Appears in Proceedings of The 11 th International Conference on Machine Learning ( ML - 94 )
نویسنده
چکیده
This paper compares two methods for reen-ing uncertain knowledge bases using propo-sitional certainty-factor rules. The rst method, implemented in the Rapture system , employs neural-network training to re-ne the certainties of existing rules but uses a symbolic technique to add new rules. The second method, based on the one used in the Kbann system, initially adds a complete set of potential new rules with very low certainty and allows neural-network training to lter and adjust these rules. Experimental results indicate that the former method results in signiicantly faster training and produces much simpler reened rule bases with slightly greater accuracy.
منابع مشابه
Proceedings of the 6th International Conference on Science and Social Research (CSSR)(Malaysia)
متن کامل
A Comparative Study of English-Persian Translation of Neural Google Translation
Many studies abroad have focused on neural machine translation and almost all concluded that this method was much closer to humanistic translation than machine translation. Therefore, this paper aimed at investigating whether neural machine translation was more acceptable in English-Persian translation in comparison with machine translation. Hence, two types of text were chosen to be translated...
متن کاملProceedings of the First MEFOMP International Conference of Medical Physics: November 2-4, 2011, Shiraz, Iran
متن کامل
In the Proceedings of the 15 th International Conference on Machine Learning ( ICML - 98 ) Well - Behaved
How can we guarantee that our software and robotic agents will behave as we require, even after learning? Formal veri cation should play a key role but can be computationally expensive, particularly if re-veri cation follows each instance of learning. This is especially a problem if the agents need to make rapid decisions and learn quickly while online. Therefore, this paper presents novel meth...
متن کاملIn Proceedings of ICML ' 96 : The 13 th International Conference on Machine Learning
Developing the ability to recognize a landmark from a visual image of a robot's current location is a fundamental problem in robotics. We describe a way in which the landmark matching problem can be mapped to that of learning a one-dimensional geometric pattern. We present an eecient noise-tolerant algorithm (designed using the statistical query model) to PAC-learn the class of one-dimensional ...
متن کامل