Botos Csabi

326 posts

Botos Csabi banner
Botos Csabi

Botos Csabi

@csaba_botos

Encode Fellow of the 2025 cohort. PhD from University of Oxford https://t.co/WzU6h3qhag

Oxford, England Entrou em Mayıs 2017
273 Seguindo217 Seguidores
Botos Csabi
Botos Csabi@csaba_botos·
Isn't harness building, context editing and prompt search analogous to LoRA, but in a semantically meaningful subspace? It's just cheaper (no need to host the model), safer (catastrophic forgetting less likely), more interpretable (you can read now what lora would do!)
English
0
0
1
80
Botos Csabi
Botos Csabi@csaba_botos·
it's funny to see how mankind got advantage by fencing off a plot of land, taking it into his head to say ‘This is mine’ and now thousands of years later to advance science we need to remove the metaphorical fence, let messy nature back in to access more fruits of our land
English
0
0
0
56
Botos Csabi
Botos Csabi@csaba_botos·
shower thought: how would you go about teaching a dog to teach other dogs a trick?
English
0
0
0
59
François Fleuret
François Fleuret@francoisfleuret·
Hear me out: an FFT is a linear convolutional encoder and decoder.
English
52
30
613
79.6K
Botos Csabi
Botos Csabi@csaba_botos·
all the soap opera around NeurIPS, ICML and ICLR is a direct cause of real world forces acting on an artificial community. Having a paper can get you a job, a grant, a promotion, regardless whether the paper itself is pure garbage or seminal breakthrough.
English
0
0
6
860
Micah Goldblum
Micah Goldblum@micahgoldblum·
An LLM-generated paper is in the top 17% of ICLR submissions in terms of average reviewer score, having received two 8's. The paper has tons of BS jargon and hallucinated references. Fortunately, one reviewer actually looked at the paper and gave it a zero. 1/3
English
39
146
1.5K
514.4K
Botos Csabi
Botos Csabi@csaba_botos·
the bonus bi-product would be some excellent training data for the future AI scientist agents. Same way as we learned the dark arts of ML by just reimplementing famous papers in the first year of the PhD.
English
0
0
0
77
Botos Csabi
Botos Csabi@csaba_botos·
Neurips R2 thought: why is the main objective for building AI scientists to accelerate new discovery while the reproducibility crisis is already crippling many fields. Maybe less ambitious but more robust and likely to be impactful work would be automating reproducibility check.
English
1
0
3
190
Botos Csabi
Botos Csabi@csaba_botos·
@francoisfleuret a professor making a meme about this pain just feels surreal. big time
English
0
0
0
23
François Fleuret
François Fleuret@francoisfleuret·
TFW for a second you thought that "test_accuracy: 99% 99/100" was the test accuracy while it was the progress bar of the loop computing the test accuracy, actually equal to 3%.
English
5
0
58
4.9K
Botos Csabi
Botos Csabi@csaba_botos·
Amazing experience with the prettiest view @NeurIPSConf presenting Label Delay botcs.github.io/label-delay/ the original session was 3 hours - stayed 2h more for follow-up questions, people's interest in the topic was way above my expectations :) feeling awesome!
Botos Csabi tweet mediaBotos Csabi tweet mediaBotos Csabi tweet mediaBotos Csabi tweet media
English
0
0
7
277
Botos Csabi
Botos Csabi@csaba_botos·
Yay 🦜 Heading to Vancouver on Monday to present my first ever @NeurIPSConf paper. I am super duper excited to finally join the cool kids’ gang. Check out the tutorial I procrastinated over for 3 months: youtu.be/EbPrlGkLxGo?si…
YouTube video
YouTube
Botos Csabi tweet media
English
1
0
16
433
Botos Csabi retweetou
Ferenc Huszár
Ferenc Huszár@fhuszar·
Call For Tasks for IOAI 2025: ML olympiad for high schoolers. I know you all have strong opinions on the ML topics future computer scientists should engage with to build mastery. This is a high-impact opportunity to help us by proposing tasks. ioai-official.org/call-for-tasks/
English
0
4
6
1.8K
Botos Csabi retweetou
Ferenc Huszár
Ferenc Huszár@fhuszar·
Teams of Hungary took 5th and 6th place in the first International Olympiad in AI taking the top 2 spots for silver 🥈. Congrats to the participants who brought a lot of self-taught expertise, adaptibility to the event continuing to build our tradition of problem solving. 🇭🇺🇭🇺🇭🇺
Ferenc Huszár tweet mediaFerenc Huszár tweet mediaFerenc Huszár tweet mediaFerenc Huszár tweet media
English
6
4
36
4.2K
Botos Csabi
Botos Csabi@csaba_botos·
@francoisfleuret Oh there's one more! in my undergrad I was fascinated with the hydrogen pumps that also act as tiny electrical engines in cellular biology, as it is a massive protein complex whose modules are highly unlikely to develop independently. en.m.wikipedia.org/wiki/V-ATPase
English
1
0
2
62
Botos Csabi
Botos Csabi@csaba_botos·
@fhuszar This is all the athletic people in bars at 2:30am when "don't stop me now" comes on the jukebox.
English
0
0
0
42
Ferenc Huszár
Ferenc Huszár@fhuszar·
Interdimensional Championships
English
1
1
11
3.1K
Ferenc Huszár
Ferenc Huszár@fhuszar·
Great day in Oxford joining forces with @oiwi3000 as co-examiners for the DPhil viva of @csaba_botos Congrats Csaba for the great thesis and amazing live demos.
Ferenc Huszár tweet mediaFerenc Huszár tweet mediaFerenc Huszár tweet mediaFerenc Huszár tweet media
English
2
0
16
2.1K

Descobrir

@SamuelAlbanie @francoisfleuret @micahgoldblum @AdityaHumnabad1 @karpathy @carobellum @NeurIPSConf @elonmusk