E-learning Development of Statistics and in Duex: Practical Approaches and Th...Joe Suzuki
?
This document discusses the development of e-learning courses in statistics through the Duex program. Duex is a consortium of Japanese universities and companies focused on data-related human resource development. It produces online statistics and data science courses using a low-cost, high-quality approach involving individual instructors creating video lectures using PowerPoint, scripts, and video editing software. The document outlines Duex's funding and participating institutions, and provides tips for instructors to efficiently create online video courses themselves with minimal budget and assistance from others.
E-learning Design and Development for Data Science in Osaka UniversityJoe Suzuki
?
This document discusses the development of e-learning courses for data science through the Kansai Data related Human Resource Development Consortium (KDC). KDC was established in 2017 with funding from the Japanese Ministry of Education and includes several universities. It aims to develop online statistics courses to make education more accessible and help train data science professionals. The document outlines KDC's goals, challenges in creating high-quality online courses, and strategies for increasing student enrollment and participation over the next five years as funding is scheduled to end.
1. The document proposes a regular quotient score for Bayesian network structure learning that allows for more efficient branch-and-bound search compared to the existing BDeu score.
2. The existing BDeu score violates regularity, meaning that Markov equivalent structures do not necessarily share the same BDeu score.
3. The authors propose a regular quotient score based on Jeffreys' prior that satisfies regularity, ensuring Markov equivalent structures share the score, enabling more efficient searching during branch-and-bound learning of Bayesian network structures.
E-learning Development of Statistics and in Duex: Practical Approaches and Th...Joe Suzuki
?
This document discusses the development of e-learning courses in statistics through the Duex program. Duex is a consortium of Japanese universities and companies focused on data-related human resource development. It produces online statistics and data science courses using a low-cost, high-quality approach involving individual instructors creating video lectures using PowerPoint, scripts, and video editing software. The document outlines Duex's funding and participating institutions, and provides tips for instructors to efficiently create online video courses themselves with minimal budget and assistance from others.
E-learning Design and Development for Data Science in Osaka UniversityJoe Suzuki
?
This document discusses the development of e-learning courses for data science through the Kansai Data related Human Resource Development Consortium (KDC). KDC was established in 2017 with funding from the Japanese Ministry of Education and includes several universities. It aims to develop online statistics courses to make education more accessible and help train data science professionals. The document outlines KDC's goals, challenges in creating high-quality online courses, and strategies for increasing student enrollment and participation over the next five years as funding is scheduled to end.
1. The document proposes a regular quotient score for Bayesian network structure learning that allows for more efficient branch-and-bound search compared to the existing BDeu score.
2. The existing BDeu score violates regularity, meaning that Markov equivalent structures do not necessarily share the same BDeu score.
3. The authors propose a regular quotient score based on Jeffreys' prior that satisfies regularity, ensuring Markov equivalent structures share the score, enabling more efficient searching during branch-and-bound learning of Bayesian network structures.
- The document discusses estimating mutual information and using it to learn forests and Bayesian networks from data. It presents methods for estimating mutual information, finding independence between variables, and using Kruskal's and Chow-Liu algorithms to learn tree structures that approximate joint distributions. Experiments apply these methods to Asia and Alarm datasets to learn Bayesian networks.
This document outlines a two-part course on Bayesian approaches to data compression. Part I on July 17th will cover data compression for known and unknown sources over 90 minutes, including a 45-minute exercise. Part II on July 24th will focus on learning graphical models from data based on the concepts from Part I.
A Conjecture on Strongly Consistent LearningJoe Suzuki
?
1. The document presents a conjecture about the error probability of overestimating the true order k* when learning autoregressive moving average (ARMA) models from samples.
2. The conjecture states that if the estimated order k is greater than the true order k*, the error probability is equal to the probability that a chi-squared distributed random variable with k - k* degrees of freedom is greater than (k - k*)dn, where dn is related to the sample size n.
3. The author provides evidence that a sum of squared estimated ARMA coefficients could be chi-squared distributed, lending credibility to the conjecture.
A Generalization of Nonparametric Estimation and On-Line Prediction for Stati...Joe Suzuki
?
This document presents a generalization of Ryabko's measure for universal coding of stationary ergodic sources. The generalization allows constructing a measure νn that achieves universal coding for sources without a density function, such as those represented by a measure μn on a measurable space. νn is defined by projecting the source onto increasing finer partitions and weighting the projections. If the Kullback-Leibler divergence between the source and weighting measure converges across partitions, νn achieves universal coding for any stationary ergodic source μn. Examples demonstrate how the approach extends Ryabko's histogram weighting to new source types.
Bayesian Criteria based on Universal MeasuresJoe Suzuki
?
