Fill in the blanks

1. We usually define the reinforcement learning problem as a ________________________ process (MDP).

2. When using the minimum description length (MDL) criterion to solve the problem of decision tree learning, we should consider two code lengths, which are _________ and ________. 3. Feedforward Neural Network is actually a representation form of __________.

4. The learning criterion in BP neural network learning algorithm is ________ , and the optimization algorithm is ____________ .

5. The operating principal of Hopfield neural network is to find___________ state through minimizing the energy function.

6. We can improve the ___________ of clustering algorithms through data sampling.