{"第一章机考题库":[{"type":"选择题","question":"人工智能独立性特征包括？","options":{"A":"自学习与自我决策","B":"依赖人类指令","C":"无信息处理能力","D":"无计算能力"},"answer":"A"},{"type":"选择题","question":"图灵测试的核心是：","options":{"A":"机器能否表现出人类水平的智能行为","B":"人类能否教会计算机下棋","C":"人工智能能否理解图像","D":"计算机能否做复杂计算"},"answer":"A"},{"type":"选择题","question":"AlphaGo综合使用的学派不包括：","options":{"A":"行为主义学派","B":"符号主义学派","C":"连接主义学派","D":"数学主义学派"},"answer":"D"},{"type":"选择题","question":"符号主义学派主要研究：","options":{"A":"数据挖掘","B":"神经网络","C":"逻辑与规则","D":"行为与环境"},"answer":"C"},{"type":"选择题","question":"连接主义学派的代表技术是：","options":{"A":"神经网络与深度学习","B":"强化学习","C":"专家系统","D":"蒙特卡罗树搜索"},"answer":"A"},{"type":"选择题","question":"以下哪项技术是人工智能基础科学问题的研究内容？","options":{"A":"脑科学及类脑智能","B":"股票分析","C":"文学写作","D":"音乐推荐"},"answer":"A"},{"type":"选择题","question":"下列哪项属于弱人工智能？","options":{"A":"AlphaGo围棋程序","B":"通用人工智能","C":"能自主学习各种技能的机器人","D":"具有自我意识的计算机"},"answer":"A"},{"type":"选择题","question":"机器学习核心思想是：","options":{"A":"设计专家系统","B":"模拟动物行为","C":"基于规则推理","D":"通过数据训练模型并预测"},"answer":"D"},{"type":"选择题","question":"人工智能的定义中强调的能力不包括以下哪项？","options":{"A":"感知","B":"决策","C":"创作文学作品","D":"推理"},"answer":"C"},{"type":"选择题","question":"深度学习的核心思想是：","options":{"A":"强化专家系统","B":"环境模拟","C":"手工编写规则","D":"数据驱动的多层特征学习"},"answer":"D"},{"type":"选择题","question":"第一次人工智能发展浪潮的时间是：","options":{"A":"1997—至今","B":"1956—1970年","C":"2006—2012年","D":"1980—1990年"},"answer":"B"},{"type":"选择题","question":"人工智能发展史上，哪项技术标志着自动驾驶进入实际应用阶段？","options":{"A":"Deep Blue","B":"AlphaGo","C":"Google无人驾驶汽车","D":"GPT模型"},"answer":"C"},{"type":"选择题","question":"达特茅斯会议的年份是：","options":{"A":"1949年","B":"1956年","C":"1950年","D":"1960年"},"answer":"B"},{"type":"选择题","question":"第二次浪潮失败的原因之一是：","options":{"A":"人类对AI缺乏兴趣","B":"第五代计算机项目的失败","C":"数据集不足","D":"网络技术落后"},"answer":"B"},{"type":"选择题","question":"下列属于人工智能应用示范研究的领域是：","options":{"A":"音乐创作","B":"天文学","C":"古典文学","D":"智慧医疗"},"answer":"D"},{"type":"选择题","question":"弱人工智能与强人工智能的核心区别是：","options":{"A":"是否能感知人类情感","B":"智能层级与应用领域","C":"是否能执行任何任务","D":"是否使用计算机"},"answer":"B"},{"type":"选择题","question":"第一次人工智能寒冬的原因主要是：","options":{"A":"技术过于复杂","B":"缺乏科学家参与","C":"存储能力不足及公众期望过高","D":"法律限制"},"answer":"C"},{"type":"选择题","question":"第二次人工智能发展浪潮的标志性技术是：","options":{"A":"GPT模型","B":"AlphaGo","C":"专家系统","D":"逻辑推理"},"answer":"C"},{"type":"选择题","question":"当前人工智能技术发展重点不包括：","options":{"A":"自主智能系统","B":"跨媒体协同处理","C":"群体集成智能","D":"古文翻译"},"answer":"D"},{"type":"选择题","question":"人工智能一词由谁在1956年首次提出？","options":{"A":"杰佛里·辛顿","B":"马尔文·明斯基","C":"艾伦·图灵","D":"约翰·麦卡锡"},"answer":"D"},{"type":"选择题","question":"下列哪项是强人工智能的特征？","options":{"A":"仅能进行图像识别","B":"无法跨领域应用","C":"只能解决单一领域问题","D":"具备自我意识"},"answer":"D"},{"type":"判断题","question":"AlphaGo属于强人工智能。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"达特茅斯会议被称为人工智能元年。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"人工智能可以完全自主产生原创性意识。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"强人工智能可以跨领域自主学习。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"符号主义学派面临处理不确定性问题的挑战。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"人工智能完全不受数据限制。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"人工智能的发展仅限于计算机科学领域。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"AlphaGo综合使用了符号主义、连接主义和行为主义学派的方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"人工智能可以被定义为让机器具备类似人类的智能行为。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"连接主义学派主要依靠神经网络和数据训练实现智能。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"专家系统是第二次人工智能发展浪潮的代表。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度学习技术首次提出是在2006年。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"弱人工智能具备自主决策能力。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"机器学习的核心思想是数据驱动的模型训练与预测。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"图灵测试是判定机器是否具备人工智能的标准之一。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"填空题","question":"符号主义学派认为人工智能应该基于逻辑、________进行处理，通过编写规则和逻辑来模拟人类思维过程。","answer":"符号"},{"type":"填空题","question":"人工智能（AI）是一门研究、开发用于模拟、延伸和扩展________ 的理论、方法、技术及应用系统的科学。","answer":"智能行为"},{"type":"填空题","question":"人工智能最核心的目标是使机器能够模拟人类的________ 活动，如学习、推理、规划、感知和语言理解。","answer":"认知"},{"type":"填空题","question":"连接主义学派人工智能可以模拟人脑的结构和功能通过________实现智能。","answer":"神经网络"},{"type":"填空题","question":"认为智能的本质是符号操作，可以通过“如果-那么”逻辑规则实现的人工智能学派是________。","answer":"符号主义"},{"type":"填空题","question":"当前人工智能的第三次高潮主要得益于大数据的积累、________ 的提升以及深度学习算法的突破。","answer":"计算能力"},{"type":"填空题","question":"1997年，IBM的超级计算机________ 击败了国际象棋世界冠军卡斯帕罗夫，标志着人工智能在博弈领域的重大突破。","answer":"深蓝"},{"type":"填空题","question":"1997年，IBM的超级计算机________ 击败了国际象棋世界冠军卡斯帕罗夫，标志着人工智能在博弈领域的重大突破。","answer":"深蓝"},{"type":"填空题","question":"人工智能这一术语是在_会议被正式提出。","answer":"达特茅斯"},{"type":"填空题","question":"主张智能行为不需要复杂的符号推理，只需要在环境中通过“感知-行动”反馈来实现的学派是________，其代表技术是强化学习和机器人控制。","answer":"行为主义"},{"type":"填空题","question":"人工智能发展的第一次高潮主要集中于逻辑证明和________ 领域，这一时期以符号主义为主导。","answer":"问题求解"},{"type":"填空题","question":"人工智能的三大研究学派包括符号主义、________ 和行为主义。","answer":"连接主义"},{"type":"填空题","question":"弱人工智能是指能够解决________领域问题的人工智能。","answer":"单个特定"},{"type":"填空题","question":"人工智能发展的第二次高潮（20世纪80年代）主要得益于________ 的成功应用，如MYCIN医疗诊断系统。","answer":"专家系统"},{"type":"填空题","question":"行为主义学派认为人工智能应该关注智能体的________和交互，通过感知和动作实现智能。","answer":"行为"}],"第二章机考题库":[{"type":"选择题","question":"以下哪个是知识图谱的用途？","options":{"A":"用来做菜","B":"用来浇水","C":"用来玩游戏","D":"帮助计算机更好地理解文本的语义信息"},"answer":"D"},{"type":"选择题","question":"什么是本体的定义？","options":{"A":"共享概念化的形式化显式规范说明","B":"一种新的计算机","C":"一种游戏","D":"一种交通工具"},"answer":"A"},{"type":"选择题","question":"以下哪个是知识的定义？","options":{"A":"一种游戏","B":"人类对客观世界的认识和经验积累","C":"一种动物","D":"一种食物"},"answer":"B"},{"type":"选择题","question":"知识表示是做什么的？","options":{"A":"让人类理解计算机","B":"让计算机理解人类的知识","C":"让电脑变得更小","D":"让网络连接更快"},"answer":"B"},{"type":"选择题","question":"以下哪项是过程性知识的例子？","options":{"A":"雪是白色的","B":"小明很聪明","C":"如果下雨，那么带伞","D":"炒菜包括洗菜、切菜、下锅、翻炒和装盘等过程"},"answer":"D"},{"type":"选择题","question":"什么是知识的不确定性？","options":{"A":"知识总是精确","B":"知识无法表示","C":"知识可能模糊，不精确","D":"知识总是模糊"},"answer":"C"},{"type":"选择题","question":"语义网络是用什么表示对象和概念的？","options":{"A":"数字","B":"三角形","C":"节点和节点之间的边","D":"字母"},"answer":"C"},{"type":"选择题","question":"知识表示是将什么转化为计算机可识别的形式？","options":{"A":"以上都不是","B":"人类知识","C":"机器知识","D":"电子知识"},"answer":"B"},{"type":"选择题","question":"以下哪个是知识的特点？","options":{"A":"相对正确性","B":"不可利用性","C":"导电","D":"不可表示性"},"answer":"A"},{"type":"选择题","question":"什么是知识表示？","options":{"A":"让计算机更便宜","B":"将人类知识转化为计算机可识别的形式","C":"让计算机变得更小","D":"音乐"},"answer":"B"},{"type":"选择题","question":"知识表示方法中，哪种方法是用\"如果...那么...