1. The document discusses various statistical and neural network-based models for representing words and modeling semantics, including LSI, PLSI, LDA, word2vec, and neural network language models.
2. These models represent words based on their distributional properties and contexts using techniques like matrix factorization, probabilistic modeling, and neural networks to learn vector representations.
3. Recent models like word2vec use neural networks to learn word embeddings that capture linguistic regularities and can be used for tasks like analogy-making and machine translation.
The document discusses distances between data and similarity measures in data analysis. It introduces the concept of distance between data as a quantitative measure of how different two data points are, with smaller distances indicating greater similarity. Distances are useful for tasks like clustering data, detecting anomalies, data recognition, and measuring approximation errors. The most common distance measure, Euclidean distance, is explained for vectors of any dimension using the concept of norm from geometry. Caution is advised when calculating distances between data with differing scales.
1. The document discusses principal component analysis (PCA) and explains how it can be used to determine the "true dimension" of vector data distributions.
2. PCA works by finding orthogonal bases (principal components) that best describe the variance in high-dimensional data, with the first principal component accounting for as much variance as possible.
3. The lengths of the principal component vectors indicate their importance, with longer vectors corresponding to more variance in the data. Analyzing the variances of the principal components can provide insight into the shape of the distribution and its true dimension.
This document summarizes a presentation on the bandit problem and algorithms to solve it. The presentation will:
1) Explain what the bandit problem is and provide a simple example.
2) Describe algorithms for solving the bandit problem, including epsilon-greedy and Thompson sampling.
3) Discuss how to apply bandit algorithms to problems that include contextual information.
1. The document discusses various statistical and neural network-based models for representing words and modeling semantics, including LSI, PLSI, LDA, word2vec, and neural network language models.
2. These models represent words based on their distributional properties and contexts using techniques like matrix factorization, probabilistic modeling, and neural networks to learn vector representations.
3. Recent models like word2vec use neural networks to learn word embeddings that capture linguistic regularities and can be used for tasks like analogy-making and machine translation.
The document discusses distances between data and similarity measures in data analysis. It introduces the concept of distance between data as a quantitative measure of how different two data points are, with smaller distances indicating greater similarity. Distances are useful for tasks like clustering data, detecting anomalies, data recognition, and measuring approximation errors. The most common distance measure, Euclidean distance, is explained for vectors of any dimension using the concept of norm from geometry. Caution is advised when calculating distances between data with differing scales.
1. The document discusses principal component analysis (PCA) and explains how it can be used to determine the "true dimension" of vector data distributions.
2. PCA works by finding orthogonal bases (principal components) that best describe the variance in high-dimensional data, with the first principal component accounting for as much variance as possible.
3. The lengths of the principal component vectors indicate their importance, with longer vectors corresponding to more variance in the data. Analyzing the variances of the principal components can provide insight into the shape of the distribution and its true dimension.
This document summarizes a presentation on the bandit problem and algorithms to solve it. The presentation will:
1) Explain what the bandit problem is and provide a simple example.
2) Describe algorithms for solving the bandit problem, including epsilon-greedy and Thompson sampling.
3) Discuss how to apply bandit algorithms to problems that include contextual information.
- The document discusses formal language theory and a measure-theoretic approach.
- It introduces regular languages and context-free languages. Regular languages have densities that are rational numbers, while it is an open problem whether all context-free languages have densities.
- The talk will cover regular measurable and non-measurable languages, closure properties of the class of regular measurable languages, and challenges.
The document discusses regular expressions and finite automata. It begins by defining regular expressions using operations like concatenation, sum, and star. It then discusses how to interpret regular expressions by defining the language they represent. The document goes on to discuss how finite automata can also represent languages and how regular expressions and automata are equivalent based on the fundamental theorem proved by Kleene in 1951. It provides examples of converting between regular expressions and automata.
A new technique for proving non regularity based on the measure of a languageRyoma Sin'ya
?
