Part I. Basics: Chapter 1. Introduction

Quizzes for Chapter 1

1、Turing Test is designed to provide what kind of satisfactory operational definition? 图灵测试旨在给予哪一种令人满意的操作定义
    A、machine intelligence 机器智能
    B、human intelligence 人工智能
    C、machine action 机器动作
    D、human action 人类动作
    E、human thinking 人类思考

2、Select the following true statements regarding the concept of artificial intelligence 选择以下关于人工智能概念的正确表述
    A、Artificial intelligence is the intelligence exhibited by machines or software. 人工智能是通过机器或软件展现的智能
    B、Artificial intelligence aims to create intelligent machines. 人工智能旨在创造智能机器
    C、Artificial intelligence defines itself as the study of human agents. 人工智能将其定义为人类智能体的研究
    D、Artificial intelligence is to enable the development of computers that are able to do things normally done by human. 人工智能是为了开发一类计算机使之能够完成通常由人类所能做的事
    E、Artificial intelligence is to study and construct agent programs that perform well in a given environment. 人工智能是研究和构建在给定环境下表现良好的智能体程序

3、Which of the following are the foundations of artificial intelligence? 如下学科哪些是人工智能的基础?
    A、Philosophy 哲学
    B、History 历史
    C、Mathematics 数学
    D、Medicine 医学
    E、Economics 经济学
    F、Psychology 心理学

4、Which of the following statements are true to describe strong AI(General AI)? 下列陈述中哪些是描述强AI(通用AI)的正确答案?
    A、The appropriately programmed computer with the right inputs and outputs would thereby have a mind in exactly the same sense human beings have minds. 是经过适当编程的具有正确输入和输出的计算机,因此有与人类同样判断力的头脑
    B、It defines non-sentient computer intelligence or AI that is focused on one narrow task. 其定义为无知觉的计算机智能,或专注于一个狭窄任务的AI
    C、It means a machine with the ability to apply intelligence to any problem. 指的是一种机器,具有将智能应用于任何问题的能力
    D、It means a machine with the ability to apply intelligence to a specific problem. 指的是一种机器,仅针对一个具体问题

5、Select the following computer systems which are instances of artificial intelligence. 选择下列计算机系统中属于人工智能的实例
    A、Robot 机器人
    B、Intelligent personal assistant 智能个人助理
    C、Voice-activated telephone menus 声控电话菜单
    D、Web search engines Web搜索引擎
    E、Distributed computing 超市条形码扫描器

6、Select which of the following ones are research area of artificial intelligence. 选择下列哪些是人工智能的研究领域
    A、Distributed computing 分布式计算
    B、Expert system 专家系统
    C、Software outsourcing 软件外包
    D、Machine learning 机器学习
    E、Image understanding 图像理解
    F、Face recognition 人脸识别

7、Examine some applicationsof artificial intelligence (AI) to discover which the following tasks can currently be solved by AI. 考察人工智能(AI)的一些应用,去发现目前下列哪些任务可以通过AI来解决
    A、Playing a decent game of table tennis (Ping-Pong). 打一场像样的乒乓球比赛
    B、Buying a week’s worth of groceries at the market. 在市场上购买一周的食品杂货
    C、Buying a week’s worth of groceries on the Web. 在Web上购买一周的食品杂货
    D、Playing the game of Texas Holdem Poker at a competitive level. 以竞技水平玩德州扑克游戏
    E、Translating spoken English into spoken Chinese in real time. 将英文口语实时翻译为中文口语

8、Rationally means the property of a systemwhich does the right thing given what it ________. 理性指的是一个系统的属性,即在_________的环境下做正确的事。

Part I. Basics: Chapter 2. Intelligent Agent

Quizzes for Chapter 2

1、For the following comparisons, select which approaches are more general and more amenable to scientific development. 对于如下比拟,选择哪些方法更通用并且更适合于科学开发
    A、Thinking humanly 类人思考
    B、Thinking rationally 理性思考
    C、Acting humanly 类人动作
    D、Acting rationally 理性动作

2、A rational agent is one that does the right thing, but what does it mean to do the right thing? Considering the consequences of the agent’s behavior, select the following true statements regarding the concept of "right thing". 理性智能体指的是有正确行为的智能体,但正确的行为意味着什么?考虑智能体行为的后果,选择下列有关“正确的行为”概念的真实陈述
    A、This sequence of actions causes the environment to go through a sequence of states. 这些动作序列引起环境改变而产生状态序列
    B、An agent in an environment generates a sequence of actions according to the percepts. 智能体在某个环境中依据感知生成动作序列
    C、If the sequence is desirable, then the agent has performed well. 如果该序列是所期望的,则该智能体表现良好
    D、But if the sequence is undesirable, then the agent destroys the environment. 但若该序列是不期望的,则该智能体对其环境造成破坏。

3、PEAS is a task environment description which stands for: PEAS是一种任务环境描述,它代表:
    A、Performance 性能
    B、Perceive 感知
    C、Environment 环境
    D、Exclusion 排斥
    E、Actuators 动作器
    F、Actions 动作
    G、Sense 感觉
    H、Sensors 感受器

4、Which of the following statements are true about environment types for intelligent agents? 下列中语句哪些与智能体环境类型有关
    A、Fully observable vs. partially observable 完全可观测与部分可观测
    B、Single agent vs. multi-agent 单智能体与多智能体
    C、Good agent vs. bad agent 好的智能体与差的智能体
    D、Episodic vs. sequential 阵发性与连续性
    E、Deterministic vs. stochastic 确定性与随机性
    F、Solidvs vs. liquid 固体与液体
    G、Perfect vs. imperfect 完美与不完美

5、Thinking the differences between gent functions and agent programs, select correct statements from following ones. 考虑智能体函数与智能体程序的差异,从下列陈述中选择正确的答案
    A、An agent program implements an agent function 一个智能体程序实现一个智能体函数
    B、An agent function implements an agent program. 一个智能体函数实现一个智能体程序
    C、An agent program has no any relation with an agent function. 智能体程序与智能体函数没有任何关系
    D、An agent program contains agent function. 智能体程序包含智能体函数

6、Selectthe following true ways that can be used to represent states for an agent. 选择下列哪些方法可以用于表示智能体的状态
    A、Atomic 原子式
    B、Distributed 分布式
    C、Factored 因子式
    D、Networked 网络式
    E、Molecular 模块式
    F、Structured 结构式

7、The approach of artificial intelligence (AI) which epresents information through symbols and their relationships is _______ AI, and which represents information in a distributed form within a network is Connectionist AI.凭借符号及他们之间的关系来表征信息的人工智能方法是____________AI,而用网络内部的一种分布式形式来表征信息的是联结主义AI。

8、An intelligent agent is anything that can be viewed as perceiving its environment through _______, and acting upon that environment through actuators. 智能体可看作是通过_________感知外部环境,并且通过执行器作用于外部环境的任何事物。

Part II. Searching: Chapter 3. Solving Problems by Search

Quizzes for Chapter 3

1、There are two main kinds of formulation for 8-queens problem. Which of the following items is such a formulation that starts with all 8 queens on the board and moves them around? 有两种8皇后问题的形式化方式。“初始时8个皇后都放在棋盘上,然后再进行移动”是哪一种形式化方式?
    A、Incremental formulation 增量形式化
    B、Normal formulation 正则形式化
    C、Complete-state formulation 全态形式化
    D、Half-state formulationon 半态形式化

2、Let h(n) denote heuristic function and g(n) denote cost, the evaluation function f(n) used by Greedy search is: 设h(n)表示启发式函数且g(n)表示代价,则贪婪搜索所使用的评价函数是
    A、f(n) = h(n)
    B、f(n) = g(n)
    C、f(n) = g(n) + h(n)
    D、f(n) = g(n) - h(n)

3、Let h(n) denote heuristic function and g(n) denote cost, the evaluation function f(n) used by A* search is: 设h(n)表示启发式函数且g(n)表示代价,则A*搜索所使用的评价函数是:
    A、f(n) = h(n)
    B、f(n) = g(n)
    C、f(n) = g(n) + h(n)
    D、f(n) = g(n) - h(n)