The document presents Joe Suzuki's work on generalizing Bayesian criteria to settings beyond discrete or continuous distributions. It introduces generalized density functions based on Radon-Nikodym derivatives that allow defining universal measures gn approximating true densities f. These generalized densities enable extending Bayesian criteria like comparing pgnXgnY to (1-p)gXY to assess independence, to any sample space without assuming a specific form. The approach unifies Bayesian and MDL methods under a framework of universality, with various applications like Bayesian network structure learning.
The Universal Measure for General Sources and its Application to MDL/Bayesian...Joe Suzuki
?
1) The document presents a new theory for universal coding and the MDL principle that is applicable to general sources without assuming discrete or continuous distributions.
2) It constructs a universal measure νn that satisfies certain conditions to allow generalization of universal coding and MDL.
3) This generalized framework is applied to problems that previously separated discrete and continuous cases, such as Markov order estimation using continuous data sequences and mixed discrete-continuous feature selection.
Universal Prediction without assuming either Discrete or ContinuousJoe Suzuki
?
1. The document discusses universal prediction without assuming data is either discrete or continuous. It presents a method to estimate generalized density functions to achieve universal prediction for any unknown probabilistic model.
2. A key insight is that universal prediction can be achieved by estimating the ratio between the true density function and a reference measure, without needing to directly estimate the density function. This allows universal prediction for data that is neither discrete nor continuous.
3. The method involves recursively refining partitions of the sample space to estimate the density ratio. It is shown that this ratio can be estimated universally for any density function, achieving the goal of prediction without assumptions about the data type.
Bayesian network structure estimation based on the Bayesian/MDL criteria when...Joe Suzuki
?
J. Suzuki. ``Bayesian network structure estimation based on the Bayesian/MDL criteria when both discrete and continuous variables are present". IEEE Data Compression Conference, pp. 307-316, Snowbird, Utah, April 2012.
Bayesian network structure estimation based on the Bayesian/MDL criteria when...Joe Suzuki
?
2014 9-26
1. .
.
Klein's fundamental second kind 2-form
for the Cab curves
椳j
昜寄尖
晩云方僥氏2014 定拍
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍1 / 12
2. はじめに
はじめに
旗方爆圧催(協x圭殻塀から竃k)
J. Silverman 縮娩(Brown) 眉屯恕幣鴬平
及15 指屁方サマ`スク`ル(2007、雑偏、寄廉麿)
仝N方の互い旗方爆とAbel 謹悶々
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍2 / 12
3. 眉屯爆
掲蒙旗方爆の協x圭殻塀(眉屯)
F: C 貧の(1 篳旗方) v方悶
O: 肝方1 の恙(o渤h泣)
L := ff 2 FjordQ(f ) 0;Q?= Og [ f0g
M := fordO(f )jf 2 Lg
a1; ; am 2 M: モノイドM の伏撹圷で、(a1; ; am) = 1 となるもの
x1; ; xm 2 L: ordO(xi ) = ai , i = 1; ;m となるもの
C[X1; ; Xm]: C 貧のm 篳謹塀h(X1; ; Xm: 音協猟忖)
.
アフィン旗方爆の協x圭殻塀
.
.徭隼な畠符瞥侏 : C[X1; ; Xm] ! C[x1; ; xm] のker()
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍3 / 12
4. 眉屯爆
Cab 爆
Telescopic ker() の伏撹圷がm 1 ()
ai
di
2
a1
di1
; ;
ai1
di1
, di = (a1; ; ai ), i = 2; ;m
(Suzuki, 2007)
Cab ker() の伏撹圷が1 () m = 2
.
Cab 爆
.
(x1; x2) = (x; y), (a1; a2) = (a; b),
.
