a picture of the author

a self-portrait
(A. aurantia)

about me

hello. i'm alice. i am a mathematician (recently), a programmer (historically), and an entomophile (always). i feel as though i really ought to have a location on the web, for thematic consistency, so this is it. probably i will eventually put blog posts here which are not of general enough interest to be posted elsewhere (ie, about maths).

resume or cv available on request (if you are reading this, you probably have it already).

essays

my lw account is peralice. currently i have a couple of essays up there; a post about evolution as an analogy for ai training and a post about perceptions of trait commonality. on this site you can read about my attempt to win the acx 2026 prediction contest. i am (as of april 2026) writing up a summary of some of my attempts to produce a more rigorously justified rating algorithm; this should be on this website soonish. in short, all widely used rating algorithms increase their confidence in the assigned rating even when the events which occur are predicted to be extremely unlikely; i am trying to produce an efficiently-computable algorithm which avoids this shortcoming.

i also did some work on reed's list colouring conjecture (there isn't a name agreed on by the literature, as far as i know); i might write this up too. i have a proof that any graph with vertex list sizes exceeding d+O(sqrt(d)), where d is the maximum vertex-colour degree, is list-colourable, plus some fraction of a proof of the same with d+O(ln(d)). also some other cleaner proofs of things already proven in the literature. i believe that any graph with vertex list sizes strictly exceeding d+1 is colourable, but this seems to be very challenging to prove.

v0.08

most styling from html_wysiwyg, with some modifications.

additional styling for this page: