[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-DHU2Eni9-1627034233353)(definition.assets/image-20210127213811502.png)]

T:Believe the Turing Test

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-9wrflRIi-1627034233354)(definition.assets/image-20210219094652244.png)]

F: DFS is quicker than BFS

(False, depend on data)

F: heuristic is admissible function, but A* proven to be optimal .admissible algorithm(可采纳算法。admission容许。admissible 不能形容algorithm，只能形容heuristic function？ if h(i) is admissible, then A* is an optimal algorithm )

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-rmB8L8Bi-1627034233356)(definition.assets/image-20210219094801670.png)]

这样做题最稳妥

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-5Asp6VTw-1627034233358)(definition.assets/image-20210219110410164.png)]

这个题目可能不会考，直接T/F考你

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-lJmBf4dH-1627034233360)(definition.assets/image-20210221143627803.png)]

imply == ‘=>’ S1☀️ s2 beach

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-BlDuwFCk-1627034233364)(definition.assets/image-20210221153540889.png)]

用inference rules去证明某式,必须知道用了什么rules，rule名字

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-dGfPNGkR-1627034233368)(definition.assets/image-20210225115111878.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-h3ozJIxg-1627034233370)(definition.assets/image-20210224204836810.png)]

== if we will go swimming, it is sunny,

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-XuU3H6Pq-1627034233372)(definition.assets/image-20210225110812582.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-6EFQG7jc-1627034233374)(definition.assets/image-20210225113722027.png)]

Midterm1复习：PPT至09-10（FIO）

PPT01 AI/turing test/agent/environment

1 What’s Al? Systems Think\Ack like Human\Rational.

2 Rational behavior: Doing the right thing== which is expected to maximize the expected return

3 Most general view of AI: Correct inference/uncertainty handle/resource limitation consideration/cognitive skill

4 Act humanly: The turing test(Problems: 1.not reproducible, constructive, and amenable to mathematics analysis 2. need physical interaction with interrogator and environment)

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-LwsBQdNU-1627034233376)(definition.assets/image-20210225160031694.png)]

5 Agent (perceiving its environment through sensors and acting upon that environment through its effectors to maximize progress towards its goals)==a tool for analyzing systems

6 rational action: the action that maximizes the expected value of the performance measure given the percept sequence to date.==Best/optimal

7 agent should be scalable(Perception(sequence) to Action Mapping. f: P* -> A)specifies which actions an agent ought to take at any point in time

not intelligence(not scalable because distance situation is huge)

8 environment type(1 accessible/inaccessible: sensors give access to complete state of environment

2 deterministic/nondeterministic: the next state can be determined based on the current state and the action==not random thing (Deterministic vs. stochastic

If the next state of the environment is completely determined by the current state and the action executed by the agent, then we say the environment is deterministic; otherwise, it is stochastic（随机）
In principle, an agent need not worry about uncertainty in a fully observable, deterministic environment.)

In a deterministic environment any action has a single guaranteed effect , and no failure or
uncertainty . In a non-deterministic environment , the same task performed twice may produce different results or may even fail completely.

3 episodic/non-episodic(sequential) episodic: each perceive and action pairs。IF STAIC MUST EPISODIC

the next episode does not depend on the actions taken in previous episodes 前后并无关联
In sequential environments, on the other hand, the current decision could affect all future decisions.例子：chess and taxi

the quality of action does not depend on the previous episodic and does not effect the next episode – non-episodic: chess

4 static/dynamic (environment changes during deliberation)

5 discrete/continuous(chess vs driving) state == a bunch of deliverable valuables

**9 agent type: **

Reflex: These agents select actions on the basis of the current percept, ignoring the rest of the percept history( just focus on current input to choose action)

State-based:(also consider history data)

goal-based :get maximize progress about goal

utility-based: how well the goal be achieved(other features considered)

learning: learn from mistakes

PPT02-03Problem solving and search Uninformed search

**1 problem type: **

1 single-state problem(ac;de–eg:chess;agent knows everything, can calculate optimal action sequence to reach goal state.)

