Evgenii Egorov (@eeevgen) header
Evgenii Egorov (@eeevgen) profile picture

PhD researcher University of Amsterdam / AMLab. Structure-aware ML for scientific systems: inference, efficient architectures, quantum AI4Science.

Amsterdam
Joined April 2010
Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)Photo by Evgenii Egorov (@eeevgen)
An interlude about structure computation, SSM and attention. Myosotis: arxiv.org/abs/2509.20503 I hope to tell more in this line of work later. See poster at SPIGM workshop.
2
6
3
28
4,806
I would curios to see what LLM guys think about 6.2. For me guess that it is bigger model, more training but more aggressively quantized.
80
I also do this but well from mother tongue reasons 🤪
With AI I say "want i that" instead of "do i want that?" and "what means X?" instead of "what does X mean?" and still get good service so I'm never going back.
2
175
“And here one reviewer will put 4, but other three will put 1, because everything is trivial”
9
421
My long stand proposition is to stop split pdf on pages, for all the same reasons. It is not new, just recalling papyrus or svitok.
Don't see a reason to split them anymore. Used to do it before Overleaf to avoid conflicts when using Git. Now Overleaf organises it neatly with solid implementation of collaborative writing. Double click works well, search/replace works well, etc. Exception is preamble.
138
I was attending an event, where in the end were big QR code with speech something like “we are doing this big joint effort, scan and let us in how if you want to join”. I scanned and fill a form, where was explicitly option “PhD/postdoc”. Ответ убил.
1
168
I received answer shortly that they are not interested in PhD/postdoc level people but encourage me ti share with my advisor 🤪
1
1
61
I guess fair enough your church your rules but why all this steps haha, to be honest I wasn’t ready to meet a human version of self-closing box, I just found this hilarious.
31
Replying to @a_kzna
@a_kzna just delivers 🫡 I never tired to be amazed how she combines her mathematical knowledge with power of action
1
11
333
No good words about @BabakEht . He bit me in chess far too many times.
1
3
63
The most difficult part in post modern science is to acknowledge that post modern approach can be productive and even can be more productive and as consequence any approach to movement is okay. I saw a man running with his back forward and cant catch him.
1
185
Or they have stocks!!
英伟达员工有种!!! 突发:数百名英伟达员工签署了一封内部信函,要求恢复黄仁勋担任首席执行官,并要求董事会辞职!!
Readers added context they thought people might want to know
该帖声称数百名英伟达员工签署内部信函要求恢复黄仁勋为CEO并董事会辞职不实,附图信函为虚假内容。真实员工支持通过匿名内部建议箱评论(超1600条)和民调体现,并无此类正式信函报道。 wired.com/story/nvidia-e… theverge.com/ai-artificial-…
2
143
Each time I see some “I receive submit/email from PhD students/postdoc” I am thinking that this was encouraging culture for years for everything: emails, review, rebuttle, abstract, introduction, your name it.
1
1
228
You can check this very easy: just read neurips paper from several years in beginning and further and see this homogeneous structure eating authors alive.
1
143
So LLM slop is what we worked towards all time
46
I think this the same type of question as where is quantum advantage. Problem is it quite can be that question of interest have enough locality to be brute force llm’ed. And for other we don’t care.
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHO
1
204
This thing with Huang is just illustration of an important point that banned export to China is an implicit lobby of group of people and by therefore who is spy and who is not is well quite arbitrary over time.
1
1
212
My opinion always is that only way to avoid wars is more trades more export more import more joint corporations.
1
25
I am happy to give a new present to world: idea of negative increment. Imagine how this can enrich this simple schedule, when after move k you can have positive increment and after move n negative
⏳ What will the players' clocks look like in Geneva? ♟ Salesforce FIDE World Championship 2026 🇮🇳 Gukesh D 🆚 Javokhir Sindarov 🇺🇿 📍 Geneva, Switzerland 🇨🇭 🗓 Games: November 24 – December 12, 2026 PURCHASE YOUR TICKETS HERE: tickets.fide.com #GukeshSindarov
105
“A neural quantum state is a neural network whose training loss is the sampled expectation value of the Hamiltonian (Carleo and Troyer, 2017) and which is accepted when that loss converge“
1
2
213
I can’t make sense of the paper, because such lines drive me crazy.🤪 However, chatgpt can! I also ask it to evaluate writing quality and it was great until I cited some lines directly.
1
1
67
So if you use some GIGA PROMTS to assess quality of your paper, write review or something else – be aware our human little brain lives in a very different context than gpt.
54