\"来表示知识的？","options":{"A":"语义网络","B":"框架","C":"产生式规则","D":"知识图谱"},"answer":"C"},{"type":"选择题","question":"什么是状态空间?","options":{"A":"交通工具","B":"所有可能状态的集合","C":"游戏","D":"计算机"},"answer":"B"},{"type":"选择题","question":"知识图谱能帮助计算机做什么？","options":{"A":"更好地玩游戏","B":"浇水","C":"更好地做菜","D":"更好地理解文本的语义信息"},"answer":"D"},{"type":"选择题","question":"什么是知识的可表示性？","options":{"A":"知识不能被表示","B":"知识只能用图片表示","C":"知识可以用适当的形式表示","D":"知识只能用文字表示"},"answer":"C"},{"type":"选择题","question":"什么是语义网络的节点？","options":{"A":"表示颜色","B":"表示游戏规则","C":"表示音乐","D":"表示对象和概念"},"answer":"D"},{"type":"选择题","question":"以下哪个是知识表示的原则？","options":{"A":"便于表达","B":"便于理解","C":"便于实现","D":"以上都是"},"answer":"D"},{"type":"选择题","question":"什么是框架的槽？","options":{"A":"表示音乐","B":"表示游戏","C":"表示颜色","D":"表示对象的一个属性"},"answer":"D"},{"type":"选择题","question":"什么是语义网络？","options":{"A":"一种交通工具","B":"一种游戏","C":"用节点和弧表示对象和关系的网络","D":"一种新的计算机语言"},"answer":"C"},{"type":"选择题","question":"以下哪个不是知识表示方法？","options":{"A":"烹饪","B":"框架","C":"脚本","D":"语义网络"},"answer":"A"},{"type":"选择题","question":"什么是框架表示法？","options":{"A":"音乐","B":"用槽和槽值表示对象属性的方法","C":"游戏","D":"颜色"},"answer":"B"},{"type":"选择题","question":"知识是什么？","options":{"A":"一种食物","B":"一种游戏","C":"一种动物","D":"人类对客观世界的认识和经验积累"},"answer":"D"},{"type":"选择题","question":"以下哪个不是知识的特点？","options":{"A":"相对正确性","B":"可表示性","C":"绝对正确性","D":"不确定性"},"answer":"C"},{"type":"选择题","question":"产生式规则的表示方式是什么？","options":{"A":"开始...结束对","B":"描述...说明对","C":"问题...答案对","D":"模式—动作对"},"answer":"D"},{"type":"选择题","question":"知识表示的目的是什么？","options":{"A":"让计算机变得更便宜","B":"让计算机变得更小","C":"让计算机更漂亮","D":"让计算机能够理解和利用知识"},"answer":"D"},{"type":"选择题","question":"什么是知识的相对正确性？","options":{"A":"知识总是错误","B":"知识总是正确","C":"知识在特定条件下才正确","D":"知识无法判断"},"answer":"C"},{"type":"选择题","question":"什么是语义网络的边？","options":{"A":"表示颜色","B":"表示节点之间的关系","C":"表示游戏规则","D":"表示音乐"},"answer":"B"},{"type":"选择题","question":"什么是知识图谱？","options":{"A":"一种交通工具","B":"一种新的互联网","C":"一种将知识结构化表示的方法","D":"一种游戏"},"answer":"C"},{"type":"选择题","question":"什么是知识的可利用性？","options":{"A":"知识只能被存储","B":"知识只能被遗忘","C":"知识不能被使用","D":"知识可以被用来解决问题"},"answer":"D"},{"type":"选择题","question":"知识可以分为哪几类？","options":{"A":"描述性知识","B":"判断性知识","C":"以上都是","D":"过程性知识"},"answer":"C"},{"type":"判断题","question":"知识是人类对客观世界的认识和经验积累。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"框架通常用于描述具有固定形式的对象。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"逻辑连接词包括合取连接词、析取连接词、蕴含连接词和否定连接词。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识无法被用来解决实际问题。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"一阶谓词逻辑表示法是以数理逻辑为基础构成的知识表示方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"谓词逻辑是一种知识表示方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识图谱不能用于搜索引擎。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"知识表示是将人类知识转化为计算机可识别的形式。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识可以分为描述性知识、判断性知识和过程性知识3类。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"状态空间表示法把求解的问题表示为问题状态、操作、约束、初始状态和目标状态。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"本体的概念最初用于哲学。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识只能用文字形式表示，不能用图形或神经网络表示。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"概念从属是一种表示自然语言语义的理论。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"陈述性知识表示是以数据的形式表示，将知识表示与知识运用分开处理。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"“1+1=2”在十进制条件下是正确的，在二进制条件下也是正确的。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"语义网络是一种有向图，节点表示对象和概念，边表示节点之间的关系。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"过程性知识表示是指将知识表示成一个过程，它将知识表示和知识运用相结合。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"脚本用于描述固定的事件序列。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识表示属于人工智能的研究领域。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"知识特点包括相对正确性、不确定性、可表示性和可利用性4个方面。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"框架表示法不适用于结构性知识的表示。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"知识表示不需要考虑是否便于扩充。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"产生式规则最早由爱因斯坦在1643年提出。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"正向推理是从已知事实出发，通过规则库得出结论。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"全称量词表示“所有的”，存在量词表示“存在”。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"反向推理是指从目标出发，并反向使用规则，以求得已知事实。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"填空题","question":"将不同数据源中指向同一真实世界的实体进行合并的过程称为________。","answer":"实体对齐"},{"type":"填空题","question":"从已知事实出发，不断匹配规则直到推出结论的推理方式是________推理。","answer":"正向"},{"type":"填空题","question":"从结构化、半结构化和非结构化数据中提取实体、关系和属性的过程称为________。","answer":"知识抽取"},{"type":"填空题","question":"框架表示法中，描述对象属性的栏目被称为________。","answer":"槽"},{"type":"填空题","question":"脚本表示法适用于描述____如就餐流程，清晰展现事件的前后关联。","answer":"固定事件序列"},{"type":"填空题","question":"知识抽取包括实体抽取、关系抽取、属性抽取和________抽取四种类型。","answer":"事件"},{"type":"填空题","question":"知识表示方法的“表示能力”原则指的是该方法能否准确表达知识的________和精确性。","answer":"完整性"},{"type":"填空题","question":"知识具有相对正确性，即知识只在特定的________或环境成立，例如“所有的鸟都会飞”在遇到企鹅时就不成立。","answer":"条件"},{"type":"填空题","question":"知识表示方法的“访问与扩充”原则关注知识的检索效率和________性。","answer":"可维护"},{"type":"填空题","question":"用于描述问题求解控制策略，如炒菜等过程的知识，属于____知识。","answer":"过程性"},{"type":"填空题","question":"在人工智能中，将人类知识符号化、格式化为计算机可以存储和操作的数据结构的过程，称为________。","answer":"知识表示"},{"type":"填空题","question":"从文本中识别出具体对象（如人名、地名、机构名）的过程称为________抽取。","answer":"实体"},{"type":"填空题","question":"知识图谱的基本组成单位是三元组，其形式为（实体，关系，________）","answer":"实体"},{"type":"填空题","question":"在一阶谓词逻辑中，“∀x (Bird(x) → Fly(x))”表示“________ 鸟都会飞”。","answer":"所有"},{"type":"填空题","question":"在专家系统中，产生式规则的推理方式分为三种：正向推理、反向推理和________推理。","answer":"双向"},{"type":"填空题","question":"用于表示对象和概念的特征与其相互关系的知识，属于_____知识。","answer":"描述性"},{"type":"填空题","question":"知识加工主要包括本体构建、知识推理和__________。","answer":"质量评估"},{"type":"填空题","question":"确定文本中的某个名称在特定上下文中具体指哪个实体的过程（如“苹果”指水果还是公司）称为________。","answer":"实体消歧"},{"type":"填空题","question":"知识图谱的表示技术通常包括本体表示和______。","answer":"嵌入表示"},{"type":"填空题","question":"在知识图谱构建中，定义实体类型、关系类型和属性约束的阶段称为________。","answer":"知识建模"},{"type":"填空题","question":"知识可以以适当的形式进行表示，比如通过语言、文字、图像等进行存储和传播，表现了知识具有____。","answer":"可表示性"},{"type":"填空题","question":"选择知识表示方法时，需要综合考虑三个原则：表示能力、________与实现、访问与扩充。","answer":"理解"},{"type":"填空题","question":"知识图谱的构建流程包括知识建模、知识抽取、知识融合、________和知识存储五个阶段。","answer":"知识加工"},{"type":"填空题","question":"在一阶谓词逻辑中，“∃x (Student(x) ∧ InClass(x, AI101))”其中∃是________量词。","answer":"存在"},{"type":"填空题","question":"在框架表示法中，下层概念可以自动获取上层概念的属性，这一特性称为________。","answer":"继承"},{"type":"填空题","question":"一阶谓词逻辑由谓词、________、连接词和量词四个核心部分组成。","answer":"项"},{"type":"填空题","question":"现实世界中，信息往往是不完整或不精确的，因此知识具有________，例如“如果打喷嚏，可能是感冒了”需要附带置信度。","answer":"不确定性"},{"type":"填空题","question":"先假设一个目标，然后反向寻找支持该目标的事实的推理方式是________推理。","answer":"反向"},{"type":"填空题","question":"知识加工主要包括本体构建、知识推理和__________。","answer":"质量评估"},{"type":"填空题","question":"专门用于描述特定场景下标准事件序列（如“去餐厅吃饭”）的知识表示方法是________。","answer":"脚本表示法"}],"第三章机考题库":[{"type":"选择题","question":"为了确保 A* 算法能找到最优解，启发函数需要满足可接受性和什么性质？","options":{"A":"单调性","B":"凸性","C":"稳定性","D":"线性"},"answer":"A"},{"type":"选择题","question":"搜索求解问题涉及多个方面，主要包括问题表示、搜索策略与什么？","options":{"A":"搜索算法","B":"启发函数","C":"智能体","D":"领域知识"},"answer":"A"},{"type":"选择题","question":"在与或树中，“分解”是将复杂的大问题分解为一组简单的小问题，这属于哪种逻辑关系？","options":{"A":"变换","B":"或（OR）","C":"互斥","D":"与（AND）"},"answer":"D"},{"type":"选择题","question":"在与或树中，如果一个节点是与子节点，那么该节点可解的条件是：","options":{"A":"至少有一个子节点为可解节点","B":"所有子节点均为可解节点","C":"所有子节点都不可解","D":"没有后继节点"},"answer":"B"},{"type":"选择题","question":"盲目搜索（Blind Search）又常叫作“蛮力法”，它属于哪种搜索？","options":{"A":"无信息搜索","B":"启发式搜索","C":"最佳优先搜索","D":"有信息搜索"},"answer":"A"},{"type":"选择题","question":"状态空间可以用一个四元组表示，其中 S0 代表什么？","options":{"A":"状态集合","B":"操作算子集合","C":"问题的初始状态集合","D":"目标状态集合"},"answer":"C"},{"type":"选择题","question":"贪心算法（Greedy