2 multiple-state problem(inacc;de–walking in dark room;agent doesn’t know the exact state;agent must reason about sequences of the actions and states assumed while working towards goal state)

3 contingency problem(nonde;inac–skater;must use sensors during execution,solution is a tree or policy;)

4 exploration problem(unknown state space–Maze)

2

3 BFS enqueue node to back of queue(FIFO)

Completeness: Y if b is finite

Time complexity: 1+b+b^d=O(bd)

Space complexity: O(b^d)

Optimality: Y(assuming per step cost =1 else not)

4 DFS enqueue node to the front of queue(LIFO)==stack

Completeness: No, fails in infinite state-space(yes if finite state space)

Time complexity: O(b^m) max depth

Space complexity: O(b*d)

Optimality: No

4.1 Depth-limited search

**Implementation: ** Nodes at depth l have no successors.

Complete: Y if cutoff chosen appropriately then it is guaranteed to find a solution.

Optimal: N it does not guarantee to find the least-cost solution

4.2 Iterative deepening search

complete :Y

TIME :b^d

space:bd

optimality: Y, if step cost=1

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-4KqPo9MP-1627034233379)(definition.assets/image-20210222145307081.png)]

先根，再左，再右

5USC

Uniform enqueue node queue is ordered by path cost(from root to node) A*just add h(n)

A refinement of the breadth-first strategy:

Breadth-first = uniform-cost with path cost = node depth

Completeness: Y if step cost>0

Time complexity: <=O(b^d) depth of least-cost solution

Space complexity: <=O(b^d)

Optimality: Y (path cost never decrease)

no need to write several goals, just replace it cost

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Bg7IFsYc-1627034233387)(definition.assets/image-20210222001701965.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-hinjJHRI-1627034233392)(definition.assets/image-20210222002206039.png)]

6 Bidirectional search

complete: Yes

time: b^(d/2)

space:b^(m/2)

optimality:Yes

summary

DFS will use m, others just d

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Ycbm0wrG-1627034233395)(definition.assets/image-20210222151028727.png)]

problem formulation usually requires abstracting away real-world details to define a state space that can be explored using computer algorithm

once problem is formulated in abstract form, complexity analysis helps us picking out best algorithm to solve problem

variety of uninformed search strategies; difference lies in method used to pick node that will be further expanded

Iterative deepening search only uses linear space and not much more time than other uniformed search strategies

补充 79-81

Complexity: O() and o() measures

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-oDVElUXc-1627034233400)(definition.assets/image-20210225191349468.png)]

when is o(g), it means f is fewer than g much in normal quantification.

PPT 04-05Informed search

Best-first search(greedy search/A* search)

expand most desirable unexpanded node.

1 Greedy search

h(n)= estimate of cost from n to goal()

complete: no(can get stuck in loops)

time: b^m a good heuristic can improve much

space:b^m keep all nodes in memory

optimal: no

2 A search*

f(n) = g(n)+h(n)

h(n)<h* (n) h*(n) ture cost from n to goal

A* is optimal(is also optimal if h(n) is consistent)

A consistent heuristic is thus also always admissible.

PATH COST SO FAR+ESTIMATED PATH COST TO GOAL

393-413-415-417

complete: Y, unless infinitely many node with f<=f(G)

Time: exponential in [(relative error in h)*(length of solution)]

space: keeps all nodes in memory

optimal: yes

Proof:

f(n) is increasing when approaching to goal.()

f’(n) = max(g’(n)+h’(n),f(n))

3 hill climbing

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-V1sZkYt5-1627034233403)(definition.assets/image-20210222174707107.png)]4

4 Simulated annealing algorithm

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-QGz1LGSn-1627034233404)(definition.assets/image-20210222180809402.png)]

Genetic algorithm

补充：

27（how to design admissible heuristic）

•Admissible heuristics can be derived from the exact solution cost of a relaxed version of the problem.

