Rong Xiang
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
AI×1Crypto×1
Research Timeline
2026
Structural Enforcement of Goal Integrity in AI Agents via Separation-of-Powers Architecture
The paper proposes the Policy-Execution-Authorization (PEA) architecture, a separation-of-powers system designed to structurally enforce goal integrity in AI agents, moving safety from a probabilistic behavioral property to a guaranteed system constraint.
Highlighted terms show continued research focus across papers