machine learning

“We have, I'll say, dipped our toe with one. I would like to not talk to who it is, but they actually are using machine learning where the machines actually build the algorithms versus most quants, the people build the algorithms, and a lot of times you just ge…”

“To be clear, I don't think almost anybody has a lot of that in production today, but those tools are becoming more and more useful.”

“So as a result, we have started to focus more on applying these techniques to focus on something like a dividend cut. Let's use that as an example. Build a model that just forecasts or predicts dividend cuts, and that's it.”

“You can feed all the best data in the world to whatever ML algorithm you choose. Let's say it's like a decision tree or something. And if you're trying to predict returns, it always falls apart. It never works.”

“Invariably, we always have a section on the competitive advantage the company could build and compound by looping customer feedback data through the machine learning system to generate a prediction that gets better and better over time.”

“Machine learning is at the heart of the firm in modern times, and they were making trades without realizing why they were making the trades. The system teaches itself”

“So if your machine learning is finding something that standard regression models didn't find, it's probably not a good model.”

“Machine learning is just applied statistics, and it's what you learn when you took your statistics class in college or graduate school, and you read Thomas Bayes, who lived in the 1700, and Gauss, who lived in the 1800. And there's nothing that machine learnin…”

“That's, in some ways, the opposite of machine learning, which is a path of saying, we have no idea what patterns are Real or we don't have any theory. We don't have any hypothesis, but we do have a large data set. What can we learn in that data set that will g…”

“If you let it look at every possible factor, it's going to be a giant insane data mining exercise. I believe it'll have zero out of sample power.”

“I think I probably slowed us down on some of the machine learning stuff by a couple years with a little cynicism along the lines of exactly what you're talking about, that we could turn over more of the decision to the machines.”

“I'll give you an example of something ML I don't think will help us on, and maybe I'll be proven wrong about this, but vital numbers to people like us. What is the equity risk premium? What is the premium for high quality against low quality stocks? There, you…”

“At MDT, we've been using these machine learning tools since 2001. So we have a 24 year head start on someone who is new to the game.”