4、Which of the following items are used to formulate a problem? 下列哪些项用于对问题进行形式化
    A、Environment 环境
    B、Initial state 初始状态
    C、Goal state 目标状态
    D、Perception 感知
    E、Actions 动作
    F、Actuator 执行器
    G、Transition model 转换模型
    H、Agent function 智能体函数
    I、Goal test 目标检测
    J、Path cost 路径检测

5、Select which of the following search methods are true uninformed search. 选择如下哪些搜索方法属于无信息搜索
    A、A* Search A*搜索
    B、Best-first Search 最佳优先搜索
    C、Bidirectional search 双向搜索
    D、Breadth-first search 宽度优先搜索
    E、Depth-first search 深度优先搜索
    F、Depth-limited search 深度受限搜索
    G、Greedy Search 贪婪搜索
    H、Uniform-cost search 一致代价搜索

6、The search strategies that use problem-specific knowledge beyond the definition of the problem itself are known as: 除了问题本身的定义之外,使用问题特定知识的搜索策略被认为是
    A、Binary search 二元搜索
    B、Brute force search 蛮力搜索
    C、Informed search 有信息搜索
    D、Uninformed search 无信息搜索
    E、Heuristic search 启发式搜索

7、Some ______ problems or NP-hard problems can be solved only by search. 某些_______或NP难问题只能通过搜索来求解。

8、Problem formulation is the process of deciding what _______ and states to consider, given a goal. 问题形式化是给定一个目标,决定要考虑的_______与状态的处理。

Part II. Searching: Chapter 4. Local Search and Swarm Intelligence

Quizzes for Chapter 4

1、Keeping just one node in memory might seem to be an extreme reaction to the problem of memory limitations. Local beam search keeps track of: 在内存中仅保存一个节点似乎是对内存限制问题的极端反应。局部束搜索保持:
    A、k states rather than just one. k个状态而不仅仅为一
    B、m states rather than n states. m个状态而不是n个
    C、just one rather than k states. 仅有一个状态而不是k个
    D、just m states rather than n states 仅有n状态而不是m

2、A genetic algorithm is a variant of stochastic beam search, in which successor states are generated: 遗传算法是随机束搜索的一个变体,其中后继节点的生成是由:
    A、by combining two parent states rather than by modifying a single state. 组合两个双亲状态而不是修改单一状态。
    B、by modifying a single state rather than by combining two parent states. 组合两个双亲节点而不是修改单一节点。
    C、by combining two parent nodes rather than by modifying a single node. 组合单一状态而不是组合两个双亲状态。
    D、by modifying a single node rather than by combining two parent nodes. 组合单一节点而不是组合两个双亲节点。

3、Hill climbing search is sometimes called greedy local search because it grabs a good neighbor state without thinking ahead about where to go next. Unfortunately, it often gets stuck for the three reasons: 爬山搜索有时也被称为贪婪局部搜索,因为它只顾抓住一个好的邻接点的状态,而不提前思考下一步该去哪儿。它在三种情况下经常被困:
    A、Hill 山丘
    B、Local maxima 局部最大值
    C、Mountain 高山
    D、Plateaux 高原
    E、Ridges 山岭

4、Which of the following statements are true about Simulated annealing algorithm? 以下关于模拟退火算法的陈述哪些是正确的?
    A、The inner loop of the simulated annealing algorithm is quite similar to hill climbing. 模拟退火算法的内循环与爬山法非常相似。
    B、The inner loop of the simulated annealing algorithm is very different from hill climbing. 模拟退火算法的内循环与爬山法完全不同。
    C、Instead of picking the best move, simulated annealing algorithm picks a random move. 模拟退火算法不是选择最佳行动,而是选择随机行动。
    D、Instead of picking the random move, simulated annealing algorithm picks a best move. 模拟退火算法不是选择随机行动,而是选择最佳行动。

5、Ant Colony Optimization (ACO) algorithm was inspired by the behavior of ants seeking a path between _______ and a source of food. 蚁群优化算法是受蚂蚁在_______和食物源之间寻找路径行为的启发而形成的。

6、Inspired by social behavior of birds and fishes, Particle Swarm Optimization (PSO) algorithm uses a number of ________ that constitute a swarm moving around in the search space looking for the best solution. 受鸟类和鱼类的社会行为的启发,粒子群优化算法采用若干_______构成一个围绕搜索空间移动的群体来寻找最优解。

7、Local search algorithms operate using a single ______ (rather than multiple paths) and generally move only to neighbors of that node. 局部搜索算法使用一个______(而不是多条路径),并且通常仅移动到该节点相邻的节点。

8、In addition to finding goals, local search algorithms are useful for solving pure _________, in which the aim is to find the best state according to an objective function. 除了寻找目标之外,局部搜索算法对解决纯_________也很有效。其目的是根据一个目标函数找到其最好的状态。

Part II. Searching: Chapter 5. Adversarial Search

Quizzes for Chapter 5

1、Which of the following is a true statement about games? 以下关于博弈的陈述哪个是正确的?
    A、Adversarial search problems are often known as games. 对抗搜索通常称为博弈
    B、Classical search problems are often known as games. 经典搜索问题通常称为博弈
    C、Heuristic search problems are often known as games. 启发式搜索问题通常称为博弈
    D、Local search problems are often known as games. 局部搜索问题通常称为博弈

2、___________ describes a situation in which the interacting agents' aggregate gains and losses can be less than or more than zero. ___________中智能体交互动作的总收益和损失可以小于或大于零
    A、Computer game 计算机博弈
    B、Non-zero sum game 非零和博弈
    C、Two-player game 双人博弈
    D、Zero sum game 零和博弈

3、Select the following true statements regarding the concept of minimax rule for a zero sum game. 从如下关于零和博弈maximum概念中选择正确的答案。
    A、Each player maximizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最大。
    B、Each player minimizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最小。
    C、Each player maximizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最大。
    D、Each player minimizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最小。

4、Which of the following statements are true about alpha-beta pruning? 以下关于alpha–beta剪枝的陈述哪些是正确的?
    A、Alpha–beta pruning is to decrease the number of nodes that are evaluated by the minimax algorithm in its search tree. Alpha–beta剪枝旨在减少其搜索树中由minimax算法评价的节点数量。
    B、Alpha–beta pruning is to increase the number of nodes that are evaluated by the minimax algorithm in its search tree. Alpha–beta剪枝旨在增加其搜索树中由minimax算法评价的节点数量。
    C、Alpha–beta pruning is to eliminate large parts that are evaluated by the minimax algorithm in its search tree. Alpha–beta剪枝旨在消除其搜索树中由minimax算法评价的大部分。
    D、Alpha–beta pruning is to add large parts that are evaluated by the minimax algorithm in its search tree. Alpha–beta剪枝旨在添加其搜索树中由minimax算法评价的大部分。

5、Claude Shannon proposed instead that programs should cut off the search earlier and apply a ____________evaluttion function to states in the search, effectively turning nonterminal nodes into terminal leaves. 克劳德·香农提出:程序应该早一些剪断搜索,并在搜索中对状态应用_____________评价函数,有效地将非终端节点转换为终端叶节点。 (请填写中文答案)

6、____________ is a dynamic game with probabilistic transitions played by one or more players. ____________是一种具有概率转换的动态博弈,有一个或多个玩家。(请填写中文答案)

7、Monte-Carlo methods are a broad class of computational algorithms that rely on ________________ to obtain numerical results. 蒙特卡罗方法是一大类计算算法,它凭借________________来获得数值结果。(请填写中文答案)

8、___________ tree search is on the analysis of the most promising moves, expanding the search tree based on random sampling of the search space. ___________树搜索对最有利的动作进行分析,根据搜索空间的随机采样来扩展搜索树。(请填写中文答案)

Part II. Searching: Chapter 6. Constraint Satisfaction Problem

Quizzes for Chapter 6

1、Select the following true one that is used to the state representation for constraint satisfaction problems (CSPs). 从如下用于约束满足问题 (CSP)的状态表示中选择正确的答案。
    A、atomic 原子
    B、Distributed 分布
    C、Factored 因子
    D、Networked 网络
    E、Molecular 分子
    F、Structure 结构

