2 papers
cs.AI2026
What Can Be Enforced? A Theory of Certified Runtime Safety for Tool-Using Agents
Shawn Ray
Runtime guardrails act before irreversible tool calls, but their guarantees depend on what policy state is representable, what a judge observes, and whether intervention changes fu…
cs.LG2026
Graph-SND: Sparse Aggregation for Behavioral Diversity in Multi-Agent Reinforcement Learning
Shawn Ray
System Neural Diversity (SND) measures behavioral heterogeneity in multi-agent reinforcement learning by averaging pairwise distances over all agent pairs, making ea…