Reminding five algorithms using deep learning teqnique
Overview of the Restricted Boltzmann Machine , RBM
1. Model proposed by Paul Smolensky in 1986
RBM is initially proposed under the name Harmonium by Paul Smolensky in 1986, and rose to prominence after Geoffrey Hinton
2. RBM is utilized in unsupervised learning in machine learning and it is useful in recommending video( like Netflix ) and recommendation algorithms.
3. Advancements in Boltzmann machine
4. There are visible layer( meaning I/O layer ) and hidden layer and nodes between each layer connected by 1:n and no connection at the inner layers
5. It calculates connection strength using by Back propagaition ( Gradient descent )
Overview of the Deep Belief Network, DBN
1. Announcement by professor Geoffrey Hinton in 2006
2. Made by stacking multiple layers of RBM
3. RBM highlited againg through DBN
4. Solved problem of vanishing gradient
- DBN optimizes RBM step by step.
- DBN introduced Relu instead Sigmoid which previously utilized
'공학 > 인공지능' 카테고리의 다른 글
| Generative Adversarial Network , GAN (0) | 2024.12.04 |
|---|---|
| Artifitial Neural Network 01 (0) | 2024.12.04 |
| Recurrent Neural Network, RNN with LSTM, GRU (0) | 2024.11.28 |
| Convolution Neural Network, CNN (0) | 2024.11.27 |
| Deep learning & Deep Neural Network, DNN (1) | 2024.11.25 |