ISM- Period 6
Ding, Shifei & Li, Hui & Su, Chunyang & Yu, Junzhao & Jin, Fengxiang. (2013). Evolutionary
10.1007/s10462-011-9270-6.
Evolutionary algorithms use biological phenotypes and genotypes to achieve the highest fitness
score in a given task.
Jeffrey Mo
ISM-Period 6
Bhattacharya A., Abraham A., Grosan C., Vasant P., Han S. (2006) Meta-Learning Evolutionary
Artificial Neural Network for Selecting Flexible Manufacturing Systems. In: Wang J., Yi
Z., Zurada J.M., Lu BL., Yin H. (eds) Advances in Neural Networks - ISNN 2006. ISNN
2006. Lecture Notes in Computer Science, vol 3973. Springer, Berlin, Heidelberg
o
o Training set data is equal to
▪ In the set of i index of arguments and target
▪ Such that if the index is equal to 1
• local search techniques are used to fine-tune the weights (meta-learning)
• Connection and activation functions
Hardesty, Larry, and MIT News Office. “Explained: Neural Networks.” MIT News, MIT News,
• Google’s translator:
o “deep learning”
o “Perform some task by analyzing training examples”
• Neural networks:
o Training examples are usually hand-labeled
o Object recognition system
o Thousands of images:
▪ Cars
▪ Houses
▪ Horses
▪ Coffee cups
• Modelled loosely on the human brain
• Thousands of nodes
o Densely interconnected
• “Feed forward network”
• Receive interconnections known as weights
o Each weight receives different data
o Threshold value depends on how much a node fires
• Training data is fed through input layer
• Exited through output layer
• “human neuroanatomy and cognition”
o An indication that they capture something about how the brain processes
information
• Poggio’s three-part theorical study of neural network
Artificial neural networks allow scientists to model the brain’s network structure to better
understand and solve complex problems.