时态结构中的恒常联系因果学习（英文）

人工智能研究中,行动这一概念通常在理论框架中有完全的定义。然而,现实中的行动有时难以完全刻画。智能体需要从过去的经验观察中习得行动的后果。本文提出一种基于时态结构的行动—结果学习理论。在自然数时间结构中,智能体通过观察过去的恒常联系建立因果关系。智能体依据已建立的因果关系指导将来的行动。同时,我们给出关于该理论的一个完全的逻辑演绎系统,并给出基于该逻辑的智能体行动的有效算法。     

多智能体模态逻辑系统K_n中的知识遗忘

如何让智能体像人一样具备遗忘的能力,目前仍然是人工智能所面临的最大挑战之一。遗忘在基于符号逻辑的知识表示与推理领域和基于统计的机器学习领域都有研究。特别在知识表示与推理领域,遗忘扮演着非常重要的角色。在命题逻辑、一阶谓词逻辑、模态逻辑、描述逻辑、回答集逻辑程序设计,以及情景演算等逻辑语言中都有大量的研究。模态逻辑适用于智能体的知识表示与推理。在模态逻辑中,知识遗忘首先在单智能体场景中被提出。随着多智能体系统研究的发展,多智能体模态逻辑中知识遗忘的研究也开始被关注。到目前为止,多智能体模态逻辑系统中的知识遗忘还无法有效计算。本文在多智能体模态逻辑系统K_n中对知识遗忘进一步展开研究。采用知识编译的思想,提出一种新的多智能体模态逻辑范式K_n-DNF。基于K_n-DNF,我们给出K_n中计算知识遗忘的有效算法,其时间复杂度是K_n-DNF公式长度的多项式时间。     

动词体的球态语义学

本文介绍由塔斯基的立体几何导出的球态语义学,并将其应用于自然语言中的动词体现象。球态语义学特别适合应用于英语的进行体。这种方法有以下优点（i）它扩展了区间式语义,并同时避免了其缺陷,（ii）它解决了未完成体难题,（iii）它的解决方法无需诉诸最终结果策略。逻辑方法一般被认为难于处理自然语言的动词体问题。基于点的时间结构以及建立在该结构之上的经典普莱尔时态逻辑（[18]）太弱了。而基于区间的时态语义则缺乏足够的表达力,并且难以解释进行体（[4,8]）.本文给出一种新的基于球上整体－部分关系概念的模型和时态语义。这种球态语义学建基于塔斯基1927年引入的立体几何之上。与基于点和基于区间的语义不同,在球态语义学中很多动词体区分都由统一的逻辑方法刻画。在一个由封闭球构成的论域中,可达关系由相切性概念给出。相应地,我们可定义外切、内切、外径、内径以及同心等基本概念。与区间式语义不同,球是论域的初始概念,球态语义学不是在时间段而是在球中对事件赋值。因此,仅将时间区间作为初始概念而不承认其端点初性性的问题不复存在。英语中的进行体由球上的连续行动来刻画。行动是非终止的,只要球没有由外切相离。相应地,外切相离刻车动作完成。我们区分在均匀球和非均匀球中发生事件的整体-部分关系。非持续动作视为直径为零的同心球。球态语义学根据动作或执行完成的时刻来定义时间概念,其中不需要时间端点的概念。在保持与基于区间的时间模型类似的基础上,球态语义学暗示了一种关于可能世界的定性概念,并且它有利于解决时间的循环概念问题。     

活动类连续行动的逻辑刻画

活动类连续行动的特点是其通过重复履行最小的行动单元而构成。例如约翰跑步、丽丽游泳等行动都属于活动类连续行动。本文中我们将利用STIT逻辑给出活动类连续行动的逻辑刻画。在简要说明活动类连续行动的特点以及STIT逻辑主要内容的基础上，给出刻画活动类连续行动的STIT逻辑并证明该系统的一些重要元定理。     

