Changfeng Wang

Learn More
We study tltt' problem of when to stop If'arning a class of feedforward networks-networks with linear outputs I1PUrOIl and fixed input weights-when they are trained with a gradient descent algorithm on a finite number of examples. Under general regularity conditions, it is shown that there in general three distinct phases in the generalization(More)
—Building on previous research, this study develops a research simulation model for the main factors affecting business innovation performance. It links the properties of knowledge obtained from outside network, firm's absorptive capacity and the features of a firm's knowledge network topology to predict innovation performance. We use VB.NET 2005(More)
Programs are demarcated as administrative structures established to realize planned organizational strategies through multi-project activities. Programs occupy a distinct locus in organizational hierarchy, so therefore necessitate specialized management approaches. Risks in programs tend to widen the gap between the organizational plans and the actual(More)
We consider the archetypal learning problem where a finite sample of examples generated by an underlying random process is made available to the learner who generates a hypothesis in a model class by gradient descent over the empirical loss function. In this context, we derive two criteria for machine size selection for a class of general nonlinear machines(More)
  • 1