Self-improving AI: an Analysis

Self-improvement was one of the aspects of AI proposed for study in the 1956 Dartmouth conference. Turing proposed a “child machine” which could be taught in the human manner to attain adult human-level intelligence. In latter days, the contention that an AI system could be built to learn and improve itself indefinitely has acquired the label of the bootstrap fallacy. Attempts in AI to implement such a system have met with consistent failure for half a century. Technological optimists, however, have maintained that a such system is possible, producing, if implemented, a feedback loop that would lead to a rapid exponential increase in intelligence. We examine the arguments for both positions and draw some conclusions.

Investigations into a left-structural right-substructural sequent calculus

We study a multiple-succedent sequent calculus with both of the structural rules Left Weakening and Left Contraction but neither of their counterparts on the right, for possible application to the treatment of multiplicative disjunction (fission, ‘cotensor’, par) against the background of intuitionistic logic. We find that, as Hirokawa dramatically showed in a 1996 paper with respect to the rules for implication, the rules for this connective render derivable some new structural rules, even though, unlike the rules for implication, these rules are what we call ipsilateral: applying such a rule does not make any (sub)formula change sides—from the left to the right of the sequent separator or vice versa. Some possibilities for a semantic characterization of the resulting logic are also explored. The paper concludes with three open questions.     

压密注浆在顶管脱节事故中的应用

顶管施工技术在市政管道工程中应用日益增多。因其自身施工特点,顶管施工中会遇到各种施工事故。文章就宁波某工业园区一条主干路中污水管顶管施工中遇到的顶管脱节事故进行研究,详细分析了导致顶管脱节的事故原因,并采用了压密注浆技术方案,对该顶管脱节事故进行处理,取得了良好的工程实际效果。同时,总结了压密注浆施工技术。在顶管脱节事故中的应用要点,为在以后的工程实际应用中提供一定的借鉴。     

改进的本体语义相似度计算方法

概念的语义相似度研究,是知识表示以及信息检索领域中的一个重要内容。通过分析两种传统的语义相似度计算方法,对它们存在的问题进行改进,提出了一种综合的基于本体的概念语义相似度计算方法。该方法结合本体的DAG网状结构特征和语义距离计算中的多种语义影响因素,充分利用本体中概念的语义来计算概念间的语义相似度。实验结果比较合理,验证了该方法的有效性。     

异构空间数据系统查询分解算法的研究

介绍了利用析取范式进行查询分解的算法,该分解方法在语义分析和规范化的基础上,对查询语句进行重组,可以大幅度地减少连接运算的时间,提高查询效率。     

Graded logic for decision support systems

We present the soft computing graded logic (GL) as a mathematical infrastructure for a soft computing propositional calculus, and GL aggregation. The GL aggregation of degrees of truth (or fuzzy membership) is a crucial component of complex criterion functions used in decision support systems. In the context of GL we propose aggregation functions that integrate means and t-norms/conorms, combining logic and probabilistic reasoning. We also propose a set of necessary and sufficient basic logic functions (models of simultaneity, substitutability, and complementing of degrees of truth) and their analytic forms. GL is developed as a seamless soft computing generalization of classical Boolean logic. The proposed generalizations include: (1) the continuous parameterized transition from drastic conjunction to drastic disjunction, integrating the regions of hyperconjunction, hard and soft partial conjunction, logic neutrality, soft and hard partial disjunction, and hyperdisjunction and (2) semantic generalization where we assign a degree of importance to each degree of truth, in a way that is consistent with observable properties of human intuitive reasoning. These generalizations reflect the fact that degrees of truth are not anonymous real numbers, but values that have semantic identity derived from goals and requirements of decision-maker. The basic goal of GL is to provide humancentric mathematical models of logic aggregators that can be used in the decision process of comparison and selection of alternatives based on the evaluation of each alternative. This paper is structured to present the characteristic properties of entire GL in a way that is both concise and complete.     

“X” means X: Semantics Fodor-style

InPsychosemantics Jerry Fodor offered a list of sufficient conditions for a symbol X to mean something X. The conditions are designed to reduce meaning to purely non-intentional natural relations. They are also designed to solve what Fodor has dubbed the disjunction problem. More recently, inA Theory of Content and Other Essays, Fodor has modified his list of sufficient conditions for naturalized meaning in light of objections to his earlier list. We look at his new set of conditions and give his motivation for them-tracing them to problems in the literature. Then we argue that Fodor's conditions still do not work. They are open to objections of two different varieties: they are too strong and too weak. We develop these objections and indicate why Fodor's new, improved list of conditions still do not work to naturalize meaning.We wish to thank Ray Elugardo, Pat Manfredi, and Donna Summerfield for comments on a draft of this paper, as well as other members of the CMU Mind Group (David Drebushenko, Gary Fuller, and Naomi Reshotko) for helpful discussion and support. A version of this paper was presented at the Southern Society for Philosophy and Psychology in Atlanta, March, 1991, and at the Canadian Society for History and Philosophy of Science in Kingston, Ontario, in May, 1991.     

一种基于布尔代数的秘密共享方案

秘密共享方案是在n个参与者之间共享秘密k的方法。将布尔代数中的与或逻辑引入秘密共享而提出的新方案运算速度快,并具有灵活的自适应能力和良好的可扩展性,便于软件编程和硬件固化,可以与经典的加密方法紧密结合,以提高其安全性。     

Age-trend-related differences in tasks involving conjunctive probabilistic reasoning.

Conjunctive probabilistic reasoning has been studied at different ages to ascertain whether the conjunction fallacy is due to a task demand misinterpretation. Such a misinterpretation might occur because a task that requires a comparison between a superordinate class A and a subordinate class A&B is mistakenly interpreted as requiring a comparison between the two complementary subordinate classes of A (i.e., A&B and A¬B). Children (7- and 10-year-olds) and adults were required to make conjunctive probability judgments about problems for which explicit objective probabilities were provided. The total number of A items was kept constant and the frequencies of the A&B and of the A¬B items varied across problems. When the number of A&B items was smaller than the number of A¬B items, the frequency of congruent responses increased with age. When the number of A&B items was greater or equal to that of the A¬B items, the frequency of correct answers decreased. (PsycINFO Database Record (c) 2010 APA, all rights reserved)