2、Assume that {A, B, C, D} are variables, the domain of each variable is {u, v, w}, and != denotes "not equal", select which of the following expressions is a binary constraint on CSP formalism? 设{A, B, C, D}为变量,每个变量的域是{u, v, w},且“!=”表示不等于,从如下表达式中选择那个是CSP形式化的2元约束?
    A、<(A), A = v>
    B、Diff(A, D)
    C、A + B = C
    D、Alldiff(A, B, C, D)

3、Compare CSP and state-space search, and select correct statements from following ones. 比较CSP和状态空间搜索,并从下列叙述中选择正确的答案。
    A、CSP solving system can be faster than state-space search solving system. CSP求解系统会比状态空间搜索求解系统快。
    B、CSP solving system can be slower than state-space search solving system. CSP求解系统会比状态空间搜索求解系统慢。
    C、CSP can quickly eliminate large swatches of the search space. CSP可以快速排除大的搜索空间样本。
    D、State-space search can quickly eliminate large swatches of the search space. 状态空间搜索可以快速排除大的搜索空间样本。

4、Which of the following statements are true types of types of local consistency for constraint propagation? 如下陈述中哪些是约束传播局部一致性的正确类型?
    A、Arc consistency 弧一致
    B、Graph consistency 图一致
    C、k-consistency k一致
    D、Loop consistency 环路一致
    E、Node consistency 节点一致
    F、Path consistency 路径一致
    G、Tree consistency 树一致

5、Select the following true statements regarding the concept of "Backtracking search". 从如下有关“回溯搜索”概念中选择正确的答案。
    A、It chooses values for one variable at a time and backtracks when a variable has no legal values left to assign. 每次为变量选择值并且当变量没有合法赋值时回溯。
    B、It incrementally builds candidates to the solutions, and abandons each partial candidate c, as soon as it determines that c cannot possibly be completed to a valid solution. 每次为变量选择值并且当变量有一个或多个合法赋值时回溯。
    C、It incrementally builds candidates to the solutions, and abandons each partial candidate c, as soon as it determines that c cannot possibly be completed to an invalid solution. 递增地构建解的候选,并且一旦确定部分候选c不能成为合法的解,就将c抛弃。
    D、It chooses values for one variable at a time and backtracks when a variable has one more legal value left to assign. 递增地构建解的候选,并且一旦确定部分候选c不能成为非法的解,就将c抛弃。

6、A constraint satisfaction problem (CSP) is defined as a set of objects whose state must satisfy a number of _______________. 约束满足问题 (CSP) 被定义为其状态必须满足若干_______________的一组对象。

7、In choosing a new value for a variable, min-conflicts heuristic is to select the value that results in the _________ with other variables. 在对一个变量选择一个新值时,最小冲突启发式选择导致与其它变量呈现__________的值。

8、To reduce a constraint graph to a tree structure, two approaches can be used that are ___________and tree decomposition. 为了简化约束图为树结构,有两种方法可以采用,即___________和树分解。

Part III. Reasoning: Chapter 7. Reasoning by Knowledge

Quizzes for Chapter 7

1、Which of the following has been defined with reference to information? 下列中哪一项是根据信息来定义的?
    A、Data 数据
    B、Information 信息
    C、Knowledge 知识
    D、Wisdom 智慧

2、What kind of knowledge will be used to describe how a problem is solved? 哪种知识可用于描述问题如何求解?
    A、Declarative knowledge 陈述性知识
    B、Deep knowledge 深层知识
    C、Dynamic knowledge 动态知识
    D、Heuristic knowledge 启发性知识
    E、Procedural knowledge 过程性知识
    F、Static knowledge 静态知识

3、Select which of the following approaches are true using first order logic. 选择以下哪种方法可以使用一阶逻辑。
    A、Declarative approaches 陈述性方法
    B、Functional approaches 函数型方法
    C、Objective approaches 对象型方法
    D、Procedural approaches 过程性方法

4、Select the following true statements regarding the concept of "semantic network". 选择如下关于“语义网络”概念的正确陈述。
    A、It can represent the semantic relations between concepts. 可以表示概念之间的语义关系。
    B、It can represent large domains, performance or meta-knowledge very well. 可以很好地表示大的域、性能或元知识。
    C、It can be represented as a directed or undirected graph, consisting of nodes used to represent concepts, and arcs to represent semantic relations between the concepts. 可以被表示为一种直接或间接的图,由表示概念的节点和表示概念之间语义关系的弧组成。
    D、It can be used when one has knowledge that is best understood as a set of concepts that are related to one another. 语义网络可以用于当所具有的知识可被很好地理解为一组彼此相关的概念时。

5、Select which of the following statements are true about ontology. 选择下列关于本体的陈述哪些是正确的。
    A、An ontology is a formal naming and definition of the types, properties, and interrelationships of the entities for a particular domain of discourse. 本体是一种对特定论域中实体的类型、特性和相互关系的形式化命名和定义。
    B、An ontology provides a common vocabulary of an area and define the meaning of the terms and the relationships between them. 本体提供一个领域的公共词汇,并且定义一些术语的含义和它们之间的关系。
    C、Ontologies cannot be created to organize information that will be applied to problem solving. 无法创建本体来组织可用于问题求解的信息。
    D、Ontologies cannot be divided the types of upper ontology, domain ontology, or hyper ontology. 无法将本体分为上层本体、领域本体、或混合本体。

6、The synonyms of Bayesian networks include: 贝叶斯网络的同义词包括:
    A、Belief network 信念网络
    B、Causal network 因果网络
    C、Computer network 计算机网络
    D、Neural network 神经网络
    E、Probabilistic network 概率网络
    F、Social network 社交网络

7、A ______________ system consists of a knowledge base and an inference engine, where knowledge base represents facts about the world, inference engine can reason about those facts. ______________系统由知识库和搜索引擎组成,其中,知识库表示关于世界的事实,推理引擎则可以对这些事实进行推理。

8、__________ representation focus on designing computer representations that capture knowledge about the world that can be used to solve complex problems. __________表示关注于设计计算机表示来采集关于世界的知识,可用于解决复杂的问题。

Part IV. Planning: Chapter 8. Classic and Real-world Planning

Quizzes for Chapter 8

1、Learned the differences between single-agent problem and multi-agent problem, select a correct one form following descriptions. 学习了单智能体问题和多智能体问题之间的区别后,从如下描述中选择正确的答案。
    A、Both of multi-effector problem and multi-body problem is a single-agent problem. 多效用器和多躯体这两者都是单智能体问题。
    B、Both of multi-effector problem and multi-body problem is a multi-agent problem. 多效用器和多躯体这两者都是多智能体问题。
    C、Multi-effector problem is a single-agent problem, but multi-body problem is not. 多效用器是单智能体问题,而多躯体是多智能体问题。
    D、Multi-body problem is a multi-agent problem, but multi- effector problem is not. 多躯体是单智能体问题,而多效用器是多智能体问题。

2、Select the following correct features regarding classical planning. 选择如下有关经典规划的正确特征。
    A、Fully observable in a dynamic environment. 动态环境中完全可观测。
    B、Partly observable in a static environment. 静态环境中部分可观测。
    C、Keeping a nondeterministic action at a time. 每次保持不确定性的动作。
    D、Keeping a deterministic action at a time. 每次保持确定性的动作。
    E、One agent with only one known initial state. 一个智能体仅有一个已知的初始状态。
    F、Several agents with multiple known initial states. 若干个智能体具有多个已知的初始状态。

3、Which of the following approaches are used for classical planning? 下列方法中哪些可用于经典规划?
    A、Boolean satisfiability 布尔可满足性
    B、constraint satisfaction 约束满足
    C、first-order logical deduction 一阶逻辑推理
    D、planning graph 规划图
    E、plan refinement 规划精进
    F、state-space search 状态空间搜索

4、Planners used in real world for planning and scheduling are more complex, so that select the following correct things which should be extend: 现实世界中用于规划和调度的规划者更为复杂,因此选择以下哪些应该扩展。
    A、the representation language 表示语言
    B、the planning algorithms 规划算法
    C、the way to interact with environment 与环境交互的方式
    D、the approach used for planning 用于规划的途径

