Viewing profile — dpaleka

More targeted training won't do good, but why wouldn't more search help? My understanding is that gwern above linked solid evidence in the paper for more search not being enough, a…

More search won't do good, but why wouldn't targeted training help? The way I see it is that the adversarial policy search discovers positions which are off-distribution for anythi…

That paper (ROME) was the most famous paper in the field last year :) See also new interesting developments breaking the connection between "Locating" and "Editing": https://arxiv.…

[I mean no bad faith in this comment, I'm a fan of yours.] Why answer questions about harmlessness/safety in such a roundabout way? Both OpenAI and Anthropic are clear about what w…

A weird property of the described abstractions is that as you go tighter (interval -> zonotope -> polyhedra), the trained networks counterintuitively become less robust. Why does m…

I too use large files with random notes, but I can't be bothered to write dates -- so I use git and cron to automate a searchable, persistent diary. Let me write a blog post about …

I feel the current marketing of topological data analysis and similar is directed towards people who have at least taken several graduate math courses. I'm really not an expert in …

Thought that the article would be abou the quest for a better random graph model. It's weird that Quanta didn't talk about this yet. For the other direction (ML helping graph theor…

How do models trained with Lightly compare with other approaches wrt adversarial robustness? Can using Lightly introduce additional bias in the model, since only a select few of in…

No one is reaching out in Switzerland right now, but I guess they should.

Implementation where you can try to disprove conjectures yourself: https://github.com/dpaleka/cross-entropy-for-combinatorics