วีดีโอ

生日问题是问最少需要多少人，当中两人同一天生日的概率才会过半，答案是23人。这问题有时也称生日悖论，但从引起逻辑矛盾的角度来说生日问题并非悖论，它称作悖论只因这事实与一般直觉相抵触而已。大多数人会认为23人中两人同生日的概率应该远小于一半。计算与此相关的概率称为生日问题

The self-supervised pre-training of Large Language Models (LLMs) is a primary training method used to generate powerful and versatile natural language processing models. Self-Supervised Learning is a type of learning that does not require manually labeled data, and it is achieved by automatically generating labels from unlabeled data for training

“极大似然估计(Maximum Likelihood Estimation, MLE)是一种常见的参数估计方法，它通过最大化样本数据的似然函数，来求解参数的值。在给定一个观察数据集的情况下，MLE 试图找到最适合这个数据集的概率分布参数，以便在未来的预测中能够更好地描述这个数据集” from chatgpt