Alex Shtoff
@alexshtf.bsky.social
Principal scientist @ TII
Visit my research blog at https://alexshtf.github.io
Visit my research blog at https://alexshtf.github.io
Or distance from KKT conditions.
July 27, 2025 at 5:56 PM
Or distance from KKT conditions.
4) convergence of deviation from optimality conditions.
July 27, 2025 at 5:00 PM
4) convergence of deviation from optimality conditions.
I heard that polynomials are the (complex) root of all evil.
June 26, 2025 at 8:52 AM
I heard that polynomials are the (complex) root of all evil.
Indistinguishable from magic*
April 24, 2025 at 3:22 AM
Indistinguishable from magic*
Don't care who uses meta services. Don't like when people invent imaginary threats to their privacy and spread them. Don't want to use - dont use. Want to opt out and use - dont be afraid they won't comply. Meta is afraid of the legal and reputational consequences. That's my opinion.
April 19, 2025 at 2:47 PM
Don't care who uses meta services. Don't like when people invent imaginary threats to their privacy and spread them. Don't want to use - dont use. Want to opt out and use - dont be afraid they won't comply. Meta is afraid of the legal and reputational consequences. That's my opinion.
Used to work for Yahoo. Not a giant like Meta, but also use plenty of user data to make money. Not complying with regulation was always a big no no. These companies are very afraid of the legal and reputational consequences. So I wouldn't be afraid they won't comply.
April 19, 2025 at 2:28 PM
Used to work for Yahoo. Not a giant like Meta, but also use plenty of user data to make money. Not complying with regulation was always a big no no. These companies are very afraid of the legal and reputational consequences. So I wouldn't be afraid they won't comply.
The phenomenal paper "epigraphical analysis" by Attouch and Wets was the basis for my Ph.D thesis. It was fun digging deep into epi-convergence.
April 18, 2025 at 6:37 AM
The phenomenal paper "epigraphical analysis" by Attouch and Wets was the basis for my Ph.D thesis. It was fun digging deep into epi-convergence.
Yes, I understand. They should cite and criticize it.
April 12, 2025 at 12:20 PM
Yes, I understand. They should cite and criticize it.
Why simply not cite the directly relevant prior work?
April 12, 2025 at 11:52 AM
Why simply not cite the directly relevant prior work?
What makes it a method for "fine-tuning LLMs" rather than a method for fine-tuning any neural network in general?
March 28, 2025 at 9:30 AM
What makes it a method for "fine-tuning LLMs" rather than a method for fine-tuning any neural network in general?
Is it true that log(1+exp(x)) is the infimum of the quadratic upper bound over a?
If so - it also has interesting consequences.
If so - it also has interesting consequences.
March 24, 2025 at 5:39 PM
Is it true that log(1+exp(x)) is the infimum of the quadratic upper bound over a?
If so - it also has interesting consequences.
If so - it also has interesting consequences.
Or maybe there's cultural difference of the black people, who may be more afraid of not returning a loan and may do extreme things, such as using the last of their savings, to return it.
This paper seems to focus too much on estimation, and ignores the complexities of modeling.
This paper seems to focus too much on estimation, and ignores the complexities of modeling.
March 12, 2025 at 11:42 AM
Or maybe there's cultural difference of the black people, who may be more afraid of not returning a loan and may do extreme things, such as using the last of their savings, to return it.
This paper seems to focus too much on estimation, and ignores the complexities of modeling.
This paper seems to focus too much on estimation, and ignores the complexities of modeling.
As beautiful as I can remember it.
March 2, 2025 at 6:56 PM
As beautiful as I can remember it.
You can get away without - theory papers.
February 26, 2025 at 9:17 AM
You can get away without - theory papers.
Models reflect training data. Training data reflects people.
Models are just fancy autocomplete.
People have free will, and are not fancy autocomplete of what a model has showed them.
Models are just fancy autocomplete.
People have free will, and are not fancy autocomplete of what a model has showed them.
February 15, 2025 at 2:29 PM
Models reflect training data. Training data reflects people.
Models are just fancy autocomplete.
People have free will, and are not fancy autocomplete of what a model has showed them.
Models are just fancy autocomplete.
People have free will, and are not fancy autocomplete of what a model has showed them.
The numpy function there doesn't use SGD. To the best of my knowledge, it uses QR decomposition.
Anyway, things get interesting when the degree becomes 200 :)
Anyway, things get interesting when the degree becomes 200 :)
February 14, 2025 at 7:21 PM
The numpy function there doesn't use SGD. To the best of my knowledge, it uses QR decomposition.
Anyway, things get interesting when the degree becomes 200 :)
Anyway, things get interesting when the degree becomes 200 :)
Linear regression with Legendre polynomials:
colab.research.google.com/drive/1phA7N...
Inspired by Ben's post about nonexistent overfitting, to convince my coworkers.
colab.research.google.com/drive/1phA7N...
Inspired by Ben's post about nonexistent overfitting, to convince my coworkers.
Google Colab
colab.research.google.com
February 14, 2025 at 6:08 PM
Linear regression with Legendre polynomials:
colab.research.google.com/drive/1phA7N...
Inspired by Ben's post about nonexistent overfitting, to convince my coworkers.
colab.research.google.com/drive/1phA7N...
Inspired by Ben's post about nonexistent overfitting, to convince my coworkers.
Stochastic people.
February 10, 2025 at 11:00 AM
Stochastic people.
But, somehow, there is a "uniform prior" over the integers, according to some people I met :)
February 10, 2025 at 10:51 AM
But, somehow, there is a "uniform prior" over the integers, according to some people I met :)