PPT06-07Game playing

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Dp8KufpZ-1627034233407)(definition.assets/image-20210225170829057.png)]

what’s kind of game

Abstraction: to describe a game we must capture every relevant aspect of the game

Accessible environments: such games are characterized by perfect information

Search: game-playing then consists of a search through possible game positions

Unpredictable opponent: introduces uncertainty thus game-playing must deal with contingency problems

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-YmVKPq6N-1627034233409)(definition.assets/image-20210225170924265.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-FTheuIBX-1627034233412)(definition.assets/image-20210225171012999.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-YWFvZY4i-1627034233417)(definition.assets/image-20210225171055783.png)]

how to define ef

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-zP6EnbLT-1627034233418)(definition.assets/image-20210225171226563.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-aYeWDIYM-1627034233420)(definition.assets/image-20210225171258831.png)]

run a-B pruning every state time = O(b^(m/2)

a: best choice so far for MAX b: best choice so far for MIN

max if v>=B return v min if v<=a return v #pruning

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Bmb9V74Z-1627034233423)(definition.assets/image-20210225171426276.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-w2GyqPrC-1627034233424)(definition.assets/image-20210225171601110.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-FdYsvdHL-1627034233427)(definition.assets/image-20210225171809654.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-QUIRJBfF-1627034233429)(definition.assets/image-20210226151310183.png)]

PPT 47\49+session5重点！！！

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-iEpHSQIw-1627034233431)(definition.assets/image-20210301175623003.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-aHiank6U-1627034233434)(definition.assets/image-20210301175637546.png)]

PPT08Constraint satisfaction

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-NiSVDBQb-1627034233436)(definition.assets/image-20210225172028004.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-2G7qsqyM-1627034233438)(definition.assets/image-20210225172101880.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-m3zP2Npb-1627034233440)(definition.assets/image-20210225172237511.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-KxjZxk9e-1627034233443)(definition.assets/image-20210225172330712.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-bsvePFJv-1627034233445)(definition.assets/image-20210226181227923.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-XmhQEiwz-1627034233447)(definition.assets/image-20210225172627535.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-8iw6EDus-1627034233449)(definition.assets/image-20210226181725090.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-XTckRxUH-1627034233452)(definition.assets/image-20210226182036820.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-oNpdeQew-1627034233453)(definition.assets/image-20210225180715709.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-YTX6xlAe-1627034233456)(definition.assets/image-20210226182543595.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-OU1c1f5l-1627034233459)(definition.assets/image-20210226182715252.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Z5bxOgpn-1627034233463)(definition.assets/image-20210226182729517.png)]

重点PPT38

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-3ZC3hoy4-1627034233467)(definition.assets/image-20210226184721153.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-e9slLNlO-1627034233469)(definition.assets/image-20210226184817653.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-vWbxems3-1627034233470)(definition.assets/image-20210226184947624.png)]

PPT09-10Agents that reason logically

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-Eu3Nb563-1627034233473)(definition.assets/image-20210226214426689.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-FPTUCRQX-1627034233474)(definition.assets/image-20210226221240644.png)]

syntax研究的是句法，或者说是语言的结构，这无关乎语言实体的含义。semantics研究的是语义/含义。

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-qEljOOkY-1627034233476)(definition.assets/image-20210226234923384.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-K9wV9sZX-1627034233477)(definition.assets/image-20210226222152706.png)]

definition of entailment、model

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-wBCnhL8G-1627034233479)(definition.assets/image-20210226222622269.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-uDLhktW1-1627034233480)(definition.assets/image-20210226222825932.png)]

soundness（never draw false conclusion/no mistakes）、completeness（it can be proven from KB）

