Thanks for reading Jonathan, I think there’re two answers to this. Firstly at an everyday investor level there’s definitely been an upsurge in interest in NLP/ML/AI applications in investing, it can be put down to a few things; the resources and knowledge required to implement have become ubitquitous, a general dissatisfaction with professional investment manager performance and the perception that applying these techniques to investing is novel and because they work so well for tasks like translation, question and answering and image classification they should do just as well in investing. It is a bit different at a professional investor level, while interests in this area has definitely increased since ten years ago, many still don’t know how this can be incorporated into an existing investment processes. As you would know investment managers are assessed based on their philosophy, robustness of investment process and discipline. As such, especially for a fundamental investor, it is quite hard to discard an existing “tried and proven” process in favor of adopting something more nascent such as AI/ML. Although, almost 7 decades old now, mean-variance optimization still remains the pre-eminent paradigm for portfolio construction(albeit with some improvements) and discounted cash flow is still the go to method for valuing and forecasting stocks. I have seen many studies that have produced positive results using NLP and news, and although I’m an advocate for ML/AI, I don’t think many of these results are reproducible, there’s just too much noise.

Researcher | Investor | Data Scientist | Curious Observer. Thoughts and insights from the confluence of investing and machine learning.

Love podcasts or audiobooks? Learn on the go with our new app.