Unit 8 16 Classical Planning 1 knowitvideos English Japanese

Unit 8 16 Classical Planning 1YouTube↗ IA↗(long shot)
Howto & Style 5.7K views 20 likes 0 dislikes 31 Oct 2011

knowitvideos 11.8K subscribers N/A YouTube↗

Click to Play Video

[Norvig] I want to describe a notation which we call classical planning,

古典的プランニングについてお話ししましょう

which is a representation language for dealing with states and actions and plans,

これは状態、行為、プランを扱う表現言語であり

and it's also an approach for dealing with the problem of complexity

世界を変数に分解することによって

by factoring the world into variables.

複雑な問題を扱うアプローチ方法でもあります

So under classical planning, a state space consists of all the possible assignments

古典的プランニングでは状態空間を構成するすべての可能な値は

to k Boolean variables.

k個の二値変数で表現することができます

So that means they'll be 2 to the k states in that state space.

つまりこの状態空間は2のk乗で表されます

And if we think about the 2 location vacuum world,

2つの場所を持つ掃除機の世界の場合は

we would have 3 Boolean variables.

3つの二値変数を持つことになります

We could have dirt in location A, dirt in location B, and vacuum in location A.

例えばAにゴミがあるか、Bにゴミがあるか、そしてAに掃除機があるかです

The vacuum has to be in either A or B.

掃除機はAまたはBのどちらかになければなりません

So these 3 variables will do, and there will be 8 possible states in that world,

この場合3つの変数が取る可能な世界状態は8つありますが

but they can be succinctly represented through the 3 variables.

それをこの3つの変数で簡潔に表すことができるのです

And then a world state consists of a complete assignment of true or false

世界状態は3つの変数それぞれの真偽値の

through each of the 3 variables.

完全割り当てで構成されます

And then a belief state.

続いて信念状態について説明します

Just as in problem solving, the belief state depends on

問題解決における信念状態は

what type of environment you want to deal with.

扱う環境のタイプに依存します

In the core classical planning, the belief state had to be a complete assignment,

コアな古典的プランニングでは信念状態は完全割り当てでなければならず

and that was useful for dealing with deterministic fully observable domains.

これは決定論的で観察可能な定義域を扱うのに役立ちました

But we can easily extend classical planning,

しかし古典的なプランニングは容易に発展させることができ

and we can deal with belief states that are partial assignments--

部分的な割り当てつまり一部の変数は値を持ち

that is, some of the variables have values and others don't.

その他の変数は値を持たないという信念状態を扱うこともできます

So we could have the belief state consisting of vacuum in A is true

ですからAに掃除機がある信念状態を真とし

and the others are unknown, and that small formula represents 4 possible world states.

その他は未知とするとこれらの簡単な式が 4つの世界状態を表現するのです

We can even have a belief state which is an arbitrary formula in Boolean logic,

またブール論理では任意の式を取る信念状態を持つこともでき

and that can represent anything we want.

発展させることができます

So that's what states look like.

ここまでが状態についての説明です

Now we have to figure out what actions look like

次に“行為”がどんなものかということと

and what the results of those actions look like.

その行為の“結果”について学びます

These are represented in classical planning by something called an action schema.

古典的プランニングではそれを “行為スキーマ”で表現します

It's called a schema because it represents many possible actions that are similar to each other.

“スキーマ”は多くの似たような可能な行為を意味しています

So let's take an example of we want to send cargo around the world,

例えば世界中に貨物を送るとします

and we've got a bunch of planes in airports, and we have cargo and so on.

空港にはたくさんの飛行機と貨物があります

I'll show you the action for having a plane fly from one location to another.

ではここである地点から別の地点へ飛行機を飛ばす時の行為を紹介しましょう

Here's one possible representation.

まず1つは次のように書きます

We say it's an action schema, so we write the word Action

これは行為スキーマですから Actionという単語を書き

and then we write the action operator and its arguments,

続けて行為のオペレータとその引数を書きます

so it's a Fly of P from X to Y.

Fly （p，x，y）です

And then we list the preconditions,

この行為を実行するために

what needs to be true in order to be able to execute this action.

真でなければならない前提条件を列挙します

We can say something like P better be a plane.

pは飛行機であることが望ましいですね

It's no good trying to fly a truck or a submarine.

トラックや潜水艦を飛ばすわけにはいきません

And we'll use the And formula from Boolean propositional logic.

そしてここにブールの命題論理の “かつ”を意味する記号∧を入れます

X better be an airport.

xは空港であることが望ましいです

We don't want to try to take off from my backyard.

裏庭から離陸なんて無茶です

And similarly, Y better be an airport.

同様にyも空港であることが望ましいでしょう

And, most importantly, P better be at airport X in order to take off from there.

そして最も重要なのは飛行機pが離陸のために空港xにあることです

And then we represent the effects of the action by saying

そして次にこの行為からどのような効果が得られるか列挙します

what's going to happen.

つまり行為のあと何が起きるかです

Once we fly from X to Y,

一度xからyへ飛び立てば

the plane is no longer at X,

飛行機はもうxにはありません

so we say not at P,X--the plane is no longer at X--

ですから￢At（p，x）になり飛行機はもうxにはなく

and the plane is now at Y.

∧At（p，y）でyにあります

This is called an action schema.