一个用于表达因果关系的ATL的扩展

CTL模型检测技术已被广泛应用于形式验证领域。交互时态逻辑（ATL）是对CTL的一个扩展,用于表达多主体博弈结构上的性质。ATL使用合作算子来表达多个主体能够通过合作保证系统的设计目标。在实际应用中,我们需要知道主体的行动与系统的输出状态之间具有因果关系。在本文中,我们通过引入新的模态算子扩展ATL,使得这种因果关系得到表达。我们使用两种方式的扩展。其中之一是从主体的能力出发,直观上,如果一些主体可以通过合作的行动来保证系统进入某个状态,同时,这些主体也可以通过合作的行动保证系统不进入这个状态,则这些主体的行动与该系统状态间具有更强的因果关系。我们使用的另一种方式是从系统状态出发。我们考虑要想使系统进入某状态,哪些主体的行动的必不可少的,哪些主体的行动是充分的但非必要的条件。在本文中,我们扩展后的逻辑CATL和SATL表达力强于ATL,但计算复杂性与ATL相同。     

面向人工智能的形式伦理及其逻辑基础

欲使人工智能体具备道德决策能力,一种可设想的工作是基于形式伦理,即把伦理原则或规则形式化。形式伦理学的初衷并不是面向人工智能,但其做法却适用于人工智能。伦理规范的形式表达必须基于某种逻辑语言及其推理规则,经典逻辑是不够的,所以有必要发展和道德、行动相关的逻辑来适应人工智能道德决策的需求。最为著名道义逻辑为形式伦理提供了一定的基础,应该把它和心智逻辑结合起来。从人工智能体实施行动的过程来看,能动性逻辑是表达机器的道德行动与推理比较有前景的逻辑。     

论休谟的因果关系思想

休谟把洛克和贝克莱的经验论发展到了逻辑终局，核心内容就是他的因果关系理论。他认为一切知识都是建立在经验基础之上的，经验的推断就是因果关系，这种关系是根源于人的习惯的，习惯是一种心理信念。     

语境同义性逻辑(英文)

众所周知,基于可能世界语义的内涵逻辑由于对意义的刻画过于粗粝而导致了所谓的"超内涵问题"。为了解决超内涵问题,出现了各种超内涵逻辑,其中由Suszko提出的带等词的命题逻辑(SCI)是超内涵逻辑中最基本的一种。本文是对SCI的精炼,其动机是语境同义性论题(CST)。该论题认为,同义性标准具有语境依赖性。基于认知语境主义,我们给出了CST的一个论证。通过将SCI中的二元等词修改为一个三元结构,用来表示两个陈述相对某个语境表达同一命题,我们给出了CST的希尔伯特式公理系统。我们证明了该系统相对一个代数模型类是可靠的和完全的。该代数模型的论域由命题构成,同时附带一组命题上的全等关系,用以刻画相对于语境的命题同一性。我们运用该逻辑部分解决了分析悖论这一困扰逻辑学家多年的问题。与我们之前的基于相同动机的论文[17]相比,本文给出的形式语言更加丰富,从而能够表达不同语境之间以及不同语境的同义性之间的关系。     

IF多值逻辑及博弈语义

本文基于经典一阶逻辑句法的逻辑优先性分析,把Hintikka的独立联结词和独立量词扩展到多值逻辑中。我们给出IF多值逻辑的句法,并使用不完全信息的语义赋值博弈解释了IF多值逻辑。     

行动理论中的说明问题

行动理论是传统分析哲学的一门新分支。本世纪中叶以来,许多学者认识到无论是自然科学的研究(包括心理学、脑科学),还是社会科学的研究都与人的行动密切相关,廓清对行动的模糊认识,探讨心理状态和行动之间的关系,说明行动者的行动为什么产生,研究行动与社会存在的相互作用,应用行动的研究成果等对社会科学的发展都具有十分重要的意义。许多著名哲学家,纷纷撰文讨论提出了许多独到的见解。从几十年来的行动研究的著述来看,概括起来,主要集中如下5个方面: (一)行动的定义。包括1、行动和身体运动的关系;2、行动和行为的区分;3、行动本身的划分等内容。 (二)行动与行动者的心理状态的关系。行动理论研究者们关注的是心理状态和行动是否具有因果性的问题。因果关系论者在认定心理状态是行动产生的原因这一点上是一致的,但哪些心理状态构成行动产生的原因却各不相同。反之,否认心理状态和行动具有因果关系的研究者们,给出种种反对的理由,并提出他们的理论,诸如情况论和行动者理论等等。     

