Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Já tem conta? Entre
Researcher with public evidence across AI research.
Interesting piece in which @itsneuronal.bsky.social suggests dropping the common bit-per-spike metric & reporting the duration of held-out data to validate a model Seems to be just 120ms of data for a sharply tuned, high firing rate head-direction cell, i.e. just 3-4 spikes ar…
**Four** positions in AI/ML for biology at @dukemedschool.bsky.social, including in Neurobiology. @scottsoderling.bsky.social and David Page are leading a tremendous initiative in Discovery AI. Wonderful colleagues, wonderful place.
The difference between Σ s_i² and Σ s_i seems minor, but intuitively the former (CKA/RSA) places greater emphasis on the top PCs. Procrustes is often a more stringent notion of similarity, and this paper by Nathan Cloos shows that the difference can matter. proceedings.iclr.cc…
Your observation that "if s_i are the singular values of the centered cross-covariance, CKA weights shared modes through Σ s_i², while Procrustes uses Σ s_i" is exactly what we discuss here, alongside equivalence of Procrustes to Bures distance on Gram matrices proceedings.mlr…
§A.13.3 notes RSA and CKA differ over the diagonal vs row/column means. Indeed, adding a mean-centering step to RSA makes the two equivalent, for linear and nonlinear variants as shown here: proceedings.mlr.press/v285/william...
Enjoying reading this! Your Lemma 22 bounds Procrustes error by CKA in one direction, with a rank factor. Proposition 3 from our paper derives a two-sided version with a sharp constant that depends on participation ratio: proceedings.mlr.press/v285/harvey2...
Hiring postdocs @EPFL: please reach out and/or apply to the AI Center fellowship www.epfl.ch/research/fun.... My #NeuroAI group is excited about testing and building models of brain and behavior, from sensory through language to cognition, and from healthy to dysfunction (trea…
Center = Alex Williams. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.
Add the Who’s Who of AI badge to your site or bio. It links back to this profile.
Markdown: [](https://aiweekly.co/whos-who/person/itsneuronal-bsky-social)