Algorithm）总是做出在当前看来最好的选择，它寻找的是哪种解？","options":{"A":"整体最优解","B":"随机解","C":"唯一最优解","D":"局部最优解"},"answer":"D"},{"type":"选择题","question":"博弈问题求解描述博弈过程的与或树被称为：","options":{"A":"博弈树","B":"决策树","C":"归约树","D":"状态空间图"},"answer":"A"},{"type":"选择题","question":"深度优先搜索（DFS）通常通过什么数据结构实现？","options":{"A":"队列（Queue）","B":"栈（Stack）","C":"链表","D":"数组"},"answer":"B"},{"type":"选择题","question":"A* 搜索算法的评估函数是 f(n)=g(n)+h(n)，其中 h(n) 代表什么？","options":{"A":"实际代价","B":"从当前节点 n 到目标节点的启发式代价估计","C":"总代价","D":"节点数量"},"answer":"B"},{"type":"选择题","question":"Dijkstra 算法是哪种算法的代表，主要用于解决最短路径问题？","options":{"A":"A* 搜索算法","B":"盲目搜索","C":"深度优先搜索（DFS）","D":"贪心算法"},"answer":"D"},{"type":"选择题","question":"启发函数中，哪种距离适用于连续空间？","options":{"A":"汉明距离","B":"切比雪夫距离","C":"欧几里得距离","D":"曼哈顿距离"},"answer":"C"},{"type":"选择题","question":"哪种盲目搜索算法通过逐层遍历，优先处理距离起始节点较近的节点？","options":{"A":"贪心算法","B":"广度优先搜索（BFS）","C":"极小极大分析","D":"深度优先搜索（DFS）"},"answer":"B"},{"type":"选择题","question":"搜索求解问题中，如果要求搜索的操作序列总代价最低，则称为：","options":{"A":"启发式搜索问题","B":"最优搜索问题","C":"盲目搜索问题","D":"简单搜索问题"},"answer":"B"},{"type":"选择题","question":"A* 搜索算法的评估函数是 f(n)=g(n)+h(n)，其中 g(n) 代表什么？","options":{"A":"目标节点的真实代价","B":"启发式代价估计","C":"总代价","D":"从起始节点到当前节点 n 的实际代价"},"answer":"D"},{"type":"选择题","question":"广度优先搜索（BFS）和深度优先搜索（DFS）是哪种搜索策略的典型代表？","options":{"A":"启发式搜索","B":"盲目搜索","C":"A* 搜索","D":"贪心算法"},"answer":"B"},{"type":"选择题","question":"启发式搜索（Heuristic Search）又称为：","options":{"A":"暴力搜索","B":"随机搜索","C":"无信息搜索","D":"有信息搜索"},"answer":"D"},{"type":"选择题","question":"启发函数中，哪种距离适用于离散空间？","options":{"A":"切比雪夫距离","B":"汉明距离","C":"欧几里得距离","D":"曼哈顿距离"},"answer":"D"},{"type":"选择题","question":"广度优先搜索（BFS）通常通过什么数据结构实现？","options":{"A":"队列（Queue）","B":"数组","C":"链表","D":"栈（Stack）"},"answer":"A"},{"type":"选择题","question":"在极小极大分析法中，对于“或”节点（己方行动），应选择其子节点中的哪种得分作为父节点的得分？","options":{"A":"最小得分","B":"平均得分","C":"最大得分","D":"随机得分"},"answer":"C"},{"type":"选择题","question":"状态空间表示法通常适用于表示哪一类问题？","options":{"A":"比较简单的问题","B":"多智能体问题","C":"无限状态的问题","D":"极度复杂的问题"},"answer":"A"},{"type":"选择题","question":"搜索求解问题的六个特征中，第一个特征是：","options":{"A":"状态空间连续","B":"环境动态","C":"初始状态确定","D":"路径代价函数未知"},"answer":"C"},{"type":"选择题","question":"问题表示方法在搜索求解问题中，常见的问题表示方法包括状态空间表示法和哪一种方法？","options":{"A":"神经网络法","B":"逻辑推理法","C":"决策图表示法","D":"与或树表示法"},"answer":"D"},{"type":"选择题","question":"搜索求解问题的本质是通过结构化策略在给定的什么中寻找最优解？","options":{"A":"知识库","B":"神经网络","C":"搜索空间","D":"记忆单元"},"answer":"C"},{"type":"选择题","question":"香农的玩具老鼠穿越迷宫的行为，使用的是哪一种盲目搜索方法？","options":{"A":"贪心算法","B":"广度优先搜索","C":"A* 搜索","D":"深度优先搜索"},"answer":"D"},{"type":"选择题","question":"搜索策略与算法常见的搜索策略包括盲目搜索和哪种搜索？","options":{"A":"暴力搜索","B":"随机搜索","C":"启发式搜索","D":"优化搜索"},"answer":"C"},{"type":"选择题","question":"在八数码问题中，所有的摆法构成的集合称为：","options":{"A":"求解路径","B":"状态空间 S","C":"目标状态 G","D":"操作算子 O"},"answer":"B"},{"type":"选择题","question":"搜索与博弈是人工智能实现什么决策的核心方法？","options":{"A":"随机决策","B":"道德决策","C":"智能决策","D":"经验决策"},"answer":"C"},{"type":"选择题","question":"在状态空间表示法中，用来将一个状态转换为另一个状态的元素是？","options":{"A":"状态集合 S","B":"操作算子集合 O","C":"初始状态集合 S0","D":"目标状态集合 G"},"answer":"B"},{"type":"选择题","question":"启发式搜索利用了问题的领域知识和什么来指导搜索方向？","options":{"A":"过去的经验","B":"搜索时间","C":"随机数","D":"启发函数"},"answer":"D"},{"type":"选择题","question":"与或树（AND OR Tree）的搜索过程可以看作一个问题求解的：","options":{"A":"穷举过程","B":"优化过程","C":"归约过程","D":"扩展过程"},"answer":"C"},{"type":"选择题","question":"在与或树中，“变换”是指将困难的问题变换为容易的等价问题，这属于哪种逻辑关系？","options":{"A":"互斥","B":"或（OR）","C":"与（AND）","D":"变换"},"answer":"B"},{"type":"判断题","question":"Dijkstra 算法是启发式搜索的代表算法。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"在状态空间四元组 (S, O, S0, G) 中，S是问题的状态集合。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"在状态空间四元组 (S, O, S0, G) 中，S0是问题的初始状态集合。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"在状态空间四元组 (S, O, S0, G) 中，O 是问题的目标状态集合。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"在与或树中，与节点是不可解节点，当且仅当它的所有子节点都是不可解节点。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"在与或树中，叶节点是可解节点。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"α-β 剪枝法的目的是为了降低极小极大分析法的运算效率","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"搜索求解问题具有六个特征，其中之一是路径代价函数必须已知。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"广度优先搜索（BFS）和深度优先搜索（DFS）都是盲目式搜索的典型代表。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"极小极大分析法中，评估博弈树端节点的得分称为倒推值。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"启发式搜索利用了领域知识和评估函数来指导搜索方向。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"广度优先搜索（BFS）能够有效地找到无权图中的最短路径。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度优先搜索（DFS）可能会错过最短路径。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"曼哈顿距离只能水平或垂直移动，适用于网格地图类的问题。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"状态空间表示法中，状态是指在系统中决定系统状态最小数目变量的有序集合。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"搜索策略盲目搜索又叫作“蛮力法”，是一无信息搜索。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"A* 算法的可接受性要求启发函数 h(n) 不能低估从当前节点 n 到目标节点的代价。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"在状态空间四元组 (S, O, S0, G) 中，G是问题的初始状态集合。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"贪心算法不追求整体最优解，而寻找特定意义上的局部最优解。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"A* 搜索算法是根据广度优先搜索（BFS）算法改进的。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"Dijkstra 算法是贪心式搜索的代表算法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"搜索策略盲目搜索又叫作“蛮力法”，是一种有信息搜索。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"在 A* 搜索算法中，每个节点的总代价是 f(n)，其算法优先扩展 f(n) 最小的节点。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"博弈问题求解在博弈树中，己方扩展的节点之间是“或”关系。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"广度优先搜索（BFS）和深度优先搜索（DFS）都是启发式搜索的典型代表。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"A* 算法的可接受性要求启发函数 h(n) 不能高估从当前节点 n 到目标节点的代价。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"α-β 剪枝法的目的是为了提高极小极大分析法的运算效率","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度优先搜索（DFS）的优点是空间复杂度较低。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"在状态空间四元组 (S, O, S0, G) 中，G是问题的目标状态集合。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"填空题","question":"描述博弈过程中所有可能状态和行动的树形结构称为________树。","answer":"博弈"},{"type":"填空题","question":"广度优先搜索（BFS）采用________数据结构来管理待扩展的节点，按层逐层扩展。","answer":"队列"},{"type":"填空题","question":"深度优先搜索（DFS）采用________数据结构来管理待扩展的节点，一条路走到黑。","answer":"栈"},{"type":"填空题","question":"八数码问题中，所有可能的数字排列构成了问题的________空间。","answer":"状态"},{"type":"填空题","question":"启发式搜索利用与问题相关的额外信息（称为________函数）来引导搜索方向。","answer":"启发"},{"type":"填空题","question":"DFS的主要优点是________复杂度低，只需存储当前路径上的节点。","answer":"空间"},{"type":"填空题","question":"贪心最佳优先搜索在选择节点时只考虑启发函数h(n)的值，而不考虑从起点到当前节点的________代价。","answer":"实际"},{"type":"填空题","question":"一个搜索问题通常包含四个要素：初始状态集合、目标状态集合、状态集合和________。","answer":"操作算子集合"},{"type":"填空题","question":"在允许八方向移动（含对角线）的网格中，常用的启发函数是________距离。","answer":"切比雪夫"},{"type":"填空题","question":"与或树中的“或节点”对应博弈树中的________节点，因为两者都只需要一个子节点成功即可。","answer":"MAX"},{"type":"填空题","question":"围棋、国际象棋等棋类游戏属于________信息博弈。","answer":"完全"},{"type":"填空题","question":"在连续空间中，常用的启发函数是________距离。","answer":"欧几里得"},{"type":"填空题","question":"或树的核心思想是______。","answer":"替换"},{"type":"填空题","question":"当α ≥ β时，当前分支可以被________，因为已经不可能产生更好的结果。","answer":"剪枝"},{"type":"填空题","question":"从家到公司的路线规划中，“家的位置”属于搜索问题三要素中的________状态。","answer":"初始"},{"type":"填空题","question":"A*算法保证找到最优解的条件是启发函数h(n)满足________性，即不会高估实际代价。","answer":"可接受"},{"type":"填空题","question":"α-β剪枝是对极小极大算法的优化，通过剪掉不可能影响最终决策的分支来减少_______。","answer":"搜索量"},{"type":"填空题","question":"BFS的主要缺点是________复杂度高，需要存储大量节点。","answer":"空间"},{"type":"填空题","question":"德州扑克属于________信息博弈，因为玩家不知道对手的底牌。","answer":"不完全"},{"type":"填空题","question":"小极大算法中，MAX节点取子节点中的________值，MIN节点取子节点中的最小值。","answer":"最大"},{"type":"填空题","question":"在博弈树中，轮到MAX玩家决策的节点称为________节点，轮到MIN玩家决策的节点称为MIN节点。","answer":"Max"},{"type":"填空题","question":"在A*算法中，评价函数f(n) = g(n) + h(n)，其中g(n)表示从起点到节点n的________代价。","answer":"实际"},{"type":"填空题","question":"极小极大算法从叶子节点开始，递归向上传递值，最终得到根节点的________值。","answer":"最优"},{"type":"填空题","question":"在α-β剪枝中，如果当前节点是MIN节点，且某个子节点的返回值小于或等于当前的________值，则可以剪掉该节点的其余子节点。","answer":"α"},{"type":"填空题","question":"在网格地图中，如果只允许上下左右移动，常用的启发函数是________距离。","answer":"曼哈顿"},{"type":"填空题","question":"在α-β剪枝中，α表示MAX节点当前已找到的最佳________值，β表示MIN节点当前已找到的最佳上界。","answer":"下界"},{"type":"填空题","question":"BFS和DFS属于_____搜索算法。","answer":"盲目"},{"type":"填空题","question":"与树的核心思想是_______。","answer":"分解"},{"type":"填空题","question":"单智能体决策问题通常称为________问题，而多智能体相互影响的决策问题称为博弈问题。","answer":"搜索"}],"第四章机考题库":[{"type":"选择题","question":"群智能算法在不依赖中心控制的情况下，通过什么机制来寻找解决方案？","options":{"A":"单一的","B":"固定的","C":"强硬的","D":"动态的和自适应的"},"answer":"D"},{"type":"选择题","question":"蚁群算法的优点之一是正反馈性显著，这指的是什么？","options":{"A":"算法随机性强","B":"寻优次数减少","C":"信息素挥发快","D":"信息素积累使好路径被选择的概率更大"},"answer":"D"},{"type":"选择题","question":"在遗传算法中，用于评估每个个体优劣的函数是什么？","options":{"A":"目标函数","B":"惩罚函数","C":"损失函数","D":"适应度函数"},"answer":"D"},{"type":"选择题","question":"人工蜂群算法中，负责搜索蜜源并将信息分享给跟随蜂的角色是？","options":{"A":"工蜂","B":"雌蜂","C":"跟随蜂","D":"侦察蜂"},"answer":"A"},{"type":"选择题","question":"群智能算法除了应用于函数优化、组合优化和机器学习外，还被广泛应用于哪个领域？","options":{"A":"高能物理","B":"量子计算","C":"气象预测","D":"机器人控制"},"answer":"D"},{"type":"选择题","question":"群智能算法通过群体中个体之间的简单交互和信息传递，展现出复杂的哪两种行为？","options":{"A":"自组织和协同","B":"随机和集中","C":"单中心和被动","D":"暴力和穷举"},"answer":"A"},{"type":"选择题","question":"遗传算法中的哪一操作是指对个体的基因进行随机修改，以防止算法陷入局部最优？","options":{"A":"交叉","B":"重组","C":"选择","D":"变异"},"answer":"D"},{"type":"选择题","question":"二进制编码的缺点之一，即相邻整数的二进制编码可能具有较大的汉明距离，这又被称为？","options":{"A":"汉明悬崖","B":"精度损失","C":"编码失真","D":"局部最优"},"answer":"A"},{"type":"选择题","question":"在粒子群优化算法中，解空间中的每个可能解被看作一个什么？","options":{"A":"染色体","B":"粒子","C":"基因","D":"细胞"},"answer":"B"},{"type":"选择题","question":"在遗传算法中，用于模拟个体（解）的结构的是？","options":{"A":"器官","B":"染色体","C":"神经元","D":"细胞"},"answer":"B"},{"type":"选择题","question":"PSO速度更新公式中，ω 代表什么，它用于控制前一步速度的影响？","options":{"A":"速度向量","B":"学习因子","C":"随机数","D":"惯性权重"},"answer":"D"},{"type":"选择题","question":"遗传算法（GA）的设计灵感主要来源于哪种理论？","options":{"A":"相对论","B":"量子力学","C":"牛顿定律","D":"生物进化理论"},"answer":"D"},{"type":"选择题","question":"蚁群优化算法（ACO）是模拟自然界中哪种生物寻找食物的过程？","options":{"A":"鸟","B":"鱼","C":"蜜蜂","D":"蚂蚁"},"answer":"D"},{"type":"选择题","question":"蚁群优化算法特别适合解决哪一类问题？","options":{"A":"组合优化问题","B":"函数优化问题","C":"机器学习问题","D":"机器人控制问题"},"answer":"A"},{"type":"选择题","question":"根据群智能算法的定义，其中“群”指的是一组什么样的群体？","options":{"A":"只能进行间接通信的群体","B":"必须由一个中心控制器协调的群体","C":"仅指具有高度智能的个体组成的群体","D":"相互之间可以进行直接或间接通信的群体"},"answer":"D"},{"type":"选择题","question":"实数编码因其使用决策变量的真实值表示，又被称为哪种编码？","options":{"A":"格雷码编码","B":"浮点数编码","C":"二进制编码","D":"真值编码"},"answer":"D"},{"type":"选择题","question":"粒子群优化算法（PSO）是受到自然界中哪种群体的觅食行为启发而提出的？","options":{"A":"鸟群","B":"鱼群","C":"蚁群","D":"蜂群"},"answer":"A"},{"type":"选择题","question":"人工蜂群算法（ABC）是受哪种生物的觅食行为启发的？","options":{"A":"鱼群","B":"蜘蛛","C":"蜜蜂","D":"蚂蚁"},"answer":"C"},{"type":"选择题","question":"群智能算法通常不依赖于哪种控制机制来寻找解决方案？","options":{"A":"自适应机制","B":"中心控制","C":"分散机制","D":"动态机制"},"answer":"B"},{"type":"选择题","question":"哪种编码方式可以直接表示连续变量或实数参数，避免了将连续空间离散化的过程？","options":{"A":"二进制编码","B":"实数编码","C":"符号编码","D":"混合编码"},"answer":"B"},{"type":"选择题","question":"群智能（SI）算法是受到自然界中哪种对象的行为模式启发而发展起来的一类计算智能算法？","options":{"A":"生物群体","B":"神经网络","C":"物理系统","D":"单个智能体"},"answer":"A"},{"type":"选择题","question":"遗传算法的核心思想借鉴了达尔文的哪一原理？","options":{"A":"物质守恒原理","B":"万有引力原理","C":"能量转换原理","D":"自然选择原理"},"answer":"D"},{"type":"选择题","question":"群智能算法的机制通常具有哪三个主要特点，以解决复杂的优化问题？","options":{"A":"分散的、静态的和刚性的","B":"分散的、动态的和自适应的","C":"间接的、被动的和单一的","D":"集中的、固定的和预设的"},"answer":"B"},{"type":"选择题","question":"人工鱼群算法（AFSA）模拟了鱼群的觅食、避敌和哪种行为？","options":{"A":"群体协作行为","B":"迁徙行为","C":"捕猎行为","D":"筑巢行为"},"answer":"A"},{"type":"选择题","question":"每个粒子在移动时需要考虑自身的历史最优位置（pbest）和哪一个位置？","options":{"A":"初始位置","B":"整个粒子群体的历史最优位置","C":"邻居粒子位置","D":"随机位置"},"answer":"B"},{"type":"选择题","question":"在蚁群优化算法求解旅行商问题（TSP）时，禁忌列表（Tabu List）的作用是什么？","options":{"A":"确定信息素挥发系数","B":"记录信息素强度","C":"记录蚂蚁已经访问过的城市","D":"计算路径长度"},"answer":"C"},{"type":"选择题","question":"蚁群优化算法通过什么物质的积累和挥发来寻找最优路径？","options":{"A":"气味","B":"能量","C":"信息素","D":"蜜源"},"answer":"C"},{"type":"选择题","question":"在遗传算法中，通过交换两个父代个体的部分基因以生成新的子代个体的操作是什么？","options":{"A":"变异","B":"选择","C":"交叉","D":"重组"},"answer":"C"},{"type":"选择题","question":"人工鱼群算法的四种基本行为包括觅食、聚集、跟随和哪一种行为？","options":{"A":"产卵行为","B":"静止行为","C":"随机行为","D":"避敌行为"},"answer":"C"},{"type":"选择题","question":"下列哪种算法是受到生物进化论和自然选择原理启发而发展起来的？","options":{"A":"人工鱼群算法（AFSA）","B":"蚁群优化算法（ACO）","C":"遗传算法（GA）","D":"粒子群优化算法（PSO）"},"answer":"C"},{"type":"选择题","question":"蚁群优化算法中，用于控制路径的距离对蚂蚁决策的影响程度的因子是？","options":{"A":"虚拟信息素强度 Q","B":"能见度重要程度因子 β","C":"信息素挥发系数 ρ","D":"信息素重要程度因子 α"},"answer":"B"},{"type":"选择题","question":"常见的遗传算法编码表示方法包括二进制编码和哪一种编码？","options":{"A":"字符编码","B":"索引编码","C":"实数编码","D":"整数编码"},"answer":"C"},{"id":33,"type":"判断题","question":"人工鱼群算法（AFSA）的“随机行为”的目的是为了扩大觅食范围。","options":{"A":"对","B":"错"},"answer":"A"},{"id":34,"type":"判断题","question":"人工蜂群算法（ABC）是模拟蜜蜂寻找蜜源和采蜜的过程。","options":{"A":"对","B":"错"},"answer":"A"},{"id":35,"type":"判断题","question":"在PSO算法中，个体最优位置 pbest 指的是单个粒子自身找到的最佳位置。","options":{"A":"对","B":"错"},"answer":"A"},{"id":36,"type":"判断题","question":"蚁群优化算法非常适合解决组合优化问题，例如旅行商问题。","options":{"A":"对","B":"错"},"answer":"A"},{"id":37,"type":"判断题","question":"在人工鱼群算法中，跟随行为是鱼群躲避危害、保证群体生存的方式。","options":{"A":"对","B":"错"},"answer":"B"},{"id":38,"type":"判断题","question":"人工蜂群算法中的侦察蜂负责在蜂巢附近寻找新的食物源。","options":{"A":"对","B":"错"},"answer":"B"},{"id":39,"type":"判断题","question":"“群智能”是指无智能或低智能的个体通过随机移动表现出高智能行为的特性。","options":{"A":"对","B":"错"},"answer":"B"},{"id":40,"type":"判断题","question":"粒子群优化算法（PSO）是通过粒子在搜索空间中的位置和加速度更新来寻找最优解。","options":{"A":"对","B":"错"},"answer":"B"},{"id":41,"type":"判断题","question":"汉明悬崖是描述实数编码在处理连续空间时出现的精度损失问题。","options":{"A":"对","B":"错"},"answer":"B"},{"id":42,"type":"判断题","question":"蚁群优化算法中的禁忌列表用于记录蚂蚁已访问的城市，防止其重复访问。","options":{"A":"对","B":"错"},"answer":"A"},{"id":43,"type":"判断题","question":"人工鱼群算法（AFSA）的“觅食行为”的目的是为了扩大觅食范围。","options":{"A":"对","B":"错"},"answer":"B"},{"id":44,"type":"判断题","question":"粒子群优化（PSO）算法是模拟了蚁群的觅食行为而发展起来的群智能算法。","options":{"A":"对","B":"错"},"answer":"B"},{"id":45,"type":"判断题","question":"群智能算法通过模拟生物群体的协作机制，为解决复杂的组合优化问题提供了新的思路和方法。","options":{"A":"对","B":"错"},"answer":"A"},{"id":46,"type":"判断题","question":"在遗传算法中，个体的适应度越高，其被选择用于繁殖的可能性就越小。","options":{"A":"对","B":"错"},"answer":"B"},{"id":47,"type":"判断题","question":"群智能算法模拟了鸟群、蚁群、鱼群等生物群体的觅食、迁徙、筑巢等行为。","options":{"A":"对","B":"错"},"answer":"A"},{"id":48,"type":"判断题","question":"蚁群优化算法（ACO）的提出时间晚于粒子群优化算法（PSO）。","options":{"A":"对","B":"错"},"answer":"B"},{"id":49,"type":"判断题","question":"遗传算法的核心组成部分包括选择、交叉和变异的机制。","options":{"A":"对","B":"错"},"answer":"A"},{"id":50,"type":"判断题","question":"蚁群优化算法具有鲁棒性强、正反馈性显著、搜索智能化等优点。","options":{"A":"对","B":"错"},"answer":"A"},{"id":51,"type":"判断题","question":"惯性权重 