Agent causation and the alleged impossibility of rational free action

Galen Strawson has claimed that “the impossibility of free will and ultimate moral responsibility can be proved with complete certainty.” Strawson, I take it, thinks that this conclusion can be established by one argument which he has developed. In this argument, he claims that rational free actions would require an infinite regress of rational choices, which is, of course, impossible for human beings. In my paper, I argue that agent causation theorists need not be worried by Strawson’s argument. For agent causation theorists are able to deny a key principle which drives the regress. Oversimplifying things a bit, the principle states that if one is responsible for her rational actions, then she was antecedently responsible for the reasons on which she acted.     

Agency,Teleological Control and Robust Causation

I propose a compatibilist theory of agency and responsibility, according to which an agent is responsible for an effect, if and only if, she is the earliest source of robust causation over it, via an action she carried out in the service of her long term interests. This theory deploys a notion of teleological control, which is a type of guidance-control of the agent over the effect and it involves action plans and means-end reasoning. The theory makes room for degrees of responsibility, and accounts for the distinction between compulsion and determination. The teleological control view is informed by neuroscience and cognitive theory, and while it is indifferent to the distinction between determinism and indeterminism, it contends that the property of natural laws relevant to agency is the presence of successive stages of attractor and bifurcation dynamics. While the former grounds robust causation over effects of actions, the latter limits the temporal range of robustness, allowing us to characterize responsibility in terms of the earliest sources of robust causation.     

Agent causation as the solution to all the compatibilist’s problems

In a recent paper I argued that agent causation theorists should be compatibilists. In this paper, I argue that compatibilists should be agent causation theorists. I consider six of the main problems facing compatibilism: (i) the powerful intuition that one can’t be responsible for actions that were somehow determined before one was born; (ii) Peter van Inwagen’s modal argument, involving the inference rule (β); (iii) the objection to compatibilism that is based on claiming that the ability to do otherwise is a necessary condition for freedom; (iv) “manipulation arguments,” involving cases in which an agent is manipulated by some powerful being into doing something that he or she would not normally do, but in such a way that the compatibilist’s favorite conditions for a free action are satisfied; (v) the problem of constitutive luck; and (vi) the claim that it is not fair to blame someone for an action if that person was determined by forces outside of his or her control to perform that action. And in the case of each of these problems, I argue that the compatibilist has a much more plausible response to that problem if she endorses the theory of agent causation than she does otherwise.     

Interactions between Knowledge,Action and Commitment within Agent Dynamic Logic

This paper considers a new class of agent dynamic logics which provide a formal means of specifying and reasoning about the agents activities and informational, motivational and practical aspects of the behaviour of the agents. We present a Hilbert-style deductive system for a basic agent dynamic logic and consider a number of extensions of this logic with axiom schemata formalising interactions between knowledge and commitment (expressing an agent s awareness of her commitments), and interactions between knowledge and actions (expressing no learning and persistence of knowledge after actions). The deductive systems are proved sound and complete with respect to a Kripke-style semantics. Each of the considered logics is shown to have the small model property and therefore decidable.Presented by Ewa Orlowska     

At one with our actions,but at two with our bodies

Jennifer Hornsby's account of human action frees us from the temptation to think of the person who acts as ‘doing’ the events that are her actions, and thereby removes much of the allure of ‘agent causation’. But her account is spoiled by the claim that physical actions are ‘tryings’ that cause bodily movements. It would be better to think of physical actions and bodily movements as identical; but Hornsby refuses to do this, seemingly because she thinks that to do so would be to endorse the so–called ‘standard causal story’. But Hornsby misses a possibility here, for we can insist on this identity claim without endorsing the standard story if we embrace an account which parallels the disjunctive account in the philosophy of perception. This will leave us with a picture of physical action that saves the insights of Hornsby's account without succumbing to its distortions.     

