没有合适的资源?快使用搜索试试~ 我知道了~
首页Natural Language Processing Language Modelling and Machine Translation
Natural Language Processing Language Modelling and Machine Trans...
需积分: 10 109 浏览量
更新于2023-05-30
评论 1
收藏 6.66MB PDF 举报
Natural Language Processing Language Modelling and Machine Translation 深度学习暑期学校2017
资源详情
资源评论
资源推荐

Natural Language Processing
Linguistics
Why are human languages the way that they are?
How does the brain map from raw linguistic input to
meaning and back again? And how do children learn
language so quickly?
Computational Linguistics
Computational models of language and
computational tools for studying language.
Natural Language Processing
Building tools for processing language and
applications that use language:
•
Intrinsic: Parsing, Language Modelling, etc.
•
Extrinsic: ASR, MT, QA/Dialogue, etc.

Language models
A language model assigns a probability to a sequence of words,
such that
P
w∈Σ
∗
p(w) = 1:
Given the observed training text, how probable is this new
utterance?
Thus we can compare different orderings of words
(e.g. Translation):
p(he likes apples) > p(apples likes he)
or choice of words (e.g. Speech Recognition):
p(he likes apples) > p(he licks apples)

History: cryptography

Language models
Much of Natural Language Processing can be structured as
(conditional) language modelling:
Translation
p
lm
(Les chiens aiment les os ||| Dogs love bones)
Question Answering
p
lm
(What do dogs love? ||| bones . |β)
Dialogue
p
lm
(How are you? ||| Fine thanks. And you? |β)
剩余90页未读,继续阅读


















安全验证
文档复制为VIP权益,开通VIP直接复制

评论0