ω 在PSO速度更新公式中用于控制前一步速度对当前速度的影响。","options":{"A":"对","B":"错"},"answer":"A"},{"id":52,"type":"判断题","question":"在蚁群算法中，信息素重要程度因子 α 值越大，蚂蚁选择信息素较高路线的概率就越大，算法的随机性就越强。","options":{"A":"对","B":"错"},"answer":"B"},{"id":53,"type":"判断题","question":"汉明悬崖是描述二进制编码的缺点。","options":{"A":"对","B":"错"},"answer":"A"},{"id":54,"type":"判断题","question":"交叉操作在遗传算法中通过交换基因，是形成生物多样性的重要因素。","options":{"A":"对","B":"错"},"answer":"A"},{"id":55,"type":"判断题","question":"人工鱼群算法（AFSA）是一种并行随机搜索算法。","options":{"A":"对","B":"错"},"answer":"A"},{"id":56,"type":"判断题","question":"人工鱼群算法（AFSA）的“觅食行为”的目的是在视域内尝试更好的状态。","options":{"A":"对","B":"错"},"answer":"A"},{"id":57,"type":"判断题","question":"人工蜂群算法中的跟随蜂负责在蜂巢附近寻找新的食物源。","options":{"A":"对","B":"错"},"answer":"B"},{"id":58,"type":"判断题","question":"二进制编码具有类似生物染色体结构的优点，便于理解和进行遗传操作。","options":{"A":"对","B":"错"},"answer":"A"},{"id":59,"type":"判断题","question":"变异操作发生的频率很高，是遗传算法提高收敛速度的主要手段。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"填空题","question":"遗传算法是模拟生物________机制而设计的一类优化算法，由霍兰德于1975年提出。","answer":"进化"},{"type":"填空题","question":"粒子群算法中，整个群体中所有粒子找到的最好位置称为________。","answer":"全局最优"},{"type":"填空题","question":"遗传算法中，如果种群过早地失去多样性而陷入局部最优，这种现象被称为________收敛。","answer":"早熟"},{"type":"填空题","question":"PSO优化算法是受________觅食行为启发而设计的优化算法，由Kennedy和Eberhart于1995年提出。","answer":"鸟群"},{"type":"填空题","question":"人工鱼群算法中的四种典型行为包括：觅食行为、聚群行为、________行为和随机行为。","answer":"跟随"},{"type":"填空题","question":"PSO算法中每个粒子的速度有_____作用。","answer":"控制移动方向"},{"type":"填空题","question":"在蚁群算法中，蚂蚁在路径上留下的化学物质称为________，其浓度随时间和经过的蚂蚁数量变化。","answer":"信息素"},{"type":"填空题","question":"蚁群优化算法是受蚂蚁觅食过程中________的沉积与挥发机制启发而设计的算法，由Dorigo于1992年提出。","answer":"信息素"},{"type":"填空题","question":"在人工蜂群算法中，负责在蜜源附近进行局部搜索的蜜蜂称为________蜂。","answer":"工蜂"},{"type":"填空题","question":"群智能算法中没有中央控制器，每个个体仅根据局部信息做出决策，这体现了群智能的________特性。","answer":"分布"},{"type":"填空题","question":"模拟“优胜劣汰”机制，从当前种群中挑选优秀个体作为父代的操作称为________。","answer":"选择"},{"type":"填空题","question":"蚁群优化算法中，路径长度的倒数称为________信息。","answer":"启发"},{"type":"填空题","question":"人工蜂群算法通过________机制自动跳出局部最优，这是其区别于其他群智能算法的重要特点。","answer":"侦察"},{"type":"填空题","question":"为了防止算法过早收敛，蚁群算法引入了________机制，使所有路径上的信息素按一定比例减少。","answer":"挥发"},{"type":"填空题","question":"个体之间的简单交互最终产生出复杂的全局行为，这种现象被称为________。","answer":"涌现"},{"type":"填空题","question":"群智能是指无智能或______的个体通过合作表现出高智能行为特性。","answer":"低智能"},{"type":"填空题","question":"两个父代个体交换部分基因片段以产生子代个体的操作称为________。","answer":"交叉"},{"type":"填空题","question":"粒子群算法中，每个粒子记录自己历史上找到的最好位置，这个位置称为________。","answer":"个体最优"},{"type":"填空题","question":"惯性权重w用于平衡粒子的________搜索能力和局部搜索能力。","answer":"全局"},{"type":"填空题","question":"在遗传算法中，一个候选解被称为________，通常用二进制串、实数串或排列进行编码。","answer":"染色体"},{"type":"填空题","question":"群智能算法是受自然界中________的社会行为启发而设计的一类优化算法。","answer":"生物群体"},{"type":"填空题","question":"粒子群算法的速度更新公式包含三个部分：惯性项、________和社会项。","answer":"认知项"},{"type":"填空题","question":"遗传算法的三个核心操作是：选择、________和变异。","answer":"交叉"},{"type":"填空题","question":"群智能算法具有三个典型特征：分布性、自组织性和________性。","answer":"涌现"},{"type":"填空题","question":"蚁群算法中，蚂蚁从城市i转移到城市j的概率与路径上的信息素浓度和________信息有关。","answer":"启发"},{"type":"填空题","question":"在粒子群算法中，每个候选解被称为一个________，它在搜索空间中飞行。","answer":"粒子"},{"type":"填空题","question":"PSO算法中的每个粒子在搜索空间中具有位置和_______属性。","answer":"速度"},{"type":"填空题","question":"在遗传算法中，评估个体好坏程度的函数称为________函数。","answer":"适应度"},{"type":"填空题","question":"以较小的概率随机改变个体中某些基因的操作称为________，其作用是保持种群多样性、防止早熟收敛。","answer":"变异"}],"第五章机考题库":[{"type":"判断题","question":"语音识别是将语音信号转换为文本的技术。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"SLAM 技术用于机器人在未知环境中的定位与建图。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"SIFT 特征具有尺度不变性和旋转不变性。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"图像预处理是计算机视觉中的高级处理阶段。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"语音合成是将文本转换为语音的技术。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"词法分析是自然语言处理的基础任务之一。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"机器翻译只能基于规则方法实现。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"卷积神经网络（CNN）只能用于图像分类任务。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"统计学习理论主要研究有限样本下的学习问题。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"硬间隔SVM适用于所有数据集。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"自然语言处理（NLP）只研究文本，不涉及语音。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"监督学习需要标注数据。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"特征空间中的线性可分性总是优于原始空间。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"专家系统的核心是知识库和推理机。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"语义分析关注语句在特定语境中的实际意义。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"无监督学习不需要任何先验知识。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"专家系统不需要推理机即可工作。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"过学习是指模型在训练集上表现好，但在测试集上表现差。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"核函数的作用是隐式地将数据映射到高维空间。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"模型部署后不需要监控和维护。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"计算机视觉的目标是让计算机像人类一样'看世界'。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"训练集用于评估模型的最终性能。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"深度学习技术不能用于智能机器人的视觉识别。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"支持向量是距离分类超平面最近的样本点。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"特征提取的目的是简化图像数据并提取关键信息。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"语音识别系统中的语言模型用于判断句子合理性。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"特征提取是从原始数据中提取有用信息的过程。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"数据清洗是数据准备阶段的关键步骤。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"半监督学习结合了有标签和无标签数据。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"直方图均衡化是一种对比度增强方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"强化学习属于监督学习的一种。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"自然语言处理中的语用分析关注字面意义。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"机器翻译的发展经历了规则、统计和深度学习三个阶段。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"回归问题中，损失函数常用均方误差。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"机器学习模型的性能只取决于算法选择。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"模型选择只依赖于数据规模。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"软间隔SVM允许部分样本违反间隔约束。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"分类问题中，输出是连续值。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"智能机器人只能在工业领域应用。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"语音合成系统中的声码器用于将频谱转换为波形。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"核函数必须是对称且半正定的。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"语音识别系统不需要语言模型。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"专家系统只能用于医疗诊断。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"句法分析用于识别句子中词语之间的语法关系。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"经验风险最小化（ERM）总是能保证模型泛化能力。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"机器学习模型的训练过程就是最小化经验风险。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"支持向量机（SVM）只能用于线性可分数据。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"智能机器人可以不具备感知能力。