Rewriting

fexed · Nov 13, 2023 · 32d6746 · 32d6746
1 parent 28c9c85
commit 32d6746
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/UniPi_CSMaster_AI/HumanLanguageTechnologies/humanlanguagetechnologies.pdf b/UniPi_CSMaster_AI/HumanLanguageTechnologies/humanlanguagetechnologies.pdf
diff --git a/UniPi_CSMaster_AI/HumanLanguageTechnologies/humanlanguagetechnologies.tex b/UniPi_CSMaster_AI/HumanLanguageTechnologies/humanlanguagetechnologies.tex
@@ -1255,8 +1255,8 @@ \subsection{Statistical Machine Translation}
 \begin{center}
 	\includegraphics[scale=0.5]{42.png}
 \end{center}
-\paragraph{Fluency $P(T)$} How to measure that a sentence is more fluent than another? E.g. "That car was almost crash onto me" is less fluent than "That car almost hit me".\\
-The answer is language models, $n$-grams! For example, $P($hit$\:|\:$almost$)>P($was$\:|\:$almost$)$.\\
+\paragraph{Fluency $P(T)$} How to measure if a sentence is more fluent than another? E.g. "That car was almost crash onto me" is less fluent than "That car almost hit me".\\
+The answer is language models, N-grams! For example, $P($hit$\:|\:$almost$)>P($was$\:|\:$almost$)$.\\
 But can use any other more sophisticated model of grammar. Advantage: it's monolingual knowledge.
 \paragraph{Faithfulness $P(S\:|\:T)$} How to quantify? Intuition: degree to which words in one sentence are plausible translations of words in the other sentence. Product of probabilities that each word in target sentence would generate each word in source sentence.\\
 Need to know, for every target language word, the probability of it mapping to every source language word. How to learn this? Parallel texts: lots of times we have two texts that are translations of each other. If we knew which word in Source text mapped to each word in Target text, we could just count.