This document summarizes a talk on a new technique for proving non-regularity of languages based on their measure (probability). It introduces the zero-one law for regular languages, which states that a regular language is either almost empty (measure 0) or almost full (measure 1). It then presents the zero lemma, which says that an almost empty language that does not have a forbidden word cannot be regular. This provides a new necessary condition for regularity. Examples of non-regular languages that are proven non-regular using this technique include palindromes, the Dyck language, and the language of prime numbers.
26. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
27. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
葉を左から並べると
(書き換えた結果の)?
語になっている
A = {<,>}
G = ({s}, {(s, <>), (s, ss), (s, <s>)}, s)
s → ss → <s>s → <<>>s → <<>><>s → <>
s
< >
s
s s
< s >
< >
< >
Example:
28. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
葉を左から並べると
(書き換えた結果の)?
語になっている
A = {<,>}
G = ({s}, {(s, <>), (s, ss), (s, <s>)}, s)
s → ss → <s>s → <<>>s → <<>><>s → <>
s
< >
s
s s
< s >
< >
< >
Example:
29. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
「構文木を作ること」を構文解析と呼ぶ.
構文解析を行うアルゴリズムが存在する?
→ 語の所属問題が解ける.?
(構文解析をせず語の所属問題を解く?
方法もある: オートマトンなど)
44. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
構文木が複数ある場合も文法によってはありえ
る!!
45. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
構文木が複数ある場合も文法によってはありえ
る!!
A = {<,>}
G = ({s}, {(s, <>), (s, ss), (s, <s>)}, s)
s
ss
< > s
< >
s
< >
s
s
< >
s
s
< >
s
< >
どちらも <><><> の構文木!!
59. 構文木
文法Gと語 w ∈ [G] について, wの(G-)構文木
とは w を生成するGの書き換え規則の有限適
用列を木で表現したもの.
構文木が複数ある場合も文法によってはありえ
る!!
A = {<,>}
G = ({s}, {(s,ε), (s, ss), (s, <s>)}, s)
s
ss
< > s
< >
s
< >
s
s
< >
s
s
< >
s
< >
どちらも <><><> の構文木!!
60. A = {<,>}
G = ({s}, {(s,ε), (s, ss), (s, <s>)}, s)
61. A = {<,>}
G = ({s}, {(s,ε), (s, ss), (s, <s>)}, s)
92. Goldstine言語
L(z)
L ?
3.1.
?
3.7 ([12] ). Goldstine
A = {a, b} G
G = {an1
ban2
b · · · anp
b | p ≥ 1, ni ?= i for some i}.
( )Goldstine G
3.1
Goldstine G A?
1. a (a + b)?
a,
2. b
G′
= {ε, ab, abaab, abaabaaab, · · · }
G = A?
(a + b)?
a G′
G G
G(z) =
1
1 ? 2z
?
z
1 ? 2z
? G′
(z)
=
1 ? z
1 ? 2z
? zn(n+1)/2?1
(14)
3.4 L
L
f, g
f(n) ~ g(n) n → ∞ f(n)/g(n) 1
3.1
3.6 (Puiseux-Transfert). S(z)
S(z) zn
[zn
]S(z)
[zn
]S(z)
[zn
]S(z) ~
αn
ns
Γ(s + 1)
m
i=0
Ciωn
i
?4 [13] Appendix B.1 “Alge-
braic elimination” .
Goldstine G A?
1. a (a + b)?
a,
2. b
G′
= {ε, ab, abaab, abaabaaab, · · · }
G = A?
(a + b)?
a G′
G G
G(z) =
1
1 ? 2z
?
z
1 ? 2z
? G′
(z)
=
1 ? z
1 ? 2z
?
n≥1
zn(n+1)/2?1
(14)
(14)
G(z)′
= n≥1
zn(n+1)/2?1
|z| = 1
(natural boundary) |z| = 1
(G′
(z) )
G
?
93. 1
— —
trices with coe?cients in suitable algebras.
— Jacques Sakarovitch
1. (3 )
2. (4 )
3. (5 )
Webにてサーベイ論文が公開中
https://www.jstage.jst.go.jp/article/jssst/34/3/34_3_3/_article/-char/ja/