Induction Confirmation Paradoxes and Reinforcement Propagation
نویسنده
چکیده
In this paper many of the classical paradoxes of induction confirmation are briefly discussed. Most of them are originated by qualitative views of confirmation. Some proposals to solve them are recalled, especially those which are based on the notion of converse coverage or entailment, jointly with a selection criterion. However, these solutions introduce new problems in the context of confirmation. The constructive reinforcement propagation method introduced in [4][7] is proposed as a solution to the problem of confirmation and analysed in front of all of these paradoxes. It is shown that this new measure is robust to all of them.
منابع مشابه
Conditions for the validity of Faraday’s law of induction and their experimental confirmation
Abstract This paper, as its main didactic objective, shows the conditions needed for the validity of Faraday’s law of induction. Inadequate comprehension of these conditions has given rise to several paradoxes about the issue; some are analysed and solved in this paper in the light of the theoretical deduction of the induction law. Furthermore, an experimental set-up, in which such conditions a...
متن کاملOn the equivalence of Goodman's and Hempel's paradoxes.
Historically, Nelson Goodman's paradox involving the predicates 'grue' and 'bleen' has been taken to furnish a serious blow to Carl Hempel's theory of confirmation in particular and to purely formal theories of confirmation in general. In this paper, I argue that Goodman's paradox is no more serious of a threat to Hempel's theory of confirmation than is Hempel's own paradox of the ravens. I pro...
متن کاملA Synthesis of Hempelian and Hypothetico-Deductive Confirmation
This paper synthesizes confirmation by instances and confirmation by successful predictions, and thereby the Hempelian and the hypotheticodeductive traditions in confirmation theory. Our merger of these two approaches solves classical problems such as the tacking paradoxes and the raven paradox and is subsequently extended to the piecemeal confirmation of entire theories. We compare our proposa...
متن کاملHypothetico-Deductive Confirmation
Hypothetico-deductive (H-D) confirmation builds on the idea that confirming evidence consists of successful predictions that deductively follow from the hypothesis under test. This article reviews scope, history and recent development of the venerable H-D account: First, we motivate the approach and clarify its relationship to Bayesian confirmation theory. Second, we explain and discuss the tac...
متن کاملDynamic Obstacle Avoidance by Distributed Algorithm based on Reinforcement Learning (RESEARCH NOTE)
In this paper we focus on the application of reinforcement learning to obstacle avoidance in dynamic Environments in wireless sensor networks. A distributed algorithm based on reinforcement learning is developed for sensor networks to guide mobile robot through the dynamic obstacles. The sensor network models the danger of the area under coverage as obstacles, and has the property of adoption o...
متن کامل