5、The characteristics of multi-agent are: 多智能体的特征是:
    A、Autonomy 自主性
    B、Centralized control 集中控制
    C、Decentralization 分散化
    D、Global views 全局视野

6、To simulate the flocking behavior of birds, the rules should consist of 要仿真鸟群的群体行为,其规则应包括:
    A、Alignment 对齐
    B、Cohesion 聚集
    C、Communication 通信
    D、Separation 分离
    E、Centralization 集中

7、There are two approaches to searching for a plan, one is ________________ search, and the other is backward relevant-states search. 有两种搜索计划的方式,一个是_____________搜索,而另一个是后向状态空间搜索。

8、Markov Decision Process (MDP) is a _______________ process, means action outcomes depend only on the current state. 马可夫决策过程(MDP)是一种_______________过程,意味着动作结果仅仅依赖于当前状态。

9、There are two optimal policies for Dynamic Programming, one is ______________, and the other is policy iteration. 动态规划有两种优化策略,一个是___________,而另一种是策略迭代。

Part V. Learning: Chapter 9. Perspectives about Machine Leaning

Quizzes for Chapter 9

1、Which of the following is used to discover general facts from training examples? 下列中哪个用于从训练样本中发现一般的事实?
    A、Deductive learning 演绎学习
    B、Inductive learning 归纳学习
    C、Reinforcement learning 强化学习
    D、Supervised learning 有监督学习
    E、Unsupervised learning 无监督学习
    F、None of the above is correct 以上都不正确

2、There is a set of Web pages identified as interested or uninterested for the user that can be used to solve Web pages filtering problem by machine learning, select which of the following element/feature is the formal definition for those Web pages. 有一组供用户标识为感兴趣或不感兴趣的网页,可被机器学习用来解决网页过滤问题,选择下列特征中哪个是通过训练这些网页学来的?
    A、Task (T) 任务
    B、Performance (P) 性能
    C、Experience (E) 经验
    D、Examples (E) 样本

3、Which of the following is used for learned features? 以下哪一个是可用于从图像中识别视觉模式的深度神经网络?
    A、CNN CNN
    B、CVPR CVPR
    C、GLOH GLOH
    D、HOG HOG
    E、LBP LBP
    F、LESH LESH
    G、OCR OCR
    H、SIFT SIFT
    I、SURF SURF
    J、SVM SVM

4、Select the following true algorithms (methods) which have been invented in machine learning? 选择下列哪些是机器学习中发明的算法(方法)?
    A、Abstraction 抽象
    B、AdaBoost 自适应提升
    C、Decision Tree 决策树
    D、Generalization 泛化
    E、Perceptron 感知机
    F、Random Forests 随机森林

5、Select the following true parties regarding the three parties of machine learning. 选择如下哪些派别是有关机器学习的三个派别?
    A、Connectionism 联结主义
    B、Behaviorism 行为主义
    C、Logicalism 逻辑主义
    D、Realism 现实主义
    E、Subjectivism 主观主义
    F、Symbolicism 符号主义

6、Select the following true phrases regarding the tasks that have been solved by machine learning. 选择下列哪些属于可以通过机器学习解决的任务。
    A、Classification 分类
    B、Clustering 聚类
    C、Category 派生
    D、Dissection 剖析
    E、Partition 划分
    F、Regression 回归

7、Select the following true phrases regarding the typical paradigms in machine learning. 选择下列哪些是典型的机器学习范式。
    A、Augmented learning 增强学习
    B、Informal learning 非正式学习
    C、Multimedia learning 多媒体学习
    D、Reinforcement learning 强化学习
    E、Supervised learning 有监督学习
    F、Unsupervised learning 无监督学习

8、Machine learning is a ______ of artificial intelligence, concerns the construction and study of systems that can learn from data. 机器学习是人工智能的一个______,从事构建和研究可以从数据中学习的系统。

Part V. Learning: Chapter 10. Tasks in Machine Learning

Quizzes for Chapter 10

1、Which statement best describes the task of “classification” in machine learning? 哪一个是机器学习中“分类”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

2、Which statement best describes the task of “regression” in machine learning? 哪一个是机器学习中“回归”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

3、Which statement best describes the task of “ranking” in machine learning? 哪一个是机器学习中“排名”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

4、Which statement best describes the task of “clustering” in machine learning? 哪一个是机器学习中“聚类”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

5、Which statement best describes the task of “density estimation” in machine learning? 哪一个是机器学习中“密度估计”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

6、Which statement best describes the task of “dimensionality reduction” in machine learning? 哪一个是机器学习中“降维”任务的准确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。

7、The output of regression is a real ______ value, and the output of classification is a discrete categories. 回归的输出是一个实际的______值,而分类的输出则是一个离散的类别。

8、Clustering is to identify similar groups for input objects _____ training data, and classification is to assign pre-defined classes for input items with training data. 聚类是为输入对象标识相似的组,_____训练数据;而分类则是给输入项分派预定义的类,有训练数据。

Part V. Learning: Chapter 11. Paradigms in Machine Learning

Quizzes for Chapter 11

1、Which statement best describes the paradigm of “Supervised learning” in machine learning? 哪一个是机器学习中“有监督学习”范式的准确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

2、Which statement best describes the paradigm of “semi-supervised learning” in machine learning? 哪一个是机器学习中“半监督学习”范式的准确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

3、Which statement best describes the paradigm of “unsupervised learning” in machine learning? 哪一个是机器学习中“无监督学习”范式的准确描述?
    A、It aims to learn some information using a limited amount of training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

4、Which statement best describes the paradigm of “weakly supervised learning” in machine learning? 哪一个是机器学习中“弱监督学习”范式的准确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

5、Which statement best describes the paradigm of “one-shot learning” in machine learning? 哪一个是机器学习中“一次性学习”范式的准确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

6、Which statement best describes the paradigm of “zero-shot learning” in machine learning? 哪一个是机器学习中“零次性学习”范式的准确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。

7、Which learning paradigm does it apply storing knowledge to a different but related problem? 那种学习范式把储备的强知识用于不同但相关的问题?
    A、Active learning 主动式学习
    B、Collaborative learning 协同式学习
    C、Ensemble learning 集成学习
    D、Transfer learning 迁移学习

8、Select the following true algorithms (methods) in which each one has been invented as an algorithm (method) of reinforcement learning? 选择下列哪些算法(方法)已被发明为专门用于强化学习的算法。
    A、Asynchronous Advantage Actor-Critic 异步优势动作者•评论者
    B、Behaviorist Psychology 行为学派心理学
    C、Deep Network 深度网络
    D、Deterministic Policy Gradients 确定性策略梯度
    E、Neural Episodic Control 神经情景控制

Part V. Learning: Chapter 12. Models in Machine Learning

Quizzes for Chapter 12

1、Select the following true words in which each one has been invented as a typical learning model for machine learning. 从如下模型中选择已被用作机器学习的代表性模型。
    A、Distributed model 分布式模型
    B、Geometric model 几何模型
    C、Lifecycle model 生命周期模型
    D、Logical model 逻辑模型
    E、Networked model 网络化模型
    F、Probabilistic model 概率模型

2、Select the following true phrases in which each one has been invented as a probabilistic model for machine learning. 从如下短语中选择已被用作机器学习的概率模型。
    A、Gaussian processes 高斯过程
    B、Hidden Markov models 隐藏马可夫模型
    C、Linear regression 线性回归
    D、Naive Bayes and Bayes Network 朴素贝叶斯和贝叶斯网络
    E、Probabilistic program 概率规划
    F、Support vector machine 支撑向量机

3、Which of the following algorithms are true ones used by manifold learning? 下列哪些是真正被流形学习所使用的算法?
    A、HMM (Hidden Markov models) HMM(隐藏马可夫模型)
    B、Isomap (Isometric mapping) Isomap(等距映射)
    C、LLE (Locally Linear Embeddings) LLE(局部线性嵌入)
    D、LE (Laplacian Eigenmaps) LE(拉普拉斯特征映射)
    E、PCA (Principal Component Analysis) PCA(主成分分析)
    F、SVM (Support Vector Machine) SVM(支撑向量机)

