The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
We show for many simple two-layer networks whose nodes compute linear threshold functions of their inputs that training is NP-complete. For any training algorithm for one of these networks there will be some sets of training data on which it performs poorly, either by running for more than an amount of time polynomial in the input length, or by producing sub-optimal weights. Thus, these networks differ...
In this paper we presented FAILSAFE-II, a system that performs adaptive search by learning on-line form its failures. The key contribution of this system is its use of the preservability of failures. Preservability assumption allows FAILSAFE-II to over-generalize the failures and discard some solutions to the problem along with the non-solutions. This leads to learning of search control rules which...
We have described a running system that embodies efficient parallel implementations of several symbolic machine learning induction operators. It functions as an “Induction Assistant” to a domain expert. First we developed an efficient, noise-tolerant, similarity-based parallel matching algorithm. This should apply to other graph-based representations of domains possessing an embedding in which the...
The goal of this research was to evaluate the potential of connectionist networks for speech recognition. The research demonstrated that solutions to a representative set of phoneme discrimination problems could be obtained for a single male speaker. There are many unanswered questions about how these network models and methods might be extended to recognition of a complete set of phonemes, spoken...
We formalize a behavior-based learning architecture for an autonomous agent. The IR oven tuning problem is introduced and is investigated as a real industrial application of this architecture. The algorithm we developed was shown to be very robust and was tested through simulation of different intelligent machines, including Genghis [9]. The distinguishing feature of this learning algorithm is that...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.