Research is
an ongoing conversation.
Research questions, technical designs and architectural notes on intelligence, environments and interaction.
01
Embodied representation: turning observations into actionable understanding
Explore joint representations of vision, language and environmental state to establish shared context for task understanding and action planning.
Read article02
Multi-agent coordination: from individual decisions to joint solutions
Study task decomposition, role allocation and execution feedback so agents can exchange information and revise plans around a shared environmental goal.
Read article03
Human–AI interaction: putting intent at the center
Explore interaction flows combining natural language, spatial cues and human confirmation. Study how systems interpret intent and accept correction.
Read article04
Action loops: learning from every action
Build a research framework from environmental perception to action feedback, focusing on constraints, exception handling and explainable execution traces.
Read articleShowing 4 resources