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"交叉验证是一种评估模型泛化能力的方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"机器学习只能用于图像和语音处理。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"图像分割只能基于灰度进行。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"无监督学习只能用于聚类任务。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"自然语言处理中的分词只适用于英文。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"MFCC 是一种常用的语音特征提取方法。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"强化学习通过奖励机制来学习策略。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"人脸识别系统通常包括人脸检测、特征点定位、特征提取和分类识别四个模块。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"智能机器人必须具备感知、决策和执行能力。","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"机器学习的目标是使模型在训练集上表现最优。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"机器学习模型的泛化能力是指对训练数据的拟合能力。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"模型评估只能在测试集上进行。","options":{"A":"对","B":"错"},"answer":"B"},{"type":"填空题","question":"在更新聚类中心时，每个聚类中心等于该类所有样本点的________。","answer":"均值"},{"type":"填空题","question":"SVM中的分类函数通常表示为：f(x)=w^Tx+b，其中w表示________。","answer":"权重向量"},{"type":"填空题","question":"常见的核函数包括线性核、多项式核和________核。","answer":"高斯"},{"type":"填空题","question":"在强化学习中，智能体所处的环境情况称为________。","answer":"状态"},{"type":"填空题","question":"在分类问题中，输出变量通常是________型变量。","answer":"离散"},{"type":"填空题","question":"强化学习中，用于评价行为好坏的信号称为________。","answer":"奖励"},{"type":"填空题","question":"无监督学习的主要任务之一是________分析（如K-means）。","answer":"聚类"},{"type":"填空题","question":"模型训练的目的是学习最优的________。","answer":"模型参数"},{"type":"填空题","question":"SVM的核心思想是_______分类间隔。","answer":"最大化"},{"type":"填空题","question":"自训练方法首先使用________数据训练初始模型。","answer":"有标签"},{"type":"填空题","question":"半监督学习结合了少量________数据和大量无标签数据。","answer":"有标签"},{"type":"填空题","question":"支持向量是指距离分类边界最近的________。","answer":"样本点"},{"type":"填空题","question":"自训练（Self-Training）是一种典型的________学习方法。","answer":"半监督"},{"type":"填空题","question":"用于衡量预测值与真实值差异的函数称为________函数。","answer":"损失"},{"type":"填空题","question":"在机器学习中，用于评估模型性能的数据集称为________集。","answer":"测试"},{"type":"填空题","question":"K-means算法属于________学习方法。","answer":"无监督"},{"type":"填空题","question":"在自训练过程中，模型会对无标签数据进行________，并选取高置信度样本加入训练集。","answer":"预测"},{"type":"填空题","question":"自训练方法的一个潜在问题是错误标签可能被不断________，从而影响模型性能。","answer":"累积"},{"type":"填空题","question":"在有监督学习中，常见的任务包括分类和________。","answer":"回归"},{"type":"填空题","question":"强化学习通过与环境交互，根据________信号进行学习。","answer":"奖励"},{"type":"填空题","question":"根据是否有标签，机器学习可分为有监督学习、________学习和半监督学习。","answer":"无监督"},{"type":"填空题","question":"机器学习是让计算机从________中自动学习规律并进行预测或决策的方法。","answer":"数据"},{"type":"填空题","question":"在软间隔SVM中，引入松弛变量是为了允许一些噪声和________存在。","answer":"异常值"},{"type":"填空题","question":"Q-learning的目标是学习最优的________函数，从而得到最优策略。","answer":"Q值"},{"type":"填空题","question":"在Q-learning中，智能体通常选择使________最大的动作作为当前策略。","answer":"Q值"},{"type":"填空题","question":"K-means算法通过不断重复“分配样本”和“更新________”两个步骤，直到结果稳定。","answer":"聚类中心"},{"type":"填空题","question":"在K-means算法中，参数K表示需要划分的________数量。","answer":"聚类"},{"type":"填空题","question":"强化学习中，智能体采取的行为称为________。","answer":"动作"},{"type":"填空题","question":"在模型训练之前，通常需要进行数据________操作。","answer":"预处理"},{"type":"填空题","question":"有监督学习的训练数据具有明确的________信息。","answer":"标签"},{"type":"填空题","question":"Q-learning是一种典型的________学习方法，通过与环境交互进行学习。","answer":"强化"},{"type":"填空题","question":"将数据划分为训练集和测试集的目的是评估模型的________能力。","answer":"泛化"},{"type":"填空题","question":"机器学习的一般流程通常从________开始。","answer":"数据收集"},{"type":"填空题","question":"支持向量机的主要任务是寻找一个最优的________来划分数据。","answer":"超平面"},{"type":"填空题","question":"在回归问题中，输出变量通常是________型变量。","answer":"连续"},{"type":"选择题","question":"以下哪项是计算机视觉中的高级处理任务？","options":{"A":"图像锐化","B":"对比度增强","C":"图像平滑","D":"图像分类"},"answer":"D"},{"type":"选择题","question":"以下哪项是欠学习的表现？","options":{"A":"训练误差大，测试误差大","B":"训练误差大，测试误差小","C":"训练误差小，测试误差大","D":"训练误差小，测试误差小"},"answer":"A"},{"type":"选择题","question":"以下哪项是半监督学习的优势？","options":{"A":"只需少量标注数据","B":"不需要模型","C":"不需要数据","D":"只需大量标注数据"},"answer":"A"},{"type":"选择题","question":"以下哪项是SVM的优点？","options":{"A":"适用于大规模数据","B":"计算复杂度低","C":"对噪声不敏感","D":"无需调参"},"answer":"C"},{"type":"选择题","question":"智能机器人中，用于感知视觉信息的传感器是：","options":{"A":"摄像头","B":"温度传感器","C":"压力传感器","D":"麦克风"},"answer":"A"},{"type":"选择题","question":"机器学习中，测试集用于：","options":{"A":"调整参数","B":"最终评估模型","C":"特征提取","D":"训练模型"},"answer":"B"},{"type":"选择题","question":"智能机器人中，用于实现抓取动作的组件是：","options":{"A":"麦克风","B":"轮子","C":"机械臂","D":"摄像头"},"answer":"C"},{"type":"选择题","question":"以下哪项是机器学习中'过学习'的直接原因？","options":{"A":"特征太少","B":"模型太简单","C":"训练数据太少","D":"模型太复杂"},"answer":"D"},{"type":"选择题","question":"以下哪项是过学习的表现？","options":{"A":"训练误差小，测试误差小","B":"训练误差大，测试误差小","C":"训练误差大，测试误差大","D":"训练误差小，测试误差大"},"answer":"D"},{"type":"选择题","question":"以下哪项是智能机器人的决策技术？","options":{"A":"机械传动","B":"图像采集","C":"语音播放","D":"深度学习"},"answer":"D"},{"type":"选择题","question":"语音识别中，用于将语音转换为文本的最终步骤是：","options":{"A":"声学建模","B":"特征提取","C":"语音采集","D":"解码"},"answer":"D"},{"type":"选择题","question":"以下哪项不是模型评估指标？","options":{"A":"F1 Score","B":"准确率","C":"召回率","D":"内存使用率"},"answer":"D"},{"type":"选择题","question":"以下哪项是硬间隔SVM的适用条件？","options":{"A":"数据带有噪声","B":"数据量小","C":"数据高维","D":"数据线性可分"},"answer":"D"},{"type":"选择题","question":"专家系统中，用于获取知识的组件是：","options":{"A":"解释器","B":"知识获取模块","C":"数据库","D":"推理机"},"answer":"B"},{"type":"选择题","question":"专家系统中，用于存储领域知识的组件是：","options":{"A":"知识库","B":"推理机","C":"解释器","D":"数据库"},"answer":"A"},{"type":"选择题","question":"计算机视觉中，图像预处理的主要目的是：","options":{"A":"提取特征","B":"进行语义分割","C":"改善图像质量","D":"进行分类"},"answer":"C"},{"type":"选择题","question":"以下哪项是软间隔SVM的适用条件？","options":{"A":"数据线性可分","B":"数据高维","C":"数据量小","D":"数据带有噪声"},"answer":"D"},{"type":"选择题","question":"以下哪项是常用的特征提取算法？","options":{"A":"决策树","B":"SVM","C":"SIFT","D":"K-means"},"answer":"C"},{"type":"选择题","question":"语音合成中，用于生成声学特征的模型是：","options":{"A":"BERT","B":"GPT","C":"Tacotron","D":"YOLO"},"answer":"C"},{"type":"选择题","question":"语音合成中，用于生成波形的神经网络模型是：","options":{"A":"WaveNet","B":"BERT","C":"GPT","D":"YOLO"},"answer":"A"},{"type":"选择题","question":"以下哪项是无监督学习的典型任务？","options":{"A":"分类","B":"聚类","C":"强化学习","D":"回归"},"answer":"B"},{"type":"选择题","question":"以下哪项是自然语言处理中的语义分析任务？","options":{"A":"句法分析","B":"情感分析","C":"分词","D":"词性标注"},"answer":"B"},{"type":"选择题","question":"自然语言处理中，用于处理文本中缺失值的步骤是：","options":{"A":"语料清洗","B":"建模","C":"特征工程","D":"分词"},"answer":"A"},{"type":"选择题","question":"以下哪项是无监督学习的算法？","options":{"A":"K-means","B":"SVM","C":"随机森林","D":"逻辑回归"},"answer":"A"},{"type":"选择题","question":"以下哪项是监督学习的算法？","options":{"A":"DBSCAN","B":"K-means","C":"决策树","D":"PCA"},"answer":"C"},{"type":"选择题","question":"语音合成中，用于将频谱转换为波形的组件是：","options":{"A":"声码器","B":"解码器","C":"分词器","D":"编码器"},"answer":"A"},{"type":"选择题","question":"专家系统中，用于解释推理过程的组件是：","options":{"A":"解释器","B":"推理机","C":"知识库","D":"数据库"},"answer":"A"},{"type":"选择题","question":"智能机器人感知环境的关键组件是：","options":{"A":"传感器","B":"电池","C":"执行器","D":"控制器"},"answer":"A"},{"type":"选择题","question":"下列哪项不是机器学习的主要类型？","options":{"A":"强化学习","B":"监督学习","C":"符号学习","D":"无监督学习"},"answer":"C"},{"type":"选择题","question":"以下哪项是特征选择的目的？","options":{"A":"减少特征维度","B":"增加模型复杂度","C":"减少计算速度","D":"增加数据量"},"answer":"A"},{