この式が行為スキーマです

It represents a set of actions for all possible planes, for all X and for all Y,

飛行機、x、yが取り得るすべての一連の行為を

represents all of those actions in one schema

1つのスキーマで表現します

that says what we need to know in order to apply the action and it says what will happen.

行為の実行にあたり知っておくべきことと実行後に何が起こるかを表したものです

In terms of the transition from state spaces, this variable will become false

状態空間からの移行でこの変数は偽になり

and this one will become true.

こちらは真になります

When we look at this formula, this looks like a term in first order logic,

この式は一階述語論理の項のように見えますが

but we're actually dealing with a completely propositional world.

ここで実際に扱っているのは完全な命題論理の世界です

It just looks like that because this is a schema.

これがスキーマなので一階述語論理のように見えるのです

We can apply this schema to specific ground states, specific world states,

このスキーマは特定の基底状態や世界状態に適用することができ

and then P and X would have specific values,

その場合pやxが特定の値を持ちます

and you could just think of it as concatenating their names all together,

そしてこれらの名前はすべて連結され

and that's just the name of one variable.

1つの変数の名前になるのです

The name just happens to have this complex form with parentheses and commas in it

名前はこのように丸括弧とカンマを持つ複雑な形式になりますが

to make it easier to write one schema that covers all the individual fly actions.

これで1つのスキーマに個々の飛行行為を簡潔に書くことができます

Similar videos:

Gundarr (Ep 12): Sausage Party Poopers (subtitulado al castellano)
daniborr
35.1K views 110 likes 10 dislikes 03 Jul 2012100% match

Unit 2, Topic 28, Optimistic Heuristics

17.7K views 26 likes 3 dislikes 05 Oct 2011100% match

Comment voit l'éveillé - 5mn
Selim Aïssel
17K views 172 likes 5 dislikes 31 Jan 2013100% match

Unit 2, Topic 29, State Spaces

16.9K views 4 likes 1 dislikes 05 Oct 2011100% match

Unit 2, Topic 14, Uniform Cost Search

62.2K views 42 likes 24 dislikes 05 Oct 2011100% match

CEOS Chatter
Fernando Garcia
1.3K views 2 likes 1 dislikes 14 Mar 2012100% match

Unit 3 13d D-Separation General Definition

20.6K views 75 likes 5 dislikes 08 Oct 2011100% match

[ENG SUB] 110811 Mnet Wide Super Junior Interview Part 3/3
yingxuan0806
38.7K views 332 likes 1 dislikes 16 Oct 2011100% match

Mr. Peabody & Sherman
TheCaptionedInternet
43.6K views 74 likes 9 dislikes 03 Mar 2014100% match

Unit 3 13b D-Separation Example

11.3K views 4 likes 1 dislikes 08 Oct 2011100% match

Instrucciones para llorar, de Julio Cortázar
Inés Andrés
5.9K views 10 likes 0 dislikes 10 Jan 2012100% match

Unit 3 2 Probabilities

19.5K views 10 likes 0 dislikes 08 Oct 2011100% match

Unit 3 8 More Complex Bayes Networks

13K views 26 likes 3 dislikes 08 Oct 2011100% match

ganar habito-vince lombardi
Josito H
15.9K views 62 likes 0 dislikes 04 Mar 2012100% match

0 views 0 likes 0 dislikes 09 Jun 2026100% match

Unit 8 4 Vacuum Cleaner Example

15.1K views 46 likes 4 dislikes 31 Oct 2011100% match

Recomendación kuentalibros - pizarra digital y niños con discapacidad auditiva
anitamc20
933 views 1 likes 0 dislikes 21 Dec 2012100% match

(ENG/CHI SUB) 140216 SM True Thailand Facebook EXO
Baby Xiumin
783 views 10 likes 0 dislikes 16 Feb 2014100% match

Unit 3 10 Different Type of Bayes Network

23.2K views 24 likes 4 dislikes 08 Oct 2011100% match

Anonymní laskavost se šíří internetem (A stranger's act of kindness goes viral)
Michal Skala
21.7K views 4 likes 0 dislikes 20 Dec 2013100% match

0 views 0 likes 0 dislikes 09 Jun 2026100% match

Unit 3 7 Bayes Rule

17.5K views 35 likes 4 dislikes 08 Oct 2011100% match

Unit 2, Topic 27, A-Star Fifth Question

22.4K views 32 likes 7 dislikes 05 Oct 2011100% match

Unit 3 11 Explaining Away

23K views 32 likes 2 dislikes 08 Oct 2011100% match

Oh ja Ah matkal
Tuuli Koitjärv
21.5K views 30 likes 4 dislikes 01 Jan 2014100% match

Gundarr (Ep 9): Meals on Heels (subtitulado al castellano)
daniborr
11.1K views 33 likes 5 dislikes 01 Jul 2012100% match

Unit 1w, 12 Chinese Translation Answer

27.3K views 16 likes 1 dislikes 05 Oct 2011100% match

TED & TEDx explained
Kyo young Chu
1.1K views 7 likes 0 dislikes 02 Apr 2012100% match

DokuWiki 02 - Creating Links and Pages
Andreas Gohr
12.1K views 22 likes 5 dislikes 25 May 2014100% match

Unit 1w, 4 Terminology

75.1K views 114 likes 4 dislikes 05 Oct 201150% match

Settings Privacy About Contact Donate