4、Select the following phrases in which each one has truly been invented or used as a logical model for machine learning. 从如下短语中选择真正地被发明为或被用作机器学习逻辑模型的方法。
    A、Adaptive boosting 自适应性提升
    B、Association rules 关联规则
    C、Convolutional neural network 卷积神经网络
    D、Decision tree 决策树
    E、First-order logic 一阶逻辑
    F、Linear regression 线性回归

5、Which of the following phrases are the artificial neural networks truly used in machine Learning? 下列短语哪些是真正用于机器学习的人工神经网络?
    A、Bayes Network 贝叶斯网络
    B、Convolutional neural network 卷积神经网络
    C、Deep auto-encoder 深度自动编码器
    D、Long short-term memory 长短期记忆
    E、Wireless network 无线网络

6、The probabilistic model is to use probability theory to express all ________________. 概率模型是采用概率论来表示所有________________。

7、A geometric model is constructed directly in instance space which can be thought as the ________________ such as Euclidean space or Riemannian space. 几何模型直接在实例空间构建,该空间可被认为是________________,例如:欧几里德空间或黎曼空间。

8、The logical models are defined in terms of easily interpretable ______________, or can be easily translated into rules that are understandable by humans. 逻辑模型被定义为易于解释的______________,或者易于转换成人类能够理解的规则。

9、An artificial neural network is an artificial representation of the ____________ that tries to simulate its learning processing. 人工神经网络是一种____________的人工表示,试图仿真其学习过程。

Final Exam of PoAI

Final Exam of PoAI

1、Turing Test is designed to provide what kind of satisfactory operational definition? 图灵测试旨在给予哪一种令人满意的操作定义?
    A、machine action 机器动作
    B、human action 人类动作
    C、machine thinking 机器思考
    D、human thinking 人类思考
    E、machine intelligence 机器智能
    F、human intelligence 人类智能
    G、None of the above is correct 以上都不正确

2、Thinking the differences between agent functions and agent programs, select correct statements from following ones. 考虑智能体函数与智能体程序的差异,从下列陈述中选择正确的答案。
    A、An agent program implements an agent function. 一个智能体程序实现一个智能体函数。
    B、An agent function implements an agent program. 一个智能体函数实现一个智能体程序。
    C、An agent program has no any relation with an agent function. 一个智能体程序与一个智能体函数没有任何关系。
    D、An agent program is same as an agent function. 一个智能体程序与一个智能体函数相同。
    E、None of the above is correct 以上都不正确

3、There are two main kinds of formulation for 8-queens problem. Which of the following one is the formulation that starts with all 8 queens on the board and moves them around? 有两种8皇后问题的形式化方式。“初始时8个皇后都放在棋盘上,然后再进行移动”属于哪一种形式化方式?
    A、Incremental formulation 增量形式化
    B、Normal formulation 正则形式化
    C、Complete-state formulation 全态形式化
    D、Half-state formulation 半态形式化
    E、None of the above is correct 以上都不正确

4、What kind of knowledge will be used to describe how a problem is solved? 哪种知识可用于描述如何求解问题?
    A、Declarative knowledge 陈述性知识
    B、Procedural knowledge 过程性知识
    C、Deep knowledge 深层知识
    D、Shallow knowledge 浅层知识
    E、Dynamic knowledge 动态知识
    F、Static knowledge 静态知识
    G、Heuristic knowledge 启发性知识
    H、None of the above is correct 以上都不正确

5、Which of the following is used to discover general facts from training examples? 下列中哪个用于训练样本中发现一般的事实?
    A、Deductive learning 演绎学习
    B、Inductive learning 归纳学习
    C、Reinforcement learning 强化学习
    D、Supervised learning 有监督学习
    E、Unsupervised learning 无监督学习
    F、None of the above is correct 以上都不正确

6、Which statement best describes the task of “classification” in machine learning? 哪一个是机器学习中“分类”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

7、Which statement best describes the task of “regression” in machine learning? 哪一个是机器学习中“回归”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

8、Which statement best describes the task of “dimensionality reduction” in machine learning? 哪一个是机器学习中“降维”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space.将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

9、Which statement best describes the paradigm of “supervised learning” in machine learning? 哪一个是机器学习中“有监督学习”范式的正确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning techniques that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。
    G、None of the above is correct 以上都不正确

10、Which statement best describes the paradigm of “unsupervised learning” in machine learning? 哪一个是机器学习中“无监督学习”范式的正确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning techniques that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。
    G、None of the above is correct 以上都不正确

11、Which learning paradigm does it apply storing knowledge to a different but related problem? 哪种学习范式把储备的知识用于不同但相关的问题?
    A、Active learning 主动式学习
    B、Collaborative learning 协同式学习
    C、Ensemble learning 集成学习
    D、Transfer learning 迁移学习
    E、Reinforcement learning 强化学习
    F、None of the above is correct 以上都不正确

12、Select the following true statements regarding the concept of artificial intelligence. 选择以下关于人工智能概念的正确表述。
    A、Artificial intelligence is the intelligence exhibited by machines or software. 人工智能是通过机器或软件展现的智能。
    B、Artificial intelligence aims to create intelligent machines. 人工智能旨在创造智能机器。
    C、Artificial intelligence defines itself as the study of human agents. 人工智能将其定义为人类智能体的研究。
    D、Artificial intelligence is to enable the development of computers that are able to do things normally done by human. 人工智能是为了开发一类计算机使之能够完成通常由人类所能做的事。
    E、Artificial intelligence is to study and construct agent programs that perform well in a given environment. 人工智能是研究和构建在给定环境下表现良好的智能体程序。

13、Which of the following statements are true to describe strong AI (General AI)? 下列陈述中哪些是描述强AI(通用AI)的正确答案?
    A、The appropriately programmed computer with the right inputs and outputs would thereby have a mind in exactly the same sense human beings have minds. 是经过适当编程的具有正确输入和输出的计算机,因此有与人类同样判断力的头脑。
    B、It defines non-sentient computer intelligence or AI that is focused on one narrow task (just one specific problem). 其定义为无意识的计算机智能,或专注于一个狭窄任务(仅针对一个具体问题)的AI。
    C、It means a machine with the ability to apply intelligence to any problem. 指的是一种机器,具有将智能应用于任何问题的能力。
    D、It is a hypothetical agent that possesses intelligence far surpassing that of the brightest and most gifted human minds. 是一个假定的智能体,拥有远远超过聪明和最有天赋的人类大脑的智能。

14、For the following comparisons, select which approaches are more general and more amenable to scientific development. 对于如下比拟,选择哪些方法更通用并且更适合于科学开发。
    A、Thinking humanly 类人思考
    B、Thinking rationally 理性思考
    C、Acting humanly 类人动作
    D、Acting rationally 理性动作

15、A rational agent is one that does the right thing, but what does it mean to do the right thing? Considering the consequences of the agent’s behavior, select the following true statements regarding the concept of "right thing". 理性智能体指的是有正确行为的智能体,但正确的行为意味着什么?考虑智能体行为的后果,选择下列有关“正确的行为”概念的真实陈述。
    A、An agent in an environment generates a sequence of actions according to the percepts. 智能体在某个环境中依据感知生成动作序列。
    B、This sequence of actions causes the environment to go through a sequence of states. 这些动作序列引起环境改变而产生状态序列。
    C、If the sequence is desirable, then the agent has performed well. 如果该序列是所期望的,则该智能体表现良好。
    D、But if the sequence is undesirable, then the agent destroys the environment. 但若该序列是不期望的,则该智能体对其环境造成破坏。

16、Select which of the following search methods are true uninformed search. 选择如下哪些搜索方法属于无信息搜索。
    A、A* Search A*搜索
    B、Best-first Search 最佳优先搜索
    C、Bidirectional search 双向搜索
    D、Breadth-first search 宽度优先搜索
    E、Depth-first search 深度优先搜索
    F、Depth-limited search 深度受限搜索
    G、Greedy Search 贪婪搜索
    H、Uniform-cost search 一致代价搜索

17、The search strategies that use problem-specific knowledge beyond the definition of the problem itself are known as: 除了问题本身的定义之外,使用问题特定知识的搜索策略被认为是:
    A、Binary search 二元搜索
    B、Brute force search 蛮力搜索
    C、Informed search 有信息搜索
    D、Uninformed search 无信息搜索
    E、Heuristic search 启发式搜

