# Truthworthy ML 下學期規劃

## Paper seminar
- 讓組員分兩組，每兩周報告 paper
- 目的：讓大家能夠鑽研整理 paper 報告技巧、實現 paper code 並應用

**電腦視覺**
- [CVPR paper with code(ref)](https://github.com/amusi/CVPR2024-Papers-with-Code/tree/master)
- 3d gaussian splatting
- TensoRF Tensorial Radiance Fields.pdf
- hexplane
- NeRF
- NeRV
- VAE
- SEAN: Image Synthesis with Semantic Region-Adaptive Normalization(GAN)
- Attention Is All You Need(transformer)
- Vision Transformer with Super Token Sampling
- [code](https://github.com/hhb072/STViT)
- Simple diffusion: End-to-end diffusion for high resolution images
- [code](https://github.com/patil-suraj/simple-diffusion)
**影像/影片/點雲壓縮**
- PCGCv1
- PCGCv2
- Sparse Tensor-based Multiscale Representation for Point Cloud Geometry Compression(pcgcv3, sparse pcgc)
- [code](https://github.com/NJUVISION/SparsePCGC)
- Variational image compression with a scale hyperprior
- (PCC-S)PointCloudCompressionwithSiblingContext andSurfacePriors
- Efficient LiDAR Point Cloud Geometry Compression
Through Neighborhood Point Attention
- Multiscale Latent-Guided Entropy Model for LiDAR Point Cloud Compression
- YOGA
- [kd tree](https://zhuanlan.zhihu.com/p/112246942)
- [oct tree](https://blog.csdn.net/xhtchina/article/details/129913312)
- Image Compression Using Augmented Normalizing Flows

-------
### Refined Proposal of GDSC

#### Overview
**Total 14 weeks:**
* Spending 5 - 7 weeks to prepare for technical background
* Video Resources
* Weekly QA
* Paper Reading and Implementation
* Weekly Meeting on Papers
* Paper/Topic Decision before last three weeks
* Groups Topic (Plausible):
* Trustworthy ML
* Natural Language Processing
* Computer Vision

#### Paper Lists:
* [Attention is all-you-need](https://proceedings.neurips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf)
* [U-Net](https://lmb.informatik.uni-freiburg.de/people/ronneber/u-net/)
* ResNet
* Diffusion Model
* GAN