Summary of cognitive computing knowledge points

One: What is cognitive computing？

Cognitive computing is one of the core technical sub fields of cognitive science. It is an important part of artificial intelligence and a computer system that simulates the cognitive process of human brain.

Two: Learning
1 Bayesian Inference
classical statistics: P(A)=f/N
Bayesian statistics
Bayesian statistics proposed the concept prior information and uses both sample information and prior information to predict population.
Probability is a way to represent an individuals degree of belief in a statement, or given evidence.
Bayesian statistics is based on the Bayesian theorem and Bayesian inference.
Choosing hypothesis, generally want the most probable
hypothesis given the training data.
In candidate hypothesis set H, searching the most likely hypothesis h for a given data D.
The MAP is determined by calculating the posterior probability of each candidate hypothesis using the Bayesian formula.

THREE: Supervised learning

Supervised learning is a method of machine learning. It studies and constructs algorithms that can learn from data and make predictions. This algorithm operates by building models from example inputs for data-driven prediction or decision-making, rather than strictly following static program instructions.
In supervised learning, each example consists of an input object and an expected output value. Supervised learning algorithm analyzes training data and generates inference function, which can be used to map new examples.

Four: Artificial Neural Network
1 Activation Function
(1) step function
(2) linear function: f(x)=kx+b
(3) sigmoid function

(4) tanh function:

2 An ANN can:

1. compute any computable function, by the appropriate selection of the network topology and weights values.
2. learn from experience. Specifically, by trial‐and‐error.

Learning by trial and error.
Continuous process of:
Trial: Processing an input to produce an output (In terms of ANN: Compute the output function of a given input).
Evaluate: Evaluating this output by comparing the actual output with the expected output.
Adjust: Adjust the weights.

3 BP
Relies on the sigmoid activation function for communication.