A Knowledge-Based Framework for Effective Probabilistic Control Strategies in Signal Understanding
| Autor: | Claudio Rullent, Roberto Gemello, Egidio P. Giachin |
|---|---|
| Rok vydání: | 1987 |
| Předmět: |
Class (computer programming)
Computer science business.industry SIGNAL (programming language) Probabilistic logic computer.software_genre Machine learning Blackboard (design pattern) Blackboard system Data mining Artificial intelligence Control (linguistics) business Set (psychology) computer Control methods |
| Zdroj: | GWAI-87 11th German Workshop on Artifical Intelligence ISBN: 9783540183884 GWAI |
| DOI: | 10.1007/978-3-642-73005-4_11 |
| Popis: | We describe a problem solving framework for a knowledge based approach to signal understanding. The risk of erroneous analysis makes advisable the use of well-experimented probabilistic control methods. Such methods have been used in the past in task-specific applications, such as speech. Here they are generalized to the case of a deduction system, thus becoming applicable to a wider class of problems still maintaining their effectiveness. That results in a blackboard based framework where every knowledge source is abstracted as a set of operators, that allow integration of different deductive processes independently evolved, either forward or backward. Such framework allows the use of admissible control strategies proposed in the literature. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|