algotradingsoc/generative_modelling

Research repo for the generative modelling team.

Generative Modelling:

Research repo for the generative modelling team.

Litature Review:

1. GP-VAE: https://arxiv.org/pdf/1907.04155.pdf. Code: https://github.com/ratschlab/GP-VAE
2. VAE: https://arxiv.org/pdf/1312.6114.pdf
3. Style Transfer: https://arxiv.org/pdf/1906.03232.pdf
4. Hierarchical PCA: https://arxiv.org/pdf/1910.02310.pdf
5. VAE-GAN: https://www.mdpi.com/1424-8220/20/13/3738/htm
6. RAE: https://arxiv.org/abs/1707.07961

---

Research Proposal:

1. Initial review of time series generative modelling implementing some key baselines for our future work to be compared against.
2. Explore viable alternative models.
3. Explore the impact of these models on financial data.
4. Explore splitting the data into predictable sub parts, wavelet and gaussian process are typical methods.

Applications:

1. Data generator for improving supervised or reinforcement learning.
2. Use data generator as a simulator for the environment to get a distribution over performance.