Fichas
Neural Computation
Fichas del conjunto (12)
- 1ANNs are worth studying because they can be extremely powerful ________.computational devices
- 2ANNs can be very efficient due to their massive ______.parallelism
- 3ANNs are both _____ tolerant and _____ tolerant.fault tolerant and noise tolerant
- 4ANNs are useful for both the scientific goal of __________, and the engineering goal of ________.modelling how real brains work, building efficient systems for real-world applications
- 5_______ is the most realistic representation of individual spike timings for ANN neurons, but to generate efficient learning algorithms, we use an approximation called ______.spike time coding
rate coding - 6The simplest ANNs consist of a set of _______, represented in a formula by indices _, _, and __. Activation flows between them via synapses with strengths ____, and ____.McCulloch-Pitts neurons
k, i, and j
w(ki), and w(ij) - 7The first step of building an ANN is to specify the problem in terms of _____, and required ______.inputs, outputs
- 8Before finding appropriate connection weights for an ANN, it is best to choose the _______ form of network that is able to solve the problem.simplest
- 9The arrangement of one input layer of activations feeding forward to one output layer of McCulloch-Pitts neurons is known as a simple ________.Perceptron
- 10The Perceptron Learning rule is an algorithm that adjusts network ______ to minimize the difference between _____ outputs and _______outputs.weights, actual, target
- 11We quantify the difference between target and actual outputs by defining an _______ function: Cross Entropy for classification, or Sum Squared Error for regression.error function
- 12—
Información
- Descripción
- Fichas sobre Neural Computation, creado por Brenna Cantwell el 18/08/2016.
- Creado por
- Brenna Cantwell
- Visitas
- 1
- Valoración