Wang et al. systematically analyzed trusted setups in their paper on , distinguishing between setup protocols and ceremonies, and evaluating the characteristics, advantages, and weaknesses of different methods. Wang等人在论文中系统化地分析了可信设置,区分了设置协议和仪式,并评估了不同方法的特点、优势和弱点。
Tan et al. proposed a novel zkPoT method in their paper on , which significantly improves performance by directly proving the correctness of the trained model, rather than the entire training process, through the introduction of an optimum vicinity concept. Tan等人在论文中提出了一种新的zkPoT方法,通过引入optimum vicinity概念,直接证明训练模型的正确性,而非整个训练过程,显著提高了性能。