18、Hill climbing search is sometimes called greedy local search because it grabs a good neighbor state without thinking ahead about where to go next. Unfortunately, it often gets stuck for the three reasons: 爬山搜索有时也被称为贪婪局部搜索,因为它只顾抓住一个好的邻接点的状态,而不提前思考下一步该去哪儿。它在三种情况下经常被困:
    A、Hill 山丘
    B、Local maxima 局部最大值
    C、Local minima 局部最小值
    D、Mountain 高山
    E、Plateaux 高原
    F、Ridges 山岭

19、Which of the following statements are true about Simulated annealing algorithm? 以下关于模拟退火算法的表述哪些是正确的?
    A、The inner loop of the simulated annealing algorithm is quite similar to hill climbing. 模拟退火算法的内循环与爬山法非常相似。
    B、The inner loop of the simulated annealing algorithm is very different from hill climbing. 模拟退火算法的内循环与爬山法完全不同。
    C、Instead of picking the best move, simulated annealing algorithm picks a random move. 模拟退火算法不是选择最佳行动,而是选择随机行动。
    D、Instead of picking the random move, simulated annealing algorithm picks a best move. 模拟退火算法不是选择随机行动,而是选择最佳行动。

20、Select the following true statements regarding the concept of minimax rule for a zero sum game. 从如下关于零和博弈minimax规则的概念中选择正确的答案。
    A、Each player maximizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最大。
    B、Each player minimizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最小。
    C、Each player maximizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最大。
    D、Each player minimizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最小。

21、Which of the following statements are true types of types of local consistency for constraint propagation? 如下陈述中哪些是约束传播局部一致性的正确类型?
    A、Arc consistency 弧一致
    B、Graph consistency 图一致
    C、k-consistency k一致
    D、Loop consistency 环路一致
    E、Node consistency 节点一致
    F、Path consistency 路径一致
    G、Tree consistency 树一致

22、Select which of the following statements are true about ontology. 选择下列关于本体的陈述哪些是正确的。
    A、An ontology is a formal naming and definition of the types, properties, and interrelationships of the entities for a particular domain of discourse. 本体是一种对特定论域中实体的类型、特性和相互关系的形式化命名和定义。
    B、An ontology provides a common vocabulary of an area and define the meaning of the terms and the relationships between them. 本体提供一个领域的公共词汇,并且定义一些术语的含义和它们之间的关系。
    C、Ontologies can be created to organize information that will be applied to problem solving. 本体可被创建为组织可用于问题求解的信息。
    D、Ontologies cannot be created to organize information that will be applied to problem solving. 本体不能被创建为组织可用于问题求解的信息。
    E、Ontologies can be divided the types of upper ontology, domain ontology, or hyper ontology. 本体可被分为上层本体、领域本体、或混合本体。
    F、Ontologies cannot be divided the types of upper ontology, domain ontology, or hyper ontology. 本体不能被分为上层本体、领域本体、或混合本体。

23、Which of the following approaches are used to classical planning? 下列方法中哪些可用于经典规划?
    A、Boolean satisfiability 布尔可满足性
    B、constraint satisfaction 约束满足
    C、first-order logical deduction 一阶逻辑推理
    D、planning graph 规划图
    E、plan refinement 规划精进
    F、state-space search 状态空间搜索

24、Select the following true parties regarding the three parties of machine learning. 选择如下哪些派别是有关机器学习的三个派别?
    A、Connectionism 联结主义
    B、Behaviorism 行为主义
    C、Logicalism 逻辑主义
    D、Realism 现实主义
    E、Subjectivism 主观主义
    F、Symbolicism 符号主义

25、Select the following true phrases regarding the typical paradigms in machine learning. 选择下列哪些是典型的机器学习范式。
    A、Augmented learning 增强学习
    B、Informal learning 非正式学习
    C、Multimedia learning 多媒体学习
    D、Reinforcement learning 强化学习
    E、Supervised learning 有监督学习
    F、Unsupervised learning 无监督学习

26、Select which ones of the following algorithms (methods) have been invented as the special algorithms dedicated to reinforcement learning. 选择下列哪些算法(方法)已被发明为专门用于强化学习的算法。
    A、Asynchronous Advantage Actor-Critic 异步优势动作者·评论者
    B、Behaviorist Psychology 行为学派心理学
    C、Deep Network 深度网络
    D、Deterministic Policy Gradients 确定性策略梯度
    E、Neural Episodic Control 神经情景控制

27、Which of the following phrases are true algorithms used by manifold learning? 下列哪些是真正被流形学习使用的算法?
    A、HMM (Hidden Markov models) HMM(隐藏马可夫模型)
    B、Isomap (Isometric mapping) Isomap(等距映射)
    C、LLE (Locally Linear Embeddings) LLE(局部线性嵌入)
    D、LE (Laplacian Eigenmaps) LE(拉普拉斯特征映射)
    E、PCA (Principal Component Analysis) PCA(主成分分析)
    F、SVM (Support Vector Machine) SVM(支撑向量机)

28、Select the following true phrases in which each one has been invented as a logical model for machine learning. 从如下短语中选择真正地被发明为或被用作机器学习逻辑模型的方法。
    A、Adaptive boosting 自适应性提升
    B、Association rules 关联规则
    C、Convolutional neural network 卷积神经网络
    D、Decision tree 决策树
    E、First-order logic 一阶逻辑
    F、Linear regression 线性回归

29、Which of the following phrases are true artificial neural networks used in machine Learning? 下列用于哪些是真正用于机器学习的人工神经网络?
    A、Bayes Network 贝叶斯网络
    B、Convolutional neural network 卷积神经网络
    C、Deep auto-encoder 深度自动编码器
    D、Long short-term memory 长短期记忆
    E、Wireless network 无线网络
    F、Generative Adversarial Networks 生成对抗网络

30、Rationally means the property of a system which does the right thing given what it ________. 理性指的是一个系统的属性,即在_________的环境下做正确的事。(中英文回答均可)

31、An intelligent agent is anything that can be viewed as perceiving its environment through sensors, and acting upon that environment through _______. 智能体可以看作是通过感受器感知外部环境,并且通过_________作用于外部环境的任何事物。(中英文回答均可)

32、Local search algorithms operate using a single current node (rather than multiple paths) and generally move only to ______ of that node. 局部搜索算法使用一个当前节点(而不是多条路径),并且通常仅移动到该节点的______。(中英文回答均可)

33、____________ is a dynamic game with probabilistic transitions played by one or more players. ____________是一种具有概率转换的动态博弈,有一个或多个玩家。(中英文回答均可)

34、___________ tree search is on the analysis of the most promising moves, expanding the search tree based on random sampling of the search space. ___________树搜索对最有利的动作进行分析,根据搜索空间的随机采样来扩展搜索树。(中英文回答均可)

35、A constraint satisfaction problem (CSP) is defined as a set of objects whose state must satisfy a number of _______________. 约束满足问题 (CSP) 被定义为其状态必须满足若干_______________的一组对象。(中英文回答均可)

36、Markov Decision Process (MDP) is a discrete time _______________ process, means action outcomes depend only on the current state. 马可夫决策过程(MDP)是一种离散时间_______________过程,意味着动作结果仅仅依赖于当前状态。(中英文回答均可)

37、Machine learning is a _______ of artificial intelligence, concerns the construction and study of systems that can learn from data. 机器学习是人工智能的一个_______,从事构建和研究可以从数据中学习的系统。(中英文回答均可)

38、Clustering is to identify similar groups for input objects _____ training data, and classification is to assign pre-defined classes for input items with training data. 聚类是为输入对象标识相似的组,_____训练数据;而分类则是给输入项分派预定义的类,有训练数据。(中英文回答均可)

39、The probabilistic model in machine learning is to use probability theory to express all forms of _____________. 机器学习的概率模型是采用概率论来表示所有____________的形式。(中英文回答均可)

40、An artificial neural network is an artificial representation of the human brain that tries to simulate its _________ processing. 人工神经网络是一种人脑的人工表示,试图仿真其_________过程。(中英文回答均可)

