We present a machine learning approach for the task of ranking previously answered questions in a question repository with respect to their relevance to a new, unanswered referenc...
Novelty detection involves identifying novel patterns. They are not usually available during training. Even if they are, the data quantity imbalance leads to a low classification ...
The method based on local features has an advantage that the important local motion feature is represented as bag-of-features, but lacks the location information. Additionally, in ...
In this article, we study the emergence of associations between words and concepts using the self-organizing map. In particular, we explore the meaning negotiations among communica...
A possible alternative to fine topology tuning for Neural Network (NN) optimization is to use Echo State Networks (ESNs), recurrent NNs built upon a large reservoir of sparsely r...