A new class of evolutionary computation processes is presented, called Learnable Evolution Model or LEM. In contrast to Darwinian-type evolution that relies on mutation, recombinat...
In many learning problems prior knowledge about pattern variations can be formalized and beneficially incorporated into the analysis system. The corresponding notion of invarianc...
Situated learning is known to be an effective didactic approach, yet, multimedia systems with built-in support for it are uncommon. We analyze the domain-independent requirements ...
This paper describes new machine learning approaches to predict the correct homepage in response to a user's homepage finding query. This involves two phases. In the first ph...
This paper describes how to automatically cross-reference documents with Wikipedia: the largest knowledge base ever known. It explains how machine learning can be used to identify...