Final Exam of PoAI

Final Exam of PoAI

1、Turing Test is designed to provide what kind of satisfactory operational definition? 图灵测试旨在给予哪一种令人满意的操作定义?
    A、machine action 机器动作
    B、human action 人类动作
    C、machine thinking 机器思考
    D、human thinking 人类思考
    E、machine intelligence 机器智能
    F、human intelligence 人类智能
    G、None of the above is correct 以上都不正确

2、Thinking the differences between agent functions and agent programs, select correct statements from following ones. 考虑智能体函数与智能体程序的差异,从下列陈述中选择正确的答案。
    A、An agent program implements an agent function. 一个智能体程序实现一个智能体函数。
    B、An agent function implements an agent program. 一个智能体函数实现一个智能体程序。
    C、An agent program has no any relation with an agent function. 一个智能体程序与一个智能体函数没有任何关系。
    D、An agent program is same as an agent function. 一个智能体程序与一个智能体函数相同。
    E、None of the above is correct 以上都不正确

3、There are two main kinds of formulation for 8-queens problem. Which of the following one is the formulation that starts with all 8 queens on the board and moves them around? 有两种8皇后问题的形式化方式。“初始时8个皇后都放在棋盘上,然后再进行移动”属于哪一种形式化方式?
    A、Incremental formulation 增量形式化
    B、Normal formulation 正则形式化
    C、Complete-state formulation 全态形式化
    D、Half-state formulation 半态形式化
    E、None of the above is correct 以上都不正确

4、What kind of knowledge will be used to describe how a problem is solved? 哪种知识可用于描述如何求解问题?
    A、Declarative knowledge 陈述性知识
    B、Procedural knowledge 过程性知识
    C、Deep knowledge 深层知识
    D、Shallow knowledge 浅层知识
    E、Dynamic knowledge 动态知识
    F、Static knowledge 静态知识
    G、Heuristic knowledge 启发性知识
    H、None of the above is correct 以上都不正确

5、Which of the following is used to discover general facts from training examples? 下列中哪个用于训练样本中发现一般的事实?
    A、Deductive learning 演绎学习
    B、Inductive learning 归纳学习
    C、Reinforcement learning 强化学习
    D、Supervised learning 有监督学习
    E、Unsupervised learning 无监督学习
    F、None of the above is correct 以上都不正确

6、Which statement best describes the task of “classification” in machine learning? 哪一个是机器学习中“分类”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

7、Which statement best describes the task of “regression” in machine learning? 哪一个是机器学习中“回归”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space. 将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

8、Which statement best describes the task of “dimensionality reduction” in machine learning? 哪一个是机器学习中“降维”任务的正确描述?
    A、To assign a category to each item. 为每个项目分配一个类别。
    B、To find the distribution of inputs in some space. 发现某个空间中输入的分布。
    C、To group data objects. 对数据对象进行分组。
    D、To order items according to some criterion. 依据某个准则对项目进行排序。
    E、To predict a real value for each item. 预测每个项目实际的值。
    F、To simplify inputs by mapping them into a lower space.将其映射到低维空间来简化输入。
    G、None of the above is correct 以上都不正确

9、Which statement best describes the paradigm of “supervised learning” in machine learning? 哪一个是机器学习中“有监督学习”范式的正确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning techniques that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。
    G、None of the above is correct 以上都不正确

10、Which statement best describes the paradigm of “unsupervised learning” in machine learning? 哪一个是机器学习中“无监督学习”范式的正确描述?
    A、It aims to learn some information from one, or only a few, training examples. 旨在从一个、或仅有的几个训练样本中学习一些信息。
    B、It aims to learn some information using a limited amount of training examples. 旨在采用有限数量的训练样本来学习一些信息。
    C、It is a class of supervised learning techniques that also make use of unlabeled data for training. 属于有监督学习算法一类,此外还利用未标记数据进行训练。
    D、It is able to solve a task despite not having received any training examples of that task. 能够求解一个任务,即使没有得到该任务的任何训练样本。
    E、It receives a set of labeled data and makes predictions for all unseen data. 接收一组标注数据并对所有的未知数据做出预测。
    F、It receives exclusively unlabeled data and makes predictions for all unseen data. 仅仅接收未标注数据并对所有的未知数据做出预测。
    G、None of the above is correct 以上都不正确

11、Which learning paradigm does it apply storing knowledge to a different but related problem? 哪种学习范式把储备的知识用于不同但相关的问题?
    A、Active learning 主动式学习
    B、Collaborative learning 协同式学习
    C、Ensemble learning 集成学习
    D、Transfer learning 迁移学习
    E、Reinforcement learning 强化学习
    F、None of the above is correct 以上都不正确

12、Select the following true statements regarding the concept of artificial intelligence. 选择以下关于人工智能概念的正确表述。
    A、Artificial intelligence is the intelligence exhibited by machines or software. 人工智能是通过机器或软件展现的智能。
    B、Artificial intelligence aims to create intelligent machines. 人工智能旨在创造智能机器。
    C、Artificial intelligence defines itself as the study of human agents. 人工智能将其定义为人类智能体的研究。
    D、Artificial intelligence is to enable the development of computers that are able to do things normally done by human. 人工智能是为了开发一类计算机使之能够完成通常由人类所能做的事。
    E、Artificial intelligence is to study and construct agent programs that perform well in a given environment. 人工智能是研究和构建在给定环境下表现良好的智能体程序。

13、Which of the following statements are true to describe strong AI (General AI)? 下列陈述中哪些是描述强AI(通用AI)的正确答案?
    A、The appropriately programmed computer with the right inputs and outputs would thereby have a mind in exactly the same sense human beings have minds. 是经过适当编程的具有正确输入和输出的计算机,因此有与人类同样判断力的头脑。
    B、It defines non-sentient computer intelligence or AI that is focused on one narrow task (just one specific problem). 其定义为无意识的计算机智能,或专注于一个狭窄任务(仅针对一个具体问题)的AI。
    C、It means a machine with the ability to apply intelligence to any problem. 指的是一种机器,具有将智能应用于任何问题的能力。
    D、It is a hypothetical agent that possesses intelligence far surpassing that of the brightest and most gifted human minds. 是一个假定的智能体,拥有远远超过聪明和最有天赋的人类大脑的智能。

14、For the following comparisons, select which approaches are more general and more amenable to scientific development. 对于如下比拟,选择哪些方法更通用并且更适合于科学开发。
    A、Thinking humanly 类人思考
    B、Thinking rationally 理性思考
    C、Acting humanly 类人动作
    D、Acting rationally 理性动作

15、A rational agent is one that does the right thing, but what does it mean to do the right thing? Considering the consequences of the agent’s behavior, select the following true statements regarding the concept of "right thing". 理性智能体指的是有正确行为的智能体,但正确的行为意味着什么?考虑智能体行为的后果,选择下列有关“正确的行为”概念的真实陈述。
    A、An agent in an environment generates a sequence of actions according to the percepts. 智能体在某个环境中依据感知生成动作序列。
    B、This sequence of actions causes the environment to go through a sequence of states. 这些动作序列引起环境改变而产生状态序列。
    C、If the sequence is desirable, then the agent has performed well. 如果该序列是所期望的,则该智能体表现良好。
    D、But if the sequence is undesirable, then the agent destroys the environment. 但若该序列是不期望的,则该智能体对其环境造成破坏。

16、Select which of the following search methods are true uninformed search. 选择如下哪些搜索方法属于无信息搜索。
    A、A* Search A*搜索
    B、Best-first Search 最佳优先搜索
    C、Bidirectional search 双向搜索
    D、Breadth-first search 宽度优先搜索
    E、Depth-first search 深度优先搜索
    F、Depth-limited search 深度受限搜索
    G、Greedy Search 贪婪搜索
    H、Uniform-cost search 一致代价搜索

17、The search strategies that use problem-specific knowledge beyond the definition of the problem itself are known as: 除了问题本身的定义之外,使用问题特定知识的搜索策略被认为是:
    A、Binary search 二元搜索
    B、Brute force search 蛮力搜索
    C、Informed search 有信息搜索
    D、Uninformed search 无信息搜索
    E、Heuristic search 启发式搜

