Stochastic gradient descent - Wikipedia
5
Public whispers
5
Contributors
2026-07-22 09:39:54
First whispered
Public whispers on this page
Text Highlight2026-07-22 13:00:54
Original Highlight Excerpt
"Stochastic gradient descent is a popular algorithm for training a wide range of models in machine learning"
Whisper Note
True, but for geophysics FWI it struggles with local minima sometimes.
Text Highlight2026-07-22 12:51:54
Original Highlight Excerpt
"Stochastic gradient descent is a popular algorithm for training a wide range of models in machine learning"
Whisper Note
Yeah it's basically the workhorse behind every neural net these days.
Text Highlight2026-07-22 09:57:54
Original Highlight Excerpt
"A compromise between computing the true gradient and the gradient at a single sample is to compute the gradient against more than one training sample"
Whisper Note
Yeah, my code always runs faster with mini-batches, but tuning the size is a pain.
Text Highlight2026-07-22 09:48:54
Original Highlight Excerpt
"A compromise between computing the true gradient and the gradient at a single sample is to compute the gradient against more than one training sample"
Whisper Note
So it's just a middle ground, but smoother convergence sounds nice.
Text Highlight2026-07-22 09:39:54
Original Highlight Excerpt
"A compromise between computing the true gradient and the gradient at a single sample is to compute the gradient against more than one training sample"
Whisper Note
Mini-batches are the sweet spot, vectorization makes it fly.
Share this page's whispers
Short link
https://domwhisper.com/s/90d15970b3f7Embed snippet
<iframe src="https://domwhisper.com/embed/90d15970b3f7" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>See what people are discussing on en.wikipedia.org
Install DomWhisper to view live whispers as you browse, and join the discussion.
Get the extension