Reading List
A curated list of foundational and frontier papers in Asset Pricing and Financial Machine Learning.
1. Asset Pricing Baseline (The Foundation)
Traditional financial models that serve as benchmarks for modern AI strategies.
| Paper | Key Relevance |
|---|---|
| Fama, E. F., & French, K. R. (2015) A five-factor asset pricing model Journal of Financial Economics | Baseline Comparison: Used to benchmark the Transformer model and demonstrate the superiority of non-linear relationships over linear factor models. |
| Carhart, M. M. (1997) On persistence in mutual fund performance The Journal of Finance | Momentum Factor: Introduces the “Momentum” factor, a key feature that is effectively captured by deep learning models. |
2. ML in Finance (The Bridge)
Pioneering works that introduced machine learning into empirical asset pricing.
| Paper | Key Relevance |
|---|---|
| Gu, S., Kelly, B., & Xiu, D. (2020) Empirical asset pricing via machine learning The Review of Financial Studies | The Bible: A must-read. Provides empirical evidence that Neural Networks significantly outperform OLS and LASSO in processing high-dimensional factors. |
| López de Prado, M. (2018) Advances in Financial Machine Learning Wiley | Industry Standard: Essential for rigorous backtesting. Focus on “Purged K-Fold Cross Validation” (Ch.7) and “Fractional Differentiation” (Ch.5) to eliminate look-ahead bias. |
3. The Frontier (Transformers & Interpretability)
Advanced deep learning architectures for time-series forecasting and event-driven analysis.
| Paper | Key Relevance |
|---|---|
| Vaswani, A., et al. (2017) Attention is all you need NIPS | Model Architecture: The origin of the Transformer model. cited to introduce the self-attention mechanism. |
| Lim, B., et al. (2021) Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting International Journal of Forecasting | TFT Model: Key reference for handling “multi-horizon” forecasting and “variable selection” in complex time series data. |
| Ding, Y., et al. (2020) Deep learning for event-driven stock prediction IJCAI | Event-Driven: A primary reference for incorporating macro events and news sentiment analysis into stock prediction models. |