If an RNN hidden layer has 4 processing neurons, how many weights (whether hidden state output or bias weights) exist to feedforward the hidden state at time t (current timestamp) to the hidden state at time t+1 (next time stamp)?

Three ECG probes are placed on 100 patients to monitor their heart pulse. Pulses from these probes are used to automatically diagnose the patients as having arrythmia or not. Each pulse is recorded for a duration of 1 minute at a sampling rate of twice per second. A many-to-one RNN is trained on the pulses to make arrythmia diagnoses and it achieves an accuracy of 95 percent. How many sensing neurons did this RNN have?

A CNN accepts a 10x10x3 image as input. How many weights (whether input or bias weights) are used in the CNN's first convolutional layer which has 1 filter with height and width both equal to 3 pixels?

What is the maximum number of unique gray levels that can be in a 100x100 grayscale image?