Presentation	2019-03-07 On the possibility of learning the order relation of familial resemblance clusters based on Information theoretic scale Tetsuya Morizumi,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	In this paper, we present an approach to incorporate the viewpoint of into an artificial intelligent system. To the question "How each mind interprets others'", “family resemblance” in Wittgenstein's language game and concepts of “Riken-no-ken (sight outside of sight)” used in Ze’ami’s treatise on Noh are designed by probabilistic models .Those who exist as stochastic variables appear as family resemblance. But on the other hand, its existence must be reconstructed from the pre-order text group to the total order cluster. The requirements for this are as follows: "The process of learning the probability distribution of the random variable that becomes the boundary of the familial resemblance clusters of texts", that is, "the process of generating a set of deterministic texts that satisfies the axiom of order relation" is stochastically analyzed . It extracts rules of order relation satisfying the familial resemblance from a large amount of text. That is, by analyzing clusters of texts having partial-order relation and total-order relationship property, the order relation of decision world is derived from stochastic world. Family similarity of probability distribution is evaluated by order relation of information theoretic scale. In this paper, we show a graphical model of Supervised Bayesian Inverse Reinforcement Learning (S - BIRL) using supervised data. S - BIRL is a model that interprets the potential random variable of Blei 's supervised LDA as a potential parameter of reinforcement learning actions and interprets Ramage' s Labeled LDA as a label of reinforcement learning state S.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	Artificial intelligence / probabilistic model / supervised inverse reinforcement learning / supervised LDA / pre-order / partial-order / family resemblance / Wittgenstein
Paper #	SITE2018-76,IA2018-68
Date of Issue	2019-02-28 (SITE, IA)

Conference Information
Committee	IA / SITE / IPSJ-IOT
Conference Date	2019/3/7(2days)
Place (in Japanese)	(See Japanese page)
Place (in English)	Grand XIV Naruto
Topics (in Japanese)	(See Japanese page)
Topics (in English)	Internet and Information Ethics Education, etc.
Chair	Katsuyoshi Iida(Hokkaido Univ.) / Tetsuya Morizumi(Kanagawa Univ.)
Vice Chair	Rei Atarashi(IIJ) / Hiroyuki Osaki(Kwansei Gakuin Univ.) / Toru Kondo(Hiroshima Univ.) / Masaru Ogawa(Kobe Gakuin Univ.) / Takushi Otani(Kibi International Univ.)
Secretary	Rei Atarashi(Tokyo Metropolitan Univ.) / Hiroyuki Osaki(TOYOTA-IT) / Toru Kondo(NEC) / Masaru Ogawa(Tokyo Health Care Univ.) / Takushi Otani(Toyo Eiwa Univ.)
Assistant	Kenji Ohira(Tokushima Univ.) / Ryohei Banno(Tokyo Inst. of Tech.) / Hisanori Kato(KDDI Research) / Nobuyuki Yoshinaga(Yamaguchi Pref Univ.) / Daisuke Suzuki(Hokuriku Univ.)

Paper Information
Registration To	Technical Committee on Internet Architecture / Technical Committee on Social Implications of Technology and Information Ethics / Special Interest Group on Internet and Operation Technology
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	On the possibility of learning the order relation of familial resemblance clusters based on Information theoretic scale
Sub Title (in English)	What is mean by "Defining the rewards of Bayesian inverse reinforcement learning as family resenblence"?
Keyword(1)	Artificial intelligence
Keyword(2)	probabilistic model
Keyword(3)	supervised inverse reinforcement learning
Keyword(4)	supervised LDA
Keyword(5)	pre-order
Keyword(6)	partial-order
Keyword(7)	family resemblance
Keyword(8)	Wittgenstein
1st Author's Name	Tetsuya Morizumi
1st Author's Affiliation	Kanagawa University(KU)
Date	2019-03-07
Paper #	SITE2018-76,IA2018-68
Volume (vol)	vol.118
Number (no)	SITE-480,IA-481
Page	pp.pp.133-140(SITE), pp.133-140(IA),
#Pages	8
Date of Issue	2019-02-28 (SITE, IA)