Top all-time · #14956
Keep the gradient flowing
Personal blog · fa.bianp.net/blogmathaidata
Keep the gradient flowing is a personal blog at fa.bianp.net about math, ai and data. Since 2016, 5 different Hacker News readers have linked it 5 times, most often “Link Between Polynomials and Optimization”. It is still writing: the last post was on 2026-08-25, and you can follow it by feed.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
HN readers link this blog for its machine learning and optimization posts, such as one on surrogate loss functions and why logistic regression uses the logit. The linked comment cites it as support.
Summarised by Claude from a sample of 1 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Most-linked posts?
- Link Between Polynomials and Optimization2 people · 2 mentions · first linked 2020-04-10
What Hacker News says?
There are good reasons other than `it models the data well' to use a logit. Of course, the perspective that much like least-squares, logit regression does best when your data is generated from a logistic GLM is true, but not the only story. In my opinion, the best justification for the logit is that it is a easily-optimizable member of the family of surrogate losses: http://fa.bianp.net/blog/2014/surrogate-loss-func…
Comment linking /blog/2014/surrogate-loss-functions-in-machine-learning/
on “The Sigmoid Function in Logistic Regression”Submitted A birds-eye view of optimization algorithms as a story · 10 points
HN discussionSubmitted Link Between Polynomials and Optimization as a story · 3 points
HN discussionSubmitted The Russian Roulette: An Unbiased Estimator of the Limit as a story · 2 points
HN discussionSubmitted Link Between Polynomials and Optimization as a story · 2 points
HN discussion