| Figure 20.1 - Undercomplete encoder-decoder architecture.png | 114.7 KB | |
| Figure 20.10 - Cumulative returns of quintile-based and long-short portfolios.png | 81.1 KB | |
| Figure 20.2 - Fashion MNIST sample images.png | 131.5 KB | |
| Figure 20.3 - Sample Fashion MNIST images, original and reconstructed.png | 194.3 KB | |
| Figure 20.4 - t-SNE visualization of the Fashion MNIST autoencoder embedding.png | 1.6 MB | |
| Figure 20.5 - Denoising input and output examples.png | 339.9 KB | |
| Figure 20.6 - Number of tickers over time and per - stock characteristic.png | 22.7 KB | |
| Figure 20.7 - Conditional autoencoder architecture designed by GKX.png | 74.4 KB | |
| Figure 20.8 - Cross-validation performance for all factor and hidden-layer size combinations.png | 497.7 KB | |
| Figure 20.9 - Mean period-wise spread by prediction quintile.png | 15.8 KB | |