Non‐Relativist Contextualism about Free Will

Abstract: Contextualist accounts of free will recently proposed by Hawthorne and Rieber imply that the same action can be both free and unfree (depending on the attributor's context). This paradoxical consequence can be avoided by thinking of contexts not as constituted by arbitrary moves in a conversation, but rather by (relatively stable) social practices (such as the practices of attributing responsibility or of giving scientific explanations). The following two conditions are suggested as each necessary and jointly sufficient for free will: (i) the agent is able to form considered practical judgements and to act accordingly, and (ii) the agent (or some agent‐involving event) is the original cause of her actions. A contextualist reformulation of the second condition is developed according to which only contexts in which responsibility is attributed are relevant for the kind of original causation required for free will, which allows for a non‐relativist contextualism about free will.     

Libertarian Free Will and the Physical Indeterminism Luck Objection

Libertarian free will is, roughly, the view that agents (or, agent-involving events) cause actions to occur or not occur: Maddy’s decision to get a beer causes her to get up off her comfortable couch to get a beer, though she almost chose not to get up. Libertarian free will notoriously faces the luck objection, according to which agential states do not determine whether an action occurs or not, so it is beyond the control of the agent, hence lucky, whether an action occurs or not: Maddy’s reasons for getting beer in equipoise with her reasons to remain in her comfortable seat do not determine that she will get up or stay seated, so it seems beyond her control, hence lucky, that she gets up. In this paper I consider a sub-set of the luck objection called the Physical Indeterminism Luck Objection, according to which indeterministic physical processes cause actions to occur or not, and agent’s lack control over these indeterministic physical processes, so agent’s lack control over, hence it is lucky, whether action occurs or not. After motivating the physical indeterminism luck objection, I consider responses from three recent event-causal libertarian models, and conclude that they fail to overcome the problem, though one promising avenue is opened up.

This paper is divided into six parts. In Section One, I minimally define libertarian free will as accepting agential indeterminism, which is the conjunction of indeterminism and agential causation, where agential indeterminism occurs when an agent’s reasons, efforts or character indeterministically cause actions. In Sections Two and Three I outline the physical indeterminism luck objection to libertarian free will, which states that sub-agential physical processes in the brain indeterministically cause actions to occur, and agents lack control over these indeterministic physical causes, so agent’s lack control over whether their actions occur. If agent’s lack control over whether actions occur, the occurrence of these actions is lucky, where this luck jeopardizes free will and moral responsibility. In Sections Four through Six I consider three recent libertarian responses to this objection—Mark Balaguer in Section Four, Chris Franklin in Section Five, and Robert Kane in Section Six. I conclude that none of these models satisfactorily overcomes the physical indeterminism luck objection, though one interpretation of Kane yields a promising avenue of reply.

     

A problem for Wegner and colleagues’ model of the sense of agency

The sense of agency, that is the sense that one is the agent of one’s bodily actions, is one component of our self-consciousness. Recently, Wegner and colleagues have developed a model of the causal history of this sense. Their model takes it that the sense of agency is elicited for an action when one infers that one or other of one’s mental states caused that action. In their terms, the sense of agency is elicited by the inference to apparent mental state causation. Here, I argue that this model is inconsistent with data from developmental psychology that suggests children can identify the agent behind an action without being capable of understanding the relationship between their intentions and actions. Furthermore, I argue that this model is inconsistent with the preserved sense of agency in autism. In general, the problem is that there are cases where subjects can experience themselves as the agent behind their actions despite lacking the resources to make the inference to apparent mental state causation.     

Causal reasoning

The main focus of this paper is the question as to what it is for an individual to think of her environment in terms of a concept of causation, or causal concepts, in contrast to some more primitive ways in which an individual might pick out or register what are in fact causal phenomena. I show how versions of this question arise in the context of two strands of work on causation, represented by Elizabeth Anscombe and Christopher Hitchcock, respectively. I then describe a central type of reasoning that, I suggest, a subject has to be able to engage in, if we are to credit her with causal concepts. I also point out that this type of reasoning turns on the idea of a physical connection between cause and effect, as articulated in recent singularist approaches of causation.