AI · Theory

Types of Neural Networks

Theory

209/215

Types of Neural Networks

Feedforward MLP: tabular data. CNN: images (local filters). RNN/LSTM: sequences, mostly replaced by transformers for language. Transformers: attention, today’s NLP default. Autoencoders: compress. GANs / diffusion: generate images.

Match architecture to data shape. Don’t put a CNN on a 5-column marks table without a reason.

Output — idea / do / check for Types of Neural Networks. Fill those three in the viva.

Exam tip

CNN vs transformer: images vs language.

Example

# Types of Neural Networks
print("idea :", "Types of Neural Networks")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Types of Neural Networks. Fill those three in the viva.

Short notes

DefMLP / CNN / RNN / Transformer.
RuleMatch to data type.

AI · Theory

Types of Neural Networks

← All stacks

Theory

209/215

Types of Neural Networks

Match architecture to data shape. Don’t put a CNN on a 5-column marks table without a reason.

Output — idea / do / check for Types of Neural Networks. Fill those three in the viva.

Exam tip

CNN vs transformer: images vs language.

Example

# Types of Neural Networks
print("idea :", "Types of Neural Networks")
print("do   : one tiny example on paper")
print("check: one failure case")

Output: idea / do / check for Types of Neural Networks. Fill those three in the viva.

Short notes

DefMLP / CNN / RNN / Transformer.
RuleMatch to data type.

Types of Neural Networks

Example

Short notes

Questions

Types of Neural Networks

Example

Short notes

Questions