The Next Revolution in AI and ML 

This was nice! Keep doing this types of videos, a little technical , good explanation and not long! Thank you!

Thank you, very interesting!

I was deciding and narrowing down the topic for my PHD and this sort of resonates with me.

Hi Dimitri, I recently discovered the work of Cynthia Rudin, PI of the Interpretable Machine Learning Lab at Duke University. She argues that we should focus on creating complex algorithms that return sparse and intrinsically interpretable models, rather than simple algorithms that return complex and less interpretable models (e.g. random forests or GBM). What do you think about it?

Hey Dimitri, as someone about to start college and who is interested in the field of quant finance, thank you for your perspective always

Hello Dimitri 😀I am a Math graduate and a Data Science/Machine Learning MSc student about to begin my thesis. Over the past 3-4 months, I have developed a strong interest in Quantitative Finance and would like to focus my thesis on this area. Although I lack professional and academic experience in Quant Finance, I am eager to explore it further. Could you please suggest a few topics for my thesis? I don't know wich of the options I can follow (Sentiment analysis of financial news, AI-Driven portfolio optimization, etc.) as I am not very aware of the market last trends Thank you very much!

Hey Dimitri! Thank you for another amazing video! I think I can speak a bit on the tech industry side. You are right that mega differences come from coming up with novel mathematical solutions to existing problems which might be optimal in various ways. But, tech actually relies a lot on tinkering. For example, even now some of the best models in NLP and CV (LLMs and diffusion) use math techniques developed decades ago. Even the idea of predicting the error and simply using MSE in diffusion is so simple that i doubt they did the math before trying it randomly out of luck. Similarly, there are so many aspects like system design and the actual working of the internet, that ideally these should fail mathematically. The entire set of routing algorithms are a lot about best effort which means the given performance is not mathematically guaranteed. Coming back. An engineer might come up with an amazing heuristic based model for complex financial modeling tasks ehich might be wrong in certain cases but at the same time it may not make oversimplifying assumptions commonly used in Math to get some solvable (empirical or analytical) solution..

Hey Dimitri. I can't thank you enough for teaching us the importance of Maths. It has literally changed the career trajectories for good. GenAI is very cool now a days but it seems that most people (obviously not all) are just throwing it around in the meetings without even knowing the actual use case for themselves. "No amount of Maths is enough"...

hi dimitri, i was just wondering what kind of math do you think we will need in the next phase of ai/ml after this generative ai bubble/hype phase passes?

Yeah, honestly I bring up similar problems often with models that don't even have speed as a requirement; that at the end of the day with all these models, odds are that they aren't even going to converge on the true population parameter. It's depressing if you're looking for explainable AI.

Hey dimitri, are you going to do a video on going from physics to quant finance any time soon?

Hey, what are your thought on Quantum computing generally and possibly for finance? Most people I talk to in industry dismiss it as it's not coming any time soon but as a MS comp sci student I find it fascinating.

4:06 talking like a true physicist there 😂

Any book recommendations on this for DL grad student?

sufficiency

Does space state models is good in financial market

You dismissed the impact that quantum computing will have as a "know-it-all" blackbox, which, fair, nothing will ever know everything. But is there no threat of a small set of quantum computers knowing everything relevant enough - and fast enough - to squeeze out all competition - including new quantum computing entrants?

"No, it's not going to be the tech industry." I laughed so hard. This is so true. Subscribed for a long time, always love your explanations on this stuff!

Why are stock exchanges not fair?

Its so weird how you say its not groundbreaking, and proceed to mention things that weren't possible 9 months ago. You need perspective on what this applied math means to our species.

AI tools may level the playing field for average performance, but you're absolutely right that that won't create barriers to entry like sophisticated math can do. Think Jim Simons.

Dmitri. There are too many sharp angles on your face

lol the irony is this script was generated by an ai lol😂

your entire thesis is fundamentally wrong. People aren’t interested in AI for financial modeling. The holy grail of AI is artificial general intelligence. A foundational model that can solve generalized problems. Models in quantitative finance are more or less glorified regression. Do I think that the current LLM/transformer stacking hype will help us achieve AGI? No. But the end goal of AI research has nothing to do with finance and honesty this video makes you look like an old boomer that doesn’t have a vision. But I wouldn’t expect someone who’s working in risk modeling at a mid tier bank to be a visionary anyways.

You know nothing about high finance

Why do you keep talking about quants when you are not one, stop embarrassing yourself. On top of that, you are not AI expert. The only thing you questionably have knowledge of is ML but no one considers you authority there. I am not trying to be rude, you just need to ground yourself into reality.