本文研究的是深度模型在有噪数据上的泛化能力，提出了MentorMix方法来克服真实和人工合成的噪声标签。 paper: https://ai.googleblog.com/2020/08/understanding-deep-learning-on.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+blogspot%2FgJZg+%28Google+AI+Blog%29source: ICML 2020code: https://github.com/google-research/google-research/tree/master/mentormix
Generative Models ：PixelRNN and PixelCNN ； Variational Autoencoders (VAE) ； Generative Adversarial Networks (GAN)
Eager execution is a feature that makes TensorFlow execute operations immediately: concrete values are returned, instead of a computational graph to be executed later.
基于cs20si的Tensorflow笔记，课程主页；本节主要内容：Basic Operations, Constants, Variables, Control Dependencies, Feeding inputs, TensorBoard
序列最小最优化(sequential minimal optimization, SMO)算法是支持向量机的学习算法，本文主要讲解了SMO算法，并且实现了一个简单基于SMO的SVM demo，最后简要了解封装SVM的scikit-learn库。
论文提出了一种隐意图对话模型（Latent Intention Dialogue Model, LIDM），通过离散的隐变量来学习对话意图，这些隐变量可以看作引导对话生成的动作决策，提高基于手工构建的状态-动作集传统强化学习模型所生成对话的多样性。