"type":"选择题","question":"以下哪项是常用的核函数？","options":{"A":"线性核","B":"多项式核","C":"以上都是","D":"高斯核"},"answer":"C"},{"type":"选择题","question":"以下哪项不是图像预处理的方法？","options":{"A":"图像锐化","B":"对比度增强","C":"特征提取","D":"平滑处理"},"answer":"C"},{"type":"选择题","question":"以下哪项是机器学习中“泛化能力”的定义？","options":{"A":"对噪声数据的容忍能力","B":"对训练数据的拟合能力","C":"对未知数据的预测能力","D":"对高维数据的处理能力"},"answer":"C"},{"type":"选择题","question":"语音识别中，用于建模语音信号时序的模型是：","options":{"A":"K-means","B":"SVM","C":"CNN","D":"RNN"},"answer":"D"},{"type":"选择题","question":"语音合成中，用于将文本转换为音素的过程是：","options":{"A":"NLP","B":"ASR","C":"G2P","D":"TTS"},"answer":"C"},{"type":"选择题","question":"以下哪项是核函数的数学性质？","options":{"A":"半正定性","B":"线性性","C":"对称性","D":"A和B"},"answer":"D"},{"type":"选择题","question":"以下哪项不是自然语言处理的研究层次？","options":{"A":"语义分析","B":"语音识别","C":"图像分类","D":"句法分析"},"answer":"C"},{"type":"选择题","question":"机器翻译中，用于将源语言转换为目标语言的模型是：","options":{"A":"自编码器","B":"编码器-解码器","C":"卷积网络","D":"生成对抗网络"},"answer":"B"},{"type":"选择题","question":"强化学习中的“奖励”用于：","options":{"A":"惩罚错误","B":"评估动作好坏","C":"标注数据","D":"选择特征"},"answer":"B"},{"type":"选择题","question":"SLAM 技术主要用于：","options":{"A":"文本生成","B":"图像分割","C":"定位与建图","D":"语音识别"},"answer":"C"},{"type":"选择题","question":"以下哪项是专家系统在医疗领域的应用？","options":{"A":"疾病诊断","B":"语音合成","C":"机器翻译","D":"图像分类"},"answer":"A"},{"type":"选择题","question":"以下哪项是模型评估中常用的回归指标？","options":{"A":"召回率","B":"准确率","C":"均方误差","D":"F1 Score"},"answer":"C"},{"type":"选择题","question":"语音识别中，用于提取语音特征的典型方法是：","options":{"A":"SIFT","B":"LBP","C":"HOG","D":"MFCC"},"answer":"D"},{"type":"选择题","question":"以下哪项是回归问题的损失函数？","options":{"A":"对数损失","B":"0-1损失","C":"均方误差","D":"交叉熵"},"answer":"C"},{"type":"选择题","question":"自然语言处理中，用于识别文本中命名实体的任务是：","options":{"A":"句法分析","B":"词性标注","C":"命名实体识别","D":"语义角色标注"},"answer":"C"},{"type":"选择题","question":"SVM 中，核函数的作用是：","options":{"A":"增加噪声","B":"隐式映射到高维空间","C":"降低计算复杂度","D":"减少数据量"},"answer":"B"},{"type":"选择题","question":"以下哪项是模型部署的步骤？","options":{"A":"序列化模型","B":"清洗数据","C":"评估模型","D":"训练模型"},"answer":"A"},{"type":"选择题","question":"以下哪项是统计学习理论的研究内容？","options":{"A":"以上都是","B":"经验风险最小化","C":"核方法","D":"泛化能力理论"},"answer":"A"},{"type":"选择题","question":"智能机器人中，用于执行动作的组件是：","options":{"A":"电池","B":"执行机构","C":"传感器","D":"控制器"},"answer":"B"},{"type":"选择题","question":"以下哪项是特征空间的作用？","options":{"A":"减少数据量","B":"增加数据维度","C":"提高线性可分性","D":"降低计算复杂度"},"answer":"C"},{"type":"选择题","question":"机器翻译中，用于训练模型的数据是：","options":{"A":"双语平行语料库","B":"图像数据集","C":"语音数据集","D":"单语语料库"},"answer":"A"},{"type":"选择题","question":"以下哪项不是计算机视觉的应用方向？","options":{"A":"语音识别","B":"语义分割","C":"图像分类","D":"目标检测"},"answer":"A"},{"type":"选择题","question":"软间隔SVM中，参数C的作用是：","options":{"A":"控制误差惩罚","B":"控制核函数类型","C":"控制间隔大小","D":"控制支持向量数量"},"answer":"A"},{"type":"选择题","question":"以下哪项是数据划分的目的？","options":{"A":"提高计算速度","B":"减少数据量","C":"增加特征","D":"评估模型泛化能力"},"answer":"D"},{"type":"选择题","question":"语音识别系统中，用于将语音信号转换为特征向量的步骤是：","options":{"A":"声学建模","B":"特征提取","C":"语音采集","D":"解码"},"answer":"B"},{"type":"选择题","question":"以下哪项是模型选择的方法？","options":{"A":"数据清洗","B":"特征提取","C":"模型部署","D":"交叉验证"},"answer":"D"},{"type":"选择题","question":"机器学习中，训练集用于：","options":{"A":"训练模型","B":"测试模型","C":"评估模型","D":"部署模型"},"answer":"A"},{"type":"选择题","question":"自然语言处理中，将句子分解为词汇单元的过程称为：","options":{"A":"语用分析","B":"词法分析","C":"语义分析","D":"句法分析"},"answer":"B"},{"type":"选择题","question":"人脸识别系统中，用于精确定位五官的步骤是：","options":{"A":"人脸检测","B":"特征点定位","C":"分类识别","D":"特征提取"},"answer":"B"},{"type":"选择题","question":"以下哪项是数据清洗的任务？","options":{"A":"缺失值处理","B":"特征提取","C":"模型训练","D":"模型部署"},"answer":"A"},{"type":"选择题","question":"专家系统的核心组成部分是：","options":{"A":"传感器和执行器","B":"数据库和界面","C":"算法和模型","D":"知识库和推理机"},"answer":"D"},{"type":"选择题","question":"以下哪项是机器学习中“损失函数”的作用？","options":{"A":"选择特征","B":"优化模型参数","C":"评估模型性能","D":"清洗数据"},"answer":"B"},{"type":"选择题","question":"以下哪项是特征提取的目的？","options":{"A":"提取有用信息","B":"删除数据","C":"降低模型复杂度","D":"增加数据量"},"answer":"A"},{"type":"选择题","question":"以下哪项是机器学习中“欠学习”的原因？","options":{"A":"特征太多","B":"模型太复杂","C":"模型太简单","D":"训练数据太多"},"answer":"C"},{"type":"选择题","question":"图像分类与识别中，用于提取特征的典型模型是：","options":{"A":"GMM","B":"CNN","C":"RNN","D":"HMM"},"answer":"B"},{"type":"选择题","question":"语音识别中，用于建模语音信号概率的模型是：","options":{"A":"CNN","B":"K-means","C":"HMM","D":"SVM"},"answer":"C"},{"type":"选择题","question":"以下哪项是模型调优的方法？","options":{"A":"删除特征","B":"调整超参数","C":"以上都是","D":"增加数据量"},"answer":"C"},{"type":"选择题","question":"以下哪项是半监督学习的算法？","options":{"A":"线性回归","B":"K-means","C":"自训练","D":"决策树"},"answer":"C"},{"type":"选择题","question":"以下哪项是机器学习应用的一般过程？","options":{"A":"模型训练 → 数据准备 → 模型选择 → 模型部署","B":"模型选择 → 数据准备 → 模型训练 → 模型部署","C":"数据准备 → 模型选择 → 模型训练 → 模型部署","D":"数据准备 → 模型训练 → 模型选择 → 模型部署"},"answer":"C"},{"type":"选择题","question":"智能机器人中，用于实现自主导航的技术是：","options":{"A":"HMM","B":"SLAM","C":"GMM","D":"CNN"},"answer":"B"},{"type":"选择题","question":"以下哪项是分类问题的损失函数？","options":{"A":"绝对值误差","B":"余弦相似度","C":"均方误差","D":"交叉熵"},"answer":"D"},{"type":"选择题","question":"自然语言处理中，用于理解句子结构的分析是：","options":{"A":"词法分析","B":"句法分析","C":"语用分析","D":"语义分析"},"answer":"B"},{"type":"选择题","question":"图像分割中，用于处理连在一起物体的算法是：","options":{"A":"分水岭算法","B":"直方图均衡化","C":"拉普拉斯滤波","D":"均值滤波"},"answer":"A"},{"type":"选择题","question":"机器翻译中，基于深度学习的典型模型是：","options":{"A":"Apriori","B":"Seq2Seq","C":"FP-Growth","D":"KNN"},"answer":"B"},{"type":"选择题","question":"以下哪项是模型评估中常用的分类指标？","options":{"A":"以上都是","B":"准确率","C":"R²","D":"均方误差"},"answer":"B"},{"type":"选择题","question":"专家系统中，用于模拟专家决策的组件是：","options":{"A":"知识库","B":"解释器","C":"推理机","D":"数据库"},"answer":"C"},{"type":"选择题","question":"以下哪项是强化学习的算法？","options":{"A":"线性回归","B":"Q-learning","C":"朴素贝叶斯","D":"KNN"},"answer":"B"},{"type":"选择题","question":"以下哪项是语音合成系统的组成部分？","options":{"A":"解码器","B":"分词器","C":"声码器","D":"语法分析器"},"answer":"C"},{"type":"选择题","question":"以下哪项是监督学习的典型任务？","options":{"A":"分类","B":"降维","C":"聚类","D":"关联规则"},"answer":"A"},{"type":"选择题","question":"以下哪项是经验风险最小化（ERM）的缺点？","options":{"A":"容易过拟合","B":"只适用于回归","C":"无法训练","D":"计算复杂"},"answer":"A"}],"第六章机考题库":[{"type":"选择题","question":"全连接层在卷积神经网络中的作用是什么？","options":{"A":"体育","B":"化学","C":"音乐","D":"结合所有局部特征"},"answer":"D"},{"type":"选择题","question":"注意力机制的作用是什么？","options":{"A":"生物","B":"美术","C":"体育","D":"让模型关注关键信息"},"answer":"D"},{"type":"选择题","question":"以下哪个是LSTM的组成部分？","options":{"A":"记忆单元","B":"以上都是","C":"输入门","D":"遗忘门"},"answer":"B"},{"type":"选择题","question":"Transformer模型的主要特点是什么？","options":{"A":"依赖注意力机制","B":"物理","C":"数学","D":"化学"},"answer":"A"},{"type":"选择题","question":"以下哪个是Transformer的常见应用？","options":{"A":"图像识别","B":"机器翻译","C":"视频分析","D":"以上都是"},"answer":"B"},{"type":"选择题","question":"激活函数的主要作用是什么？","options":{"A":"音乐","B":"引入非线性","C":"美术","D":"体育"},"answer":"B"},{"type":"选择题","question":"LSTM是为了解决RNN的什么问题而提出的？","options":{"A":"梯度消失问题","B":"游戏","C":"计算机网络","D":"计算速度慢"},"answer":"A"},{"type":"选择题","question":"池化层在卷积神经网络中的作用是什么？","options":{"A":"减少数据维度","B":"游戏","C":"减少激活函数","D":"物理"},"answer":"A"},{"type":"选择题","question":"什么是CNN的全称？","options":{"A":"数学","B":"机制","C":"语文","D":"卷积神经网络"},"answer":"D"},{"type":"选择题","question":"什么是池化层的主要作用？","options":{"A":"历史","B":"化学","C":"压缩数据和参数，防止过拟合","D":"提