+95%
48/38
VLA
小方向 小方向
Subdirections
方法族趨勢來源於 field-state-*.json（機械化，零 LLM）。加速度 = (近 7d − 前 7d) / max(前值, 1)。
方法族趋势来源于 field-state-*.json（机械化，零 LLM）。加速度 = (近 7d − 前 7d) / max(前值, 1)。
速度榜 · 前 8 名 速度榜 · 前 8 名 / Velocity Top 8
- 01 AI Context Engineering +200% 2/1
- 02 AI Vertical Agents +200% 2/1
- 03 AI Agentic Coding +183% 11/6
- 04 AI Frontier Models +150% 6/4
- 05 AI Agent Infra +100% 4/4
- 06 VLA VLA Generalist +95% 48/38
- 07 VLA World Model Dynamics +86% 40/35
- 08 VLA Hierarchical Planning +85% 17/15
新冒頭新冒头 · EMERGING · prior volume = 0
VLA · 15
+86%
40/35
VLA
World Model Dynamics vla.world.world_model_dynamics
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+85%
17/15
VLA
Hierarchical Planning vla.reasoning.hierarchical_planning
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+75%
17/17
VLA
VLA Post-Training RL vla.policy.robot_rl.vla_post_training_rl
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+70%
13/14
VLA
Tactile Sensing vla.manipulation.tactile_sensing
+60%
11/3
VLA
Mobile Manipulation vla.manipulation.mobile_manipulation
佐證論文佐证论文 M3-Tele: A Unified Multimodal Teleoperational Framework for Compliant Whole-Body Mobile Manipulation DreamTrajectory: Trajectory-Guided Action Generation with World Model Alignment for Mobile Manipulation Embodied Multimodal Grounding for Open-Vocabulary Mobile Manipulation via Semantic 3D Gaussian Splatting
+38%
7/4
VLA
Diffusion Policy vla.action.diffusion_policy
頂尖實驗室：顶尖实验室： Shanghai Jiao Tong UniversityTsinghua Shenzhen International Graduate School, Tsinghua UniversitySchool of Artificial Intelligence, University of Chinese Academy of Sciences
佐證論文佐证论文 MemCorr-DP: Counterfactual Correspondence Conditioning for a Diffusion Policy Guided by a Reference Dressing in Motion: A Human Motion-Aware Diffusion Policy for Robot-Assisted Dressing SCRIPT: Scalable Diffusion Policy with Multi-stage Training for Language-driven Physics-Based Humanoid Control
+25%
7/21
VLA
Flow Matching vla.action.flow_matching
+27%
5/10
VLA
VLA Generalist vla.foundation.vla_generalist
頂尖實驗室：顶尖实验室： University of Illinois Urbana-Champaign
+22%
4/1
VLA
Cross-Embodiment Foundation Models vla.foundation.cross_embodiment_fm
+22%
4/3
VLA
Sim-to-Real vla.sim_data.sim_to_real
+16%
3/4
VLA
Dexterous Manipulation vla.manipulation.dexterous_manipulation
佐證論文佐证论文 RoboTok: An Internet-Scale Data Engine for Human Demonstration Retrieval and Dexterous Manipulation Learning RoboTok: An Internet-Scale Data Engine for Human Demonstration Retrieval and Dexterous Manipulation Learning Does Imitation Learning Preserve Temporal Robustness in Dexterous Manipulation? An Expert-Learner Comparison Across Task Execution Speeds
+16%
3/11
VLA
Imitation Learning vla.policy.imitation_learning
佐證論文佐证论文 Does Imitation Learning Preserve Temporal Robustness in Dexterous Manipulation? An Expert-Learner Comparison Across Task Execution Speeds Rethinking Demonstration Unlearning in Imitation Learning for Robotics VERAGMIL: Virtual Environment for Scooping Granular Foods with Imitation Learning Models
+16%
3/1
VLA
Imitation Learning vla.policy.imitation_learning
佐證論文佐证论文 Does Imitation Learning Preserve Temporal Robustness in Dexterous Manipulation? An Expert-Learner Comparison Across Task Execution Speeds Rethinking Demonstration Unlearning in Imitation Learning for Robotics VERAGMIL: Virtual Environment for Scooping Granular Foods with Imitation Learning Models
+5%
1/2
VLA
3D Representation vla.perception.representation_3d
AI App · 11
+183%
11/6
AI App
Agentic Coding ai.coding.agentic_coding
頂尖實驗室：顶尖实验室： Qwen
+150%
6/4
AI App
Frontier Models ai.frontier.frontier_model
+100%
4/4
AI App
Agent Infra ai.infra.agent_infra
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+999%
2/0
AI App
MCP (Model Context Protocol) ai.tools.mcp_protocol
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+999%
2/0
AI App
Explicit Reasoning (CoT / Search) ai.reasoning.reasoning_planning
+200%
2/1
AI App
Context Engineering ai.context.context_engineering
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+200%
2/1
AI App
Vertical Agents ai.vertical.vertical_agent
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
+67%
2/3
AI App
Voice & Multimodal ai.frontier.voice_multimodal
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
0%
0/0
AI App
Multi-Agent Teams ai.multi_agent.multi_agent
頂尖實驗室：顶尖实验室： Institute of Interdisciplinary Information Sciences at Tsinghua UniversityDepartment of Computer Science, University of Freiburg
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
0%
0/0
AI App
Agent Safety ai.safety.agent_safety
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.
0%
0/0
AI App
Agent Eval ai.eval.agent_eval
No evidence papers matched in the last 90 days. This family is tracked mechanically — when the threshold fires, papers will appear here.