ManipulationRobot LearningEmbodied NavigationEmbodied Spatial ReasoningSim2Real / Cross-Embodiment 查看研究 Idea →
Multimodal AI 多模态 MLLM / VLM多模态理解与生成统一模型Vision TokenizerMultimodal ReasoningSpatial ReasoningVideo MLLM / 长视频理解Multimodal RAGMultimodal Agent多模态安全与幻觉 查看研究 Idea →
Large Language Model 大语言模型 大模型推理 ReasoningTest-Time Scaling / 推理时计算RAG / 知识增强Long Context / 长上下文大模型预训练与持续预训练大模型后训练 Post-TrainingEfficient LLM / AI Infra 查看研究 Idea →
3D Vision & Generation 3D 3D Gaussian SplattingNeRF / Neural RenderingFeed-Forward 3D ReconstructionDynamic 3D / 4D3D Generation3D Vision-Language ModelSpatial IntelligencePoint Cloud 查看研究 Idea →
AI Agent Agent Agent Planning & ReasoningAgent MemoryTool-Use AgentGUI / Computer-Use AgentWeb AgentDeep Research / Search AgentMulti-AgentSelf-Evolving Agent 查看研究 Idea →
Autonomous Driving 自动驾驶 3D PerceptionTrajectory PredictionEnd-to-End Autonomous DrivingDriving World Model 查看研究 Idea →
世界模型 World Model Video World ModelPhysical World ModelInteractive World ModelWorld Model × RoboticsWorld Model × RLWorld Model × Agent3D / 4D World Model 查看研究 Idea →
强化学习 Reinforcement Learning RL for LLM ReasoningRL for AgentMultimodal RLOffline RLMulti-Agent RLRobot RL 查看研究 Idea →
AI for Science / AI+交叉应用 AI + 医疗AI + 生物 / 分子 / 药物AI + 遥感AI + 金融AI + 推荐系统AI + 时间序列AI + Scientific Agent 查看研究 Idea →
