limited reasoning
Recently Published Documents


TOTAL DOCUMENTS

8
(FIVE YEARS 0)

H-INDEX

4
(FIVE YEARS 0)

Author(s):  
Abhishek Sharma ◽  
Keith M. Goolsbey

Cognitive systems must reason with large bodies of general knowledge to perform complex tasks in the real world. However, due to the intractability of reasoning in large, expressive knowledge bases (KBs), many AI systems have limited reasoning capabilities. Successful cognitive systems have used a variety of machine learning and axiom selection methods to improve inference. In this paper, we describe a search heuristic that uses a Monte-Carlo simulation technique to choose inference steps. We test the efficacy of this approach on a very large and expressive KB, Cyc. Experimental results on hundreds of queries show that this method is highly effective in reducing inference time and improving question-answering (Q/A) performance.


2018 ◽  
pp. 55-72
Author(s):  
Kyla Ebels-Duggan

In this chapter, Kyla Ebels-Duggan considers how Christian philosophers should decide which questions are worth asking. She provides an interpretation and defense of Alvin Plantinga’s claim that Christian philosophers should strive for autonomy, and then argues that this rules out some ways of settling on our questions. Ebels-Duggan then suggests that the questions in which Christian philosophers should take an interest are those arising from or continuous with a distinctively Christian way of life. Along the way she argues that the power of the distinctive tools of philosophy is importantly limited: reasoning alone cannot settle either which questions we should ask or which commitments we should take on.


2018 ◽  
Vol 19 (1) ◽  
Author(s):  
Nick Vikander

AbstractThis paper examines how a firm can strategically use sellouts to influence consumers’ beliefs about its product’s popularity. A monopolist faces a market of conformist consumers, whose willingness to pay is increasing in their beliefs about aggregate demand. Consumers are broadly rational but have limited strategic reasoning about the firm’s incentives. Formally, I apply the concept of a ‘cursed equilibrium’, where consumers neglect how the firm’s chosen actions might be correlated with its private information about demand. I show that in a dynamic setting, the firm may choose its price and capacity so as to generate sellouts, specifically to exploit consumers’ limited reasoning. It does so to effectively conceal unfavorable information from consumers about past demand in a way that increases future profits. Sellouts tend to occur when demand is low, rather than high, and may be accompanied by introductory pricing. The analysis also demonstrates that the firm’s ability to mislead some consumers always benefits certain others, and can result in higher overall consumer surplus.


10.29007/ndjg ◽  
2018 ◽  
Author(s):  
Giles Reger ◽  
Martin Suda

This paper describes initial experiments using the set of support strategy to improve how a saturation-based theorem prover performs theory reasoning with explicit theory axioms. When dealing with theories such as arithmetic, modern automated theorem provers often resort to adding explicit theory axioms, for example, x+y = y+x. Reasoning with such axioms can be explosive. However, little has been done to explore methods that mitigate the negative impact of theory axioms on saturation-based reasoning. The set of support strategy requires that all inferences involve a premise with an ancestor in a so-called set of support,initially taken to be a subset of the input clauses, usually those corresponding to the goal. This leads to completely goal orientated reasoning but is incomplete for practical reasoning (e.g. in the presence of ordering constraints). The idea of this paper is to apply the set of support strategy to theory axioms only, and then to explore the effect of allowing some limited reasoning within this set. The suggested approach is implemented and evaluated within the VAMPIRE theorem prover.


1987 ◽  
Vol 34 (1) ◽  
pp. 39-76 ◽  
Author(s):  
Ronald Fagin ◽  
Joseph Y. Halpern
Keyword(s):  

Sign in / Sign up

Export Citation Format

Share Document