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-VU6Y5HAz-1627034233481)(definition.assets/image-20210227000126275.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-b3mY4YyT-1627034233483)(definition.assets/image-20210226224042410.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-0GxGr3Bh-1627034233484)(definition.assets/image-20210226224102892.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-m0NefiwD-1627034233486)(definition.assets/image-20210226224136181.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-kr59WGET-1627034233488)(definition.assets/image-20210226224225107.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-crbLpvxu-1627034233490)(definition.assets/image-20210226235824120.png)]

考题方式

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-k4nb1Jto-1627034233492)(definition.assets/image-20210226235652336.png)]

or propositional inference

definition of tautologies

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-HUmwlfF3-1627034233498)(definition.assets/image-20210226235908402.png)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-tndytOlJ-1627034233500)(definition.assets/image-20210227000320944.png)]

minmax、alpha-beta、constraint satisfaction、AC3出大题、proof of proposition

-1627034233484)]

[外链图片转存中…(img-m0NefiwD-1627034233486)]

[外链图片转存中…(img-kr59WGET-1627034233488)]

[外链图片转存中…(img-crbLpvxu-1627034233490)]

考题方式

[外链图片转存中…(img-k4nb1Jto-1627034233492)]

or propositional inference

definition of tautologies

[外链图片转存中…(img-HUmwlfF3-1627034233498)]

[外链图片转存中…(img-tndytOlJ-1627034233500)]

minmax、alpha-beta、constraint satisfaction、AC3出大题、proof of proposition

刘顺琦 - CSCI 561 mid 1definition相关推荐

刘顺琦 - CSCI 561 midterm 2 basic knowledge
[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-WPpj6Rfn-1627034161223)(basic%20knowledge.assets/image-202103 ...
刘顺琦 CSCI 561 midterm3
Outline 60% midterm2 后 + 40%midterm2 前,search+game playing+contract satisfaction+logic MDP+DT+NN+BN ...
南京计算机与软件学院,南京信息工程大学计算机与软件学院刘　琦
南京信息工程大学计算机与软件学院刘琦 View 264 Download 2 Embed Size (px) 344 x 292429 x 357514 x 422599 x 487 DESCR ...
刘瑞琦单曲《离开的借口》歌词版MV上线
中新网1月15日电 15日,刘瑞琦2019年全新创作单曲,治愈系伤感情歌<离开的借口>官方歌词版MV正式上线. 受访者供图作为疗愈系情歌歌手代表,刘瑞琦高辨识度的嗓音.鲜明独特的风格.将 ...
刘顺琦 - CSCI561 Related Vocabulary
premise n.前提 on-premise 自有,on-premise private cloud is hosted within an organization. off-premise so ...
仿豆瓣（静态web布局）（全）
静态页面布局页面布局HTML <!DOCTYPE html> <html xmlns="http://w ...
python 中UnicodeEncodeError 错误
在我编写代码处理网页或者是xml格式的文件的时候,运行脚本,出现了下面的错误提示 UnicodeEncodeError: 'ascii' codec can't encode character u' ...
强者愈强！疫情拉大“数据资产”贫富差距，顶级公司数据建设靠什么
5月12日,谷歌子公司Waymo宣布获得7.5亿美元的额外投资,将无人驾驶汽车商业化.两个月内,Waymo的筹款额已经达到了30亿美元. 阿里云计算收入同比增长58%,新零售平台在疫情影响下不减反增. ...
完整约束三（学习笔记）
主外键约束FK FOREIGN KEY(FK) CONSTRAINT fk_外键名 FOREIGN KEY(列名) REFERENCES 表2名(列名)--一般表2的主键先删除 DROP TABLE ...

刘顺琦 - CSCI 561 mid 1definition

Midterm1复习：PPT至09-10（FIO）

PPT01 AI/turing test/agent/environment

PPT02-03Problem solving and search Uninformed search

补充 79-81

PPT 04-05Informed search

补充：

27（how to design admissible heuristic）

PPT06-07Game playing

PPT08Constraint satisfaction

PPT09-10Agents that reason logically

刘顺琦 - CSCI 561 mid 1definition相关推荐

最新文章

热门文章