This is just my personal notes where I put my research resources. This is basically my journey from a noob.
https://github.com/sxudan/artificial-intelligence-journey/blob/main/Classification/dogvscat.ipynb
A Generative Adversarial Network (GAN) is a type of artificial intelligence model used in unsupervised machine learning, particularly for generating new data samples from a given distribution.
The basic structure of a GAN involves two neural networks: a generator and a discriminator. Here's how it works:
Generator: This network takes random noise as input and tries to generate data samples that resemble the real data. For example, if you're training a GAN to generate images of cats, the generator network will take random noise vectors as input and output images that ideally look like real cat images.
Discriminator: This network is like a binary classifier. It takes both real data samples and generated data samples as input and tries to distinguish between them. It's trained to output a high probability if the input is real (i.e., from the true data distribution) and a low probability if the input is generated by the generator.
https://github.com/sxudan/artificial-intelligence-journey/tree/main/GAN
Fully Convolutional Networks (FCNs): These networks replace fully connected layers with convolutional layers to enable end-to-end pixel-wise prediction.
U-Net: U-Net is a popular architecture that consists of an encoder-decoder structure with skip connections. Skip connections help preserve spatial information during upsampling.
DeepLab: DeepLab is based on the atrous convolution (also known as dilated convolution) and employs techniques like atrous spatial pyramid pooling for capturing multi-scale context.
https://github.com/sxudan/artificial-intelligence-journey/tree/main/Segmentation
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent