Wei Ren

A shared state estimator can enable effective policy-gradient transfer even with differing simulator and plant dynamics, by bounding the gradient discrepancy linearly with model error.

BENCHCOMPASS is introduced, a payment-domain benchmark whose construction pipeline builds scenario-grounded tasks from typed evidence packs, applies LLM-based quality checks, creates task-input attack variants, and reserves final item admission for domain experts.

Agents can now effectively collaborate in multi-robot systems even when facing unpredictable computational delays, leading to improved learning and control performance.

Novel high-order control barrier functions enable explicit safety control in nonlinear systems, ensuring stabilization without compromising safety.