Pin Lyu

zero-shot learning (1) benchmark evaluation (1) self-supervised learning (1) kl divergence (1) code generation (1) bidirectional attention (1) policy iteration (1) semantic correlation (1) safe reinforcement learning (1) software development (1) complex reasoning (1) code agent (1) barrier function (1) jailbreak defense (1) constrained policy optimization (1) lagrangian method (1) generalized zero-shot learning (1) code execution (1) task completion (1) attribute feature (1)