1 - The Issues Surrounding Artificial Intelligence in Alternative Investment Portfolio Management
Update: 2022-12-21
Description
Angelo Calvello discusses issues surrounding AI (Artificial Intelligence) in Alternative Investment portfolios.
Transcript:
| My name is Angelo Calvello with my partner, Julie Bonifede. I co-founded Rosetta Analytics in 2016. Rosetta We're an asset manager and an asset manager that's been using advanced. I mean, we've been using deep learning since 2017 and deep reinforcement learning since 2020 to run money and life strategies for U.S. institutional investors. And by any measure, the models work, and the strategies provide investors with the investment outcomes they seek. |
| Yet we face a strong headwind in our business development because of what I can only describe as a cultural belief. And I'll get to that in a minute. Also, for a fuller delineation of the following ideas, I encourage you to read my essay and institutional investor entitled The Most Powerful Artificial Intelligence Knows Nothing About Investing, and that's perfectly okay. |
| It was published in February of 2021. Let me start with my assessment of the investment industry's use of AI or what I prefer to call machine learning or its use of an L. Today's become quite fashionable for managers to claim to be using machine learning as part of their investment process if they're truthful. And let's be clear, there's a lot of hand-waving in this space, but if they're truthful, then they have likely integrated what I would call traditional machine learning techniques, such as support vector machines, random forests, nearest neighbor into their existing investment processes. |
| These traditional machine learning techniques are generally used to augment an existing human based investment process, reducing AI at the end of the day to a handmaiden to human intelligence. And while such machine learning techniques have proven helpful, they remain bound by the constraints of human intelligence, and they fail to achieve the superhuman performance that comes with what many call a new wave of AI systems, this new wave. |
| These are programs that operate unlike traditional AI programs. They are not hardcoded and therefore they're not restricted to working within the confines of what is already known. This new wave of AI is inspired by neuroscience and is capable of learning on their own first principles. This new wave includes such systems as deep learning and deep reinforcement learning, which, unlike traditional machine learning systems, are capable of finding patterns in data directly and making predictions and decisions entirely on their own independent of human intelligence or human judgment. |
| And this is where Rosetta differs from other managers using ML. We embrace this new wave of AI to the point that our A.I. is our investment process. Our models develop their own predictions, their own decisions directly from the data. Most to get raw unstructured time series data. It is precisely this non-human dimension that makes this new wave of AI so powerful. |
| As David Silver of Google's DeepMind points out, deep learning is more powerful than previous approaches because by not using human data or human expertise in any fashion, we've removed the constraints of human knowledge and it is able to create knowledge itself. Investment managers, especially quants and quants that have adopted traditional machine learning, readily concede that these powerful algorithms can be used to solve incredibly complex problems in medicine, autonomous driving, engineering, robotics and other verticals. |
| But they staunchly deny that deep learning and deep reinforcement learning can be used to solve investment problems and build successful autonomous investment strategies. In my view, that denial will be their downfall. |
Comments
In Channel
Download from Google Play
Download from App Store
United States00:00
00:00
x
