Unit IV - Part 01
Unit IV - Part 01
• What is an auto-encoder?
Decoder
• Encoder
• Bottleneck
• Decoder
1. Input:
2. Encoder:
3. Decoder:
• Types of auto-encoders
1. Undercomplete auto-encoders
2. Sparse auto-encoders
3. Contractive auto-encoders
4. Denoising auto-encoders
2. Sparsity Constraint:
3. Loss Function:
1. Efficient Representation:
2. Feature Learning:
3. Robustness to Noise:
Types of auto-encoders
• 3. Contractive auto-encoders
• Mathematically:
𝛿ℎ/𝛿𝑥
input.
Types of auto-encoders
• Loss Function
where:
𝐿𝑟𝑒𝑐𝑜𝑛𝑠𝑡𝑟𝑢𝑐𝑡𝑖𝑜𝑛 = ∥𝑥 − 𝑥^∥2
• Robustness to Perturbations:
• Regularization:
Types of auto-encoders
• 4. Denoising autoencoders
• 5. Variational autoencoders
𝑧=𝜇+𝜎⋅𝜖
• Applications of VAEs
• 2. Image denoising
detection.