取"},"answer":"C"},{"type":"选择题","question":"什么是GRU的全称？","options":{"A":"历史","B":"语文","C":"数学","D":"门控循环单元"},"answer":"D"},{"type":"选择题","question":"什么是LSTM的全称？","options":{"A":"长期短期记忆网络","B":"体育","C":"美术","D":"生物"},"answer":"A"},{"type":"选择题","question":"人工神经网络的基本原理是什么？","options":{"A":"模仿植物生长","B":"模仿体育","C":"模仿美术","D":"模仿生物神经网络"},"answer":"D"},{"type":"选择题","question":"以下哪个是卷积神经网络的常见层？","options":{"A":"生物","B":"池化层","C":"数学","D":"语文"},"answer":"B"},{"type":"选择题","question":"以下哪个是Transformer的常见层？","options":{"A":"注意力层","B":"植物","C":"生物","D":"体育"},"answer":"A"},{"type":"选择题","question":"感知器不能处理什么问题？","options":{"A":"视频","B":"游戏","C":"测试","D":"非线性问题"},"answer":"D"},{"type":"选择题","question":"以下哪个是卷积神经网络的常见结构？","options":{"A":"卷积层","B":"激活函数层","C":"以上都是","D":"池化层"},"answer":"C"},{"type":"选择题","question":"卷积神经网络中的卷积操作是什么？","options":{"A":"游戏","B":"体育","C":"裁剪","D":"将一个可移动的小窗口与图像进行逐元素相乘再相加"},"answer":"D"},{"type":"选择题","question":"RNN与前馈神经网络的主要区别是什么？","options":{"A":"RNN处理图像数据","B":"RNN有内部记忆功能","C":"游戏","D":"RNN不使用激活函数"},"answer":"B"},{"type":"选择题","question":"卷积神经网络主要用于处理哪种类型的数据？","options":{"A":"化学","B":"图像数据","C":"数学","D":"物理"},"answer":"B"},{"type":"选择题","question":"池化层中常用的池化方法是什么？","options":{"A":"数学","B":"最大值池化和平均值池化","C":"物理","D":"化学"},"answer":"B"},{"type":"选择题","question":"人工神经网络的输入是什么？","options":{"A":"体育","B":"美术","C":"生物","D":"数据"},"answer":"D"},{"type":"选择题","question":"以下哪个是卷积神经网络的核心思想？","options":{"A":"植物","B":"生物","C":"美术","D":"模拟人类视觉处理"},"answer":"D"},{"type":"选择题","question":"什么是深度学习的\"黑箱\"性质？","options":{"A":"体育","B":"模型很难解释","C":"美术","D":"生物"},"answer":"B"},{"type":"选择题","question":"什么是BP神经网络的全称？","options":{"A":"反向传播神经网络","B":"游戏","C":"历史","D":"生物"},"answer":"A"},{"type":"选择题","question":"以下哪个是注意力机制的评分函数？","options":{"A":"生物","B":"缩放点积注意力","C":"历史","D":"美术"},"answer":"B"},{"type":"选择题","question":"以下哪个是常见激活函数？","options":{"A":"Sigmoid","B":"Tanh","C":"ReLU","D":"以上都是"},"answer":"D"},{"type":"选择题","question":"人工神经网络的灵感来自哪里？","options":{"A":"植物","B":"电脑","C":"人类的神经元","D":"骨骼"},"answer":"C"},{"type":"选择题","question":"人工神经网络中，什么负责计算输入和权重的加权和？","options":{"A":"体育","B":"神经元","C":"美术","D":"生物"},"answer":"B"},{"type":"选择题","question":"以下哪个是卷积神经网络的常见应用？","options":{"A":"生物","B":"体育","C":"图像识别","D":"美术"},"answer":"C"},{"type":"判断题","question":"AlexNet采用ReLU激活函数","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"感知器可以解决所有分类问题","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"GRU是一种改进的RNN，通过引入门控机制解决梯度消失问题","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"卷积神经网络(CNN)主要用于处理图像相关的机器学习问题","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"Sigmoid函数将任意实数映射到0到1之间的值","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度学习是一种基于多层神经网络的机器学习方法","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"ReLU是常用的激活函数","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"卷积操作是将一个可移动的小窗口与图像进行逐元素相乘再相加的过程","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"卷积神经网络只能用于人脸识别","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"LeNet-5是用于手写数字识别的卷积神经网络","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"RNN是专门用于处理序列数据的神经网络架构","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"ResNet通过引入残差连接结构解决深度神经网络的训练难题","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"梯度消失问题是深度学习面临的挑战之一","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"Transformer模型最早用于图像分类任务","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"MNIST数据集中的每个样本是一张28×28像素的灰度手写数字图像","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度学习模型总是比传统机器学习模型准确","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"感知器是一种线性模型","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"注意力机制使网络模型能够关注到关键信息","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"RNN不能处理文本数据","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"深度学习模型可以处理图像识别、视频内容理解、序列数据分析等领域的问题","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"Tanh函数解决了Sigmoid函数\"不以0为中心输出\"的问题","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"所有神经网络都使用Sigmoid作为激活函数","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"Stable Diffusion不能生成图像","options":{"A":"对","B":"错"},"answer":"B"},{"type":"判断题","question":"CNN模型主要包括输入层、卷积层、激活函数层、池化层、全连接层和输出层","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"池化层的作用包括特征不变性、特征降维和防止过拟合","options":{"A":"对","B":"错"},"answer":"A"},{"type":"判断题","question":"深度学习模型的黑箱性质是其优点","options":{"A":"对","B":"错"},"answer":"A"},{"type":"填空题","question":"深度学习相比传统机器学习，最大的优势在于能够自动进行________学习","answer":"特征"},{"type":"填空题","question":"点积注意力直接将查询与键进行________计算作为得分，效率较高，但当维度较大时容易导致梯度过小。","answer":"点积"},{"type":"填空题","question":"常用的激活函数包括Sigmoid、Tanh和ReLU等，其中________以其计算简单和缓解梯度消失的优点被广泛使用。","answer":"ReLU"},{"type":"填空题","question":"循环神经网络（RNN）专门用于处理________数据，如文本、语音和时间。","answer":"序列"},{"type":"填空题","question":"神经网络中，________函数为网络引入非线性，使得多层网络能够逼近任意复杂函数。","answer":"激活"},{"type":"填空题","question":"在卷积运算中，________的大小决定了感受野的范围，步长（stride）控制卷积核滑动的间隔。","answer":"卷积核"},{"type":"填空题","question":"注意力机制的核心思想是让模型在处理序列时，对不同位置的信息赋予不同的________，聚焦于重要部分。","answer":"权重"},{"type":"填空题","question":"Transformer的编码器由多个相同层堆叠而成，每层包含两个子层：多头自注意力层和________前馈网络层。","answer":"位置"},{"type":"填空题","question":"卷积层通过________操作提取输入数据的局部特征，每个卷积核对应一种特征提取器。","answer":"卷积"},{"type":"填空题","question":"卷积神经网络中的________是指输出特征图中的一个像素对应于输入图像上的区域大小，层数越深感受野越大。","answer":"感受野"},{"type":"填空题","question":"Transformer模型完全基于________，摒弃了传统的循环神经网络（RNN）和卷积神经网络（CNN）结构。","answer":"注意力机制"},{"type":"填空题","question":"常用的池化操作包括最大池化和________池化，其中最大池化取局部区域的最大值作为输出。","answer":"平均"},{"type":"填空题","question":"人工神经网络最基本的计算单元是________，它接收多个输入，进行加权求和后通过激活函数输出。","answer":"神经元"},{"type":"填空题","question":"加法注意力将查询和键拼接后通过一个________计算得分，适用于查询和键的维度不同的情况。","answer":"前馈神经网络"},{"type":"填空题","question":"为了解决点积注意力在维度较大时梯度消失的问题，通常采用________点积注意力，即将点积结果除以维度的平方根进行缩放。","answer":"缩放"},{"type":"填空题","question":"________网络引入了残差块（Residual Block），通过跳跃连接（skip connection）解决了深层网络的退化问题。","answer":"ResNet"},{"type":"填空题","question":"2006年，Geoffrey Hinton等人提出了________和深度自编码器，解决了深层神经网络的训练问题，被认为是深度学习的开端。","answer":"深度信念网络"},{"type":"填空题","question":"Transformer的解码器相比编码器多了一个________注意力子层，用于关注编码器的输出。","answer":"掩码"},{"type":"填空题","question":"________层通过减小特征图的空间尺寸来降低计算量，同时增强特征的平移不变性。","answer":"池化"},{"type":"填空题","question":"感知器是最早的神经网络模型，它只能解决________可分的问题，无法处理异或（XOR）问题。","answer":"线性"},{"type":"填空题","question":"在Transformer中，由于没有循环和卷积结构，需要额外加入________编码来注入序列中单词的位置信息。","answer":"位置"},{"type":"填空题","question":"2012年，AlexNet在________大规模视觉识别竞赛中取得突破性成绩，大幅领先传统方法，开启了深度学习时代。","answer":"ImageNet"},{"type":"填空题","question":"神经网络的________算法通过链式法则计算损失函数对各层参数的梯度，用于更新网络权重。","answer":"反向传播"},{"type":"填空题","question":"卷积神经网络主要由三种层构成：卷积层、________层和全连接层。","answer":"池化"},{"type":"填空题","question":"________网络通过引入“门”机制（输入门、遗忘门、输出门）和细胞状态，有效缓解了长依赖问题。","answer":"长短期记忆"},{"type":"填空题","question":"在注意力机制中，评分函数的作用是计算________与键向量之间的相似度得分，得分越高表示该位置的信息越重要。","answer":"查询向量"},{"type":"填空题","question":"RNN的隐藏层神经元不仅接收当前时刻的输入，还接收上一时刻的________状态信息，形成循环结构。","answer":"隐藏"},{"type":"填空题","question":"早期神经网络训练困难的主要原因是________问题。","answer":"梯度消失和梯度爆炸"},{"type":"填空题","question":"________激活函数的引入有效缓解了梯度消失问题，使得深层网络训练成为可能","answer":"ReLU"}]}