“任意”上下文无关语法？
长期崇拜者第一次询问者:)
我正在开发一个程序，该程序从上下文无关语法导出确定性有限状态自动机，而我分配的论文解释了如何做到这一点，一直提到“任意概率上下文” -自由语法”，但从未定义与 PCFG 相关的“任意”的含义。我认为他们的意思是“任何旧的 PCFG”，但为什么不直接说“任何 PCFG”呢？
该术语也出现在多个维基百科条目中。在 CFG 页面的顶部，有一个关于 CFG 的任意性的参考（“子句可以任意深度地嵌套在子句内”），但没有明确说明为什么有人将 PCFG 或 PCFG 子集称为任意的。
如果有人好奇，这篇论文是 Klein 和 Manning (2001) 的解析和超图；我还阅读了他们与本文相关的另外两篇论文（用于任意概率上下文无关语法的基于议程的图表解析器和经验界限、理论模型和宾夕法尼亚树库），它广泛使用该术语，但也从未对其进行解释。
Long time admirer first time inquirer :)
I'm working on a program which derives a deterministic finite-state automata from a context-free grammar, and the paper I have been assigned which explains how to do this keeps referring to "arbitrary probabilistic context-free grammars" but never defines the meaning of "arbitrary" in relation to PCFGs. I assume they mean "any old PCFG" but then why not just say "any PCFG"?
The term also turns up in several Wikipedia entries. At the top of the CFG page there is a reference to arbitrariness in relation to CFGs on ("clauses can be nested inside clauses arbitrarily deeply"), but doesn't make clear why someone would refer to a PCFG or subset of PCFGs as arbitrary.
In case anyone is curious, the paper is Parsing and Hypergraphs by Klein and Manning (2001); I've also been reading two other papers by them related to this one (An Agenda-Based Chart Parser for Arbitrary Probabilistic Context-Free Grammars and Empirical Bounds, Theoretical Models, and the Penn Treebank) which use the term extensively but never explain it either.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
“任意”意味着“任何”，就像您所想的那样。它强调他们对 PCFG 没有假设任何特殊条件。至于为什么，如果你看“任意 PCFG 的表格解析框架”这样的句子，“任意”是比“任何”更好的词选择。
"Arbitrary" means "any" like you are thinking. It emphasizes that they assume no special conditions on the PCFG. As for why, if you look at a sentence like "a tabular parsing framework for arbitrary PCFGs", "arbitrary" is a better word choice than "any".