正文: … (explanation). Should be in Chinese. The user says: “请根据以下主题和问题撰写一篇文章，使用文本返回内容。文章的第一行应该是标题，格式如下：标题：[主题] 接下来是文章的正文部分。主题：语言建模的任务是什么 问题：语言建模的任务是什么”

Thus we need to produce an article with title line “标题：语言建模的任务是什么” and then a body that explains the task of language modeling.

We need to ensure it’s a proper article, maybe with sections: 什么是语言模型, 任务定义, 目标, 应用, 关键技术, 发展历程, 未来趋势. But keep it concise but thorough.

We must not include any extra content beyond the article.

Thus final answer: Title line then article.

We must ensure it’s plain text, not markdown. But the user didn’t specify to avoid markdown. It says “使用文本返回内容”. Could be plain text. Title line as given.

Thus final answer:

标题：语言建模的任务是什么

正文: … (multiple paragraphs). Let’s write in Chinese.

We need to answer: “语言建模的任务是什么”. So we need to explain that language modeling is the task of predicting the probability distribution of the next word given previous words, or the probability of a sequence of words, i.e., modeling the likelihood of a sentence, used for various NLP tasks, etc.

Thus produce article.

We can include bullet points? Possibly plain text. But it’s okay to include bullet points.

We must ensure the title is exactly “标题：语言建模的任务是什么”. Then newline, then article.

Let’s write.

Potential sections:

– 什么是语言模型
– 语言建模的核心任务
– 训练目标
– 评价指标
– 应用场景
– 关键技术
– 发展趋势

Will produce a coherent article.

Now produce final answer.

标题：语言建模的任务是什么

语言建模（Language Modeling）是自然语言处理（NLP）领域的核心任务之一，其主要目标是建立一种统计或神经网络模型，能够对自然语言的序列进行概率估计。具体来说，语言模型的任务可以概括为以下几个方面：

1. **预测下一个词**
给定一段已经出现的词序列（上下文），语言模型要计算下一个词出现的概率分布。即模型要学习条件概率 \(P(w_t | w_{