In the present paper we propose a consistent way to integrate syntactical least general generalizations (lgg's) with semantic evaluation of the hypotheses. For this purpose we...
We explore the striking mathematical connections that exist between market scoring rules, cost function based prediction markets, and no-regret learning. We first show that any c...
There has been a recent, growing interest in classification and link prediction in structured domains. Methods such as conditional random fields and relational Markov networks sup...
In this paper, we propose a scheme to improve the performance of subspace learning by using a pattern(data) selection method as preprocessing. Generally, a training set for subspa...
Jin Hee Na, Seok Min Yun, Minsoo Kim, Jin Young Ch...
Different formal learning models address different aspects of learning. Below we compare learning via queries—interpreting learning as a one-shot process in which the learner i...