18、Hill climbing search is sometimes called greedy local search because it grabs a good neighbor state without thinking ahead about where to go next. Unfortunately, it often gets stuck for the three reasons: 爬山搜索有时也被称为贪婪局部搜索,因为它只顾抓住一个好的邻接点的状态,而不提前思考下一步该去哪儿。它在三种情况下经常被困:
    A、Hill 山丘
    B、Local maxima 局部最大值
    C、Local minima 局部最小值
    D、Mountain 高山
    E、Plateaux 高原
    F、Ridges 山岭

19、Which of the following statements are true about Simulated annealing algorithm? 以下关于模拟退火算法的表述哪些是正确的?
    A、The inner loop of the simulated annealing algorithm is quite similar to hill climbing. 模拟退火算法的内循环与爬山法非常相似。
    B、The inner loop of the simulated annealing algorithm is very different from hill climbing. 模拟退火算法的内循环与爬山法完全不同。
    C、Instead of picking the best move, simulated annealing algorithm picks a random move. 模拟退火算法不是选择最佳行动,而是选择随机行动。
    D、Instead of picking the random move, simulated annealing algorithm picks a best move. 模拟退火算法不是选择随机行动,而是选择最佳行动。

20、Select the following true statements regarding the concept of minimax rule for a zero sum game. 从如下关于零和博弈minimax规则的概念中选择正确的答案。
    A、Each player maximizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最大。
    B、Each player minimizes the maximum payoff possible for itself. 每个玩家会使自己可能的最大收益变得最小。
    C、Each player maximizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最大。
    D、Each player minimizes the maximum loss possible for the other. 每个玩家会使对手可能的最大损失变得最小。

21、Which of the following statements are true types of types of local consistency for constraint propagation? 如下陈述中哪些是约束传播局部一致性的正确类型?
    A、Arc consistency 弧一致
    B、Graph consistency 图一致
    C、k-consistency k一致
    D、Loop consistency 环路一致
    E、Node consistency 节点一致
    F、Path consistency 路径一致
    G、Tree consistency 树一致

22、Select which of the following statements are true about ontology. 选择下列关于本体的陈述哪些是正确的。
    A、An ontology is a formal naming and definition of the types, properties, and interrelationships of the entities for a particular domain of discourse. 本体是一种对特定论域中实体的类型、特性和相互关系的形式化命名和定义。
    B、An ontology provides a common vocabulary of an area and define the meaning of the terms and the relationships between them. 本体提供一个领域的公共词汇,并且定义一些术语的含义和它们之间的关系。
    C、Ontologies can be created to organize information that will be applied to problem solving. 本体可被创建为组织可用于问题求解的信息。
    D、Ontologies cannot be created to organize information that will be applied to problem solving. 本体不能被创建为组织可用于问题求解的信息。
    E、Ontologies can be divided the types of upper ontology, domain ontology, or hyper ontology. 本体可被分为上层本体、领域本体、或混合本体。
    F、Ontologies cannot be divided the types of upper ontology, domain ontology, or hyper ontology. 本体不能被分为上层本体、领域本体、或混合本体。

23、Which of the following approaches are used to classical planning? 下列方法中哪些可用于经典规划?
    A、Boolean satisfiability 布尔可满足性
    B、constraint satisfaction 约束满足
    C、first-order logical deduction 一阶逻辑推理
    D、planning graph 规划图
    E、plan refinement 规划精进
    F、state-space search 状态空间搜索

24、Select the following true parties regarding the three parties of machine learning. 选择如下哪些派别是有关机器学习的三个派别?
    A、Connectionism 联结主义
    B、Behaviorism 行为主义
    C、Logicalism 逻辑主义
    D、Realism 现实主义
    E、Subjectivism 主观主义
    F、Symbolicism 符号主义

25、Select the following true phrases regarding the typical paradigms in machine learning. 选择下列哪些是典型的机器学习范式。
    A、Augmented learning 增强学习
    B、Informal learning 非正式学习
    C、Multimedia learning 多媒体学习
    D、Reinforcement learning 强化学习
    E、Supervised learning 有监督学习
    F、Unsupervised learning 无监督学习

26、Select which ones of the following algorithms (methods) have been invented as the special algorithms dedicated to reinforcement learning. 选择下列哪些算法(方法)已被发明为专门用于强化学习的算法。
    A、Asynchronous Advantage Actor-Critic 异步优势动作者·评论者
    B、Behaviorist Psychology 行为学派心理学
    C、Deep Network 深度网络
    D、Deterministic Policy Gradients 确定性策略梯度
    E、Neural Episodic Control 神经情景控制

27、Which of the following phrases are true algorithms used by manifold learning? 下列哪些是真正被流形学习使用的算法?
    A、HMM (Hidden Markov models) HMM(隐藏马可夫模型)
    B、Isomap (Isometric mapping) Isomap(等距映射)
    C、LLE (Locally Linear Embeddings) LLE(局部线性嵌入)
    D、LE (Laplacian Eigenmaps) LE(拉普拉斯特征映射)
    E、PCA (Principal Component Analysis) PCA(主成分分析)
    F、SVM (Support Vector Machine) SVM(支撑向量机)

28、Select the following true phrases in which each one has been invented as a logical model for machine learning. 从如下短语中选择真正地被发明为或被用作机器学习逻辑模型的方法。
    A、Adaptive boosting 自适应性提升
    B、Association rules 关联规则
    C、Convolutional neural network 卷积神经网络
    D、Decision tree 决策树
    E、First-order logic 一阶逻辑
    F、Linear regression 线性回归

29、Which of the following phrases are true artificial neural networks used in machine Learning? 下列用于哪些是真正用于机器学习的人工神经网络?
    A、Bayes Network 贝叶斯网络
    B、Convolutional neural network 卷积神经网络
    C、Deep auto-encoder 深度自动编码器
    D、Long short-term memory 长短期记忆
    E、Wireless network 无线网络
    F、Generative Adversarial Networks 生成对抗网络

30、Rationally means the property of a system which does the right thing given what it ________. 理性指的是一个系统的属性,即在_________的环境下做正确的事。(中英文回答均可)

31、An intelligent agent is anything that can be viewed as perceiving its environment through sensors, and acting upon that environment through _______. 智能体可以看作是通过感受器感知外部环境,并且通过_________作用于外部环境的任何事物。(中英文回答均可)

32、Local search algorithms operate using a single current node (rather than multiple paths) and generally move only to ______ of that node. 局部搜索算法使用一个当前节点(而不是多条路径),并且通常仅移动到该节点的______。(中英文回答均可)

33、____________ is a dynamic game with probabilistic transitions played by one or more players. ____________是一种具有概率转换的动态博弈,有一个或多个玩家。(中英文回答均可)

34、___________ tree search is on the analysis of the most promising moves, expanding the search tree based on random sampling of the search space. ___________树搜索对最有利的动作进行分析,根据搜索空间的随机采样来扩展搜索树。(中英文回答均可)

35、A constraint satisfaction problem (CSP) is defined as a set of objects whose state must satisfy a number of _______________. 约束满足问题 (CSP) 被定义为其状态必须满足若干_______________的一组对象。(中英文回答均可)

36、Markov Decision Process (MDP) is a discrete time _______________ process, means action outcomes depend only on the current state. 马可夫决策过程(MDP)是一种离散时间_______________过程,意味着动作结果仅仅依赖于当前状态。(中英文回答均可)

37、Machine learning is a _______ of artificial intelligence, concerns the construction and study of systems that can learn from data. 机器学习是人工智能的一个_______,从事构建和研究可以从数据中学习的系统。(中英文回答均可)

38、Clustering is to identify similar groups for input objects _____ training data, and classification is to assign pre-defined classes for input items with training data. 聚类是为输入对象标识相似的组,_____训练数据;而分类则是给输入项分派预定义的类,有训练数据。(中英文回答均可)

39、The probabilistic model in machine learning is to use probability theory to express all forms of _____________. 机器学习的概率模型是采用概率论来表示所有____________的形式。(中英文回答均可)

40、An artificial neural network is an artificial representation of the human brain that tries to simulate its _________ processing. 人工神经网络是一种人脑的人工表示,试图仿真其_________过程。(中英文回答均可)