# PaAC Open Projects 2023 | Deep Learning

## Meet - 01
## Literature Review
- [CNN](https://d2l.ai/chapter_convolutional-neural-networks/index.html) - Dive into Deep Learning (Must Understand)
- [Complex CNN models](https://d2l.ai/chapter_convolutional-modern/index.html) - For exploring ResNets, AlexNet, DenseNet. (Optional)
- [Intro to CNN](https://arxiv.org/abs/1511.08458) - Research Paper (Must Understand)

### Videos
- [CS231n](https://youtube.com/playlist?list=PLC1qU-LWwrF64f4QKQT-Vg5Wr4qEE1Zxk&si=C3feC9JGfwYV8xB5) - Lecture Number 2, 3, 5, 6, 7. (Must)
- [Statquest](https://youtube.com/playlist?list=PLblh5JKOoLUIxGDQs4LFFD--41Vzf-ME1&si=xc2OPAn-heQNFlpf) - Good videos on DL. Fun watch.

### Books
- ["Deep Learning" by Ian Goodfellow, Yoshua Bengio, and Aaron Courville](http://www.deeplearningbook.org/)
- A comprehensive book covering various deep learning topics, including CNNs.

### Task
- Achieve high accuracy on CNN for classification of CIFAR10 Images. Try different approaches, play with the model, modify and change the structures, understand hyperparameter tuning etc etc.

## Meet - 02
### Video Lectures.
- [Arxiv Insights (VAE)](https://youtu.be/9zKuYvjFFS8?si=ozon8wzV3L3cDLxR) - Informatics
- [Basic Intro with python implementation](https://youtube.com/playlist?list=PLZsOBAyNTZwb-uK_a6ywrU3t0hy80G5QP&si=xGccKDqETLBTJycL)
- [Building VAE with Pytorch](https://youtu.be/IQpP_cH8rrA?si=_3-sGZQ89BDeojtF)
- [Mathematical Intro to VAE](https://youtu.be/iL1c1KmYPM0?si=QugwwDpYgW0_Rikk)
- [Variational inference, VAE's and normalizing flows](https://youtu.be/-hcxTS5AXW0?si=DhrUGV8Q8S20hgI4)

### Literature
- [VAE from towards data-science](https://towardsdatascience.com/understanding-variational-autoencoders-vaes-f70510919f73)
- [Blog for AE](https://www.jeremyjordan.me/autoencoders/)

### Task
Implement from scratch the VAE and AE, apply it on CIFAR-10 and Galaxy-10 datasets, and compare the results, understand the mechanism behind the differences. Get minimum reconstruction loss.