In the simple model (the Rescorla-Wagner rule) of the classical conditioning introduced in the part 26, what is going to happen to the association w when the reward is delivered with a probability 0
A. w reaches a constant value equal to 1-p.
B. w drops to zero
C. w fluctuates around p.
D. w approaches one.
查看答案
Which of the following phenomena can be modeled by unsupervised learning? 84. (part 25, easy) 105-Final
A. The integrator neurons in the oculomotor system.
B. Associative memory.
Classical conditioning.
D. Ocular dominance.
In the unsupervised learning described in the part 25, the synaptic weight change is determined by 82. (part 25, difficult) 105-Final
A. the firing rate of the postsynaptic neuron
B. the synaptic normalization
C. the correlation between inputs to different synapses.
D. the external supervised signal that carries the correct responses.
To address the problem of the basic Hebb rule in the firing rate model, several solutions have been proposed. Which of the following concepts is not one of them? 79. (part 24, medium) 105-Final
A. synaptic depression
B. sliding threshold
C. synaptic normalization
D. spike-timing dependent plasticity.
Given the simple firing rate model, τdv/dt=-v+w · u, in which v is the rate of the postsynaptic neuron, u and w are vectors representing the rates of the presynaptic neurons and the corresponding synaptic weights, respectively, which of the following equations describes the basic Hebb learning rule? *77. (part 24, easy) 105-Final
A. τwdw/dt=vu
B. w=vu
C. dw/dt=u
D. u=vw