C :
Σ
(i ;j)2D
ci ;jxi yj = 0 ; ci ;j 2 C ; D := f(i ; j)ji ; j 0; ai + bj abg
g =
(a 1)(b 1)
2
a = 2, b = 2g + 1 のとき、階卩
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍4 / 12
5. }の協塀晒
Klein's fundamental second kind 2-form
fduigg
i=1: 及1 N裏蛍
C C 貧の2-form R((x; y); (z;w))dxdz で、參和を困燭垢發
1. (z;w) = (x; y)でのみ2了のO
2. lim
(z;w)!(x;y)
R((x; y); (z;w))(x z)2 = 1
R((x; y); (z;w)) =
d
dz
Ω((x; y); (z;w)) +
Σg
i=1
dui (x; y)
dx
dri (z;w)
dz
fdrigg
i=1: O でのみOをもつ及2 N裏蛍
Ω((x; y); (z;w)): 1-form
R((x; y); (z;w)) = R((z;w); (x; y))
を祭磴垢fdrigg
i=1 と、そのときのR((x; y); (z;w))dxdz
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍5 / 12
6. }の協塀晒
階劼栽(Klein, 1888)
y2 = c2g+1x2g+1 + c2g x2g + + c1x + c0 ; c0; ; c2g+1 2 C
dui (x; y) :=
xi1dx
2y
; i = 1; ; g
Ω((x; y); (z;w)) =
y + w
2(x z)y
dx
dri (z;w) =
2Σgi
k=i
ck+1+i (k + 1 i )
zk
2w
dz ; i = 1; ; g
R((x; y); (z;w)) =
Σg
j=0 xj zjfc2j+1(x + z) + 2c2jg
(x z)2
1
2y
1
2w
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍6 / 12
7. }の協塀晒
Cab の栽(Nakayashiki, 2010)
C :
Σ
(i ;j)2D
ci ;jxi yj = 0 ; D = f(i ; j)ji ; j 0; ai + bj abg
fj :=
Σ
i :(i ;j)2D
ci ;jxi ; gj :=
Σ
i :(i ;j)2D
ci ;j zi ; hj :=
Σj1
i=0
wi yj1i
dui ;j (x; y) =
xi yj
Σa
k=1 kyk1fk
dx ; i = 1; ; g
Ω((x; y); (z;w)) =
Σa
j=0 hjgj
(x z)
Σa
k=1 yk1fk
dx
R((x; y); (z;w)) =
d
dz
Ω((x; y); (z;w)) +
Σg
i=1
dui (x; y)
dx
dri (z;w)
dz
が各となるdri ;j (z;w) =
Σ
u;v Di ;j ;u;v zuwv
Σa
k=1 kwk1gk
dz のS方fDi ;j ;u;v g が贋壓
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍7 / 12
8. 麼Y惚
麼Y惚
光(i ; j) 2 J(a; b) := f(i ; j)ji ; j 0; ai + bj ab a bg について、
dri ;j (z;w) :=
Σj+1
u=0
Σa
v=j+1
Σ
r
Σ
s
cr ;ucs;vDr ;s;u;v (i ; j)zr+si2wu+vj2
Σa
k=1 kwk1gk
dz
Dr ;s;u;v (i ; j) :=
8
:
u(s i 1) (u v = j + 1)
u(r i 1) (j + 1 = u v)
(j + 1 u)s (v j 1)r
+(i + 1)(v u) (u j ; j + 2 v)
とおくとき、R((x; y); (z;w)) が各となる。
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍8 / 12
9. 麼Y惚
箭1: 階(a = 2, b = 2g + 1)
C : y2 + y
Σg
r=0
cr ;1xr +
2Σg+1
s=0
cs;0xs = 0
dui ;0(x; y) :=
xidx
2y +
Σg
r=0 cr ;1xr ; i = 0; ; g 1
dri ;j (z;w) =
ri ;0(z;w)
2w +
Σg
r=0 cr ;1zr dz として、
ri ;0(z;w) =
Σg
r=i+2
Σg
s=0
cr ;1cs;1(r i 1)zr+si2
+
Σg
r=i+2
cr ;1(r i 1)zri2w
2Σg+1
r=2i+3
cr ;0(r 2i 2)zri2
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab cu晩rve云s方僥氏2014 定拍9 / 12
10. 麼Y惚
箭2: 蒙(cr ;u = 0, u?= 0; a)
C : ya +
Σb
s=0
cs;0xs = 0
dui ;j (x; y) :=
xi yjdx
aya1 ; (i ; j) 2 J(a; b)
dri ;j (z;w) =
ri ;j (z;w)
awa1 dz として、
ri ;j (z;w) =
Σb
r=i+2
cr ;0(ar a r ai rj)zr2iwa2j
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab c晩urv云es方僥氏2014 定拍10 / 12
11. 麼Y惚
vB冩梢
Cab またはそれ參貧の匯違晒(S方は箔めない)
Cab Nakayashiki 2010
Telescopic c勸丐t鴬平猟2013
階厰塒發eの爆
C3;4 Elibeck, Matsutani, Onishi 麿(2007)
C3;7;8, C6;13;14;15;16 Matsutani (2013)
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab c晩urv云es方僥氏2014 定拍11 / 12
12. まとめ
まとめ
.
Klein's fundamental second kind 2-form
.
.Klein 參栖、126 定ぶりの匯違晒
鮄:
巓豚佩双M、たとえば v方(u;M)(u の謹塀) のS方を麻
辛e蛍狼
旗方爆圧催
.
書瘁のn}
.
.匯違議な眉屯爆 (匯違議な]Riemann 中に) への匯違晒
椳j(昜寄尖) Klein's fundamental second kind 2-form for the Cab c晩urv云es方僥氏2014 定拍12 / 12