Critic
الناقد
في شبكة واسرشتاين التوليدية، الشبكة العصبية التي تحل محل المُميِّز التقليدي. تُخرج درجة حقيقية غير محدودة (بدون سيغمويد) تقيّم مدى واقعية المدخل، ويجب أن تحقق قيد ليبشيتز.
In a Wasserstein GAN, the neural network that replaces the traditional discriminator. It outputs an unbounded real-valued score (no sigmoid) evaluating how real the input looks, and must satisfy the Lipschitz constraint.
Also translated asشبكة الناقد، دالة التقييم
First appears in this corpus in: Wasserstein GAN (2017)
Appears in these papers
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor2018in the sky ✦
- Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor2018in the sky ✦
- Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection2023in the sky ✦
- TD3: Addressing Function Approximation Error in Actor-Critic Methods2018in the sky ✦
- Wasserstein GAN2017in the sky ✦
- Wasserstein GAN2017in the sky ✦