Based on your view, what is the most challenging research problem in data mining? If you were given a number of years and a good number of researchers and implementors, what would your plan be to make good progress toward an effective solution to such a problem?