Vai al contenuto

Dmitry Kobak

@hippopedoid.bsky.social did:plc:tr4nc2rjorkbm4b4gpnohe4e

PI at Ghent University and VIB.AI. Manifold learning, contrastive learning, self-supervised learning, genomics, transcriptomics, interpretability. Excess mortality, statistical forensics. Born but to die and reas'ning but to err. https://dkobak.github.io

Iscritto a Bluesky il

Follower
335
Seguiti
67
Post
20
Giorni in classifica
0

Riepilogo

Dmitry Kobak (@hippopedoid.bsky.social) è un utente di Bluesky iscritto il 28 novembre 2024. Ha 335 follower, segue 67 account e ha pubblicato 20 post.

Presenze in classifica

I giorni migliori in cui i post di questo account sono entrati nella classifica giornaliera di Debya.

Questo account non è ancora comparso in una classifica giornaliera.

Post piaciuti di recente

Like pubblici dalla rete Bluesky.

  • Karsten Konrad 🇪🇺@zabong69.bsky.social

    The arithmetic thing is a good example. LLMs learn adding numbers by representing addition as geometric operations in the NN and this can add numbers they have never seen in training. As this contradicts the stochastic parrot meme, Dr. Bender simply claims cheating and drops out of the discussion

    Mostra post
  • Leland McInnes@lelandmcinnes.bsky.social

    UMAP 0.6 is out! A new initialization, new optimizers and faster nearest neighbour search make UMAP much faster and far better at large scale. On 3 million arXiv papers it finishes in under 5 minutes, before UMAP 0.5 has even found its nearest neighbours.

    Mostra post
  • hikikomorphism@hikikomorphism.bsky.social

    If you can substitute "hungry ghost trapped in a jar" for "AI" in a sentence it's probably a valid use case for LLMs. Take "I have a bunch of hungry ghosts in jars, they mainly write SQL queries for me". Sure. Reasonable use case. "My girlfriend is a hungry ghost I trapped in a jar"? No. Deranged.

    Mostra post
  • Hendrik De Bie@debiehendrik.bsky.social

    Thanks, I was a bit confused about mastodon. Yes, he has said many things over the past years that I found objectionable (even though before that did not seem to be the case). When I was younger, I found Grothendieck a bit too outspoken on certain issues, but the older I get, the more I am

    Mostra post
  • Ryan O'Donnell@booleananalysis.bsky.social

    I lectured on the "ETH false" (#3SAT in subexponential time) result in my grad complexity class at CMU today. Didn't have time in 80 mins to do the full proof, but I hope I got across the main plan, and some of the basic TCS ingredients going into it. youtu.be/_HlQGMcQWhA

    Mostra post

Accedi per vedere di più

Account che blocca

I blocchi su Bluesky sono registri pubblici.

Questo account non blocca nessuno.