Germans Savcisens (Savčišens)
@germansavePostdoc @NUnetsi (@KhouryCollege), PhD @dtutweet 👾 ✨ work on epistemic stability of LLMs🌿 my plants call me daddy 🦄 he/him 🇱🇻 🇺🇦 https://t.co/65Y11pkBXe
Similar User
@jeha_paul
@sigmabayesian
@Perbbr
@sdesojo
@laura_rieger_de
@robysinatra
@FZdyb
@vrskd
@suneman
@RasmusBonnevie
My data viz image was picked for a cover of the Jan 2024 issue of the Nat Comput Sci journal 😋
📢Our January issue is now live! Highlights include a model for predicting life outcomes, a Perspective on the advantages of language models for quantum simulation, and a protein language model for signal peptide prediction. 👉nature.com/natcomputsci/v…
Now is a great time to transition to #BlueSky—it suddenly became so alive. Find me there as converge.bsky.social
"Can I change my vote?" searches have spiked in the US after most of the races were called 😅 #Election2024US
I wouldn't go as far as stating that "PhD is bad for your mental health," but rather that the PhD support system and treatment of the students are poor.
Strong evidence showing that getting a PhD is extremely bad for your mental health. A new paper uses Swedish medical records and matches them to the full population of PhD students for which the authors could get gender and birth year data from 2006 to 2017. After some exclusion…
Little victories in Boston 😋 Silver medal for the volley
Just got back from @IC2S2 🌟 Although I didn't present, I was blown away by the incredible research and met so many brilliant people (despite being super awkward in social situations). Now, I feel super energized to dive back into my work in Boston. 🧠#researchlife
btw, our paper "Using sequences of life-events to predict human lives" is now in the open-access: nature.com/articles/s4358…
This week, I am attending @IC2S2 in Philladelphia. Got welcomed with 39°C and a massive lightning show. 🌩️⚡️ Excited to explore new ideas! Let me know if you would like to meet for a coffee ☕️ #IC2S2 #Philadelphia
Hyperbolic Random Forests at been accepted to #TMLR! Random forests split data hierarchically, making for strong links with hyperbolic space. We show how to generalize forests through horospheres, resulting in powerful machine learning models. Paper: openreview.net/pdf?id=pjKcIzv…
The best interpretation of the claim that GPT-5 will have a PhD level intelligence 😂
Strong evidence showing that getting a PhD is extremely bad for your mental health. A new paper uses Swedish medical records and matches them to the full population of PhD students for which the authors could get gender and birth year data from 2006 to 2017. After some exclusion…
I just found a book about #life2vec, written by a chatbot (probably), and sold on Amazon. I guess #life2vec has a life of its own at this point. 😁🫠
New preprint 🚀 When should we favour a model with higher-order interactions over classical pairwise ones? "Functional reducibility of higher-order networks" Arxiv: arxiv.org/abs/2404.08547 Code: t.ly/pdELP w/ @l_gajo @ArshamGhavasieh @fede7j @manlius84 👇🧵
New postdoc opening in my group! If you're interested in using cutting edge methods to understand how our social networks contribute to life outcomes through analyzing with one of the most amazing datasets in the world, this is the job for you! efzu.fa.em2.oraclecloud.com/hcmUI/Candidat…
#life2vec by @germansave himself, it's not a death calculator #databeers #databeerscph @databeers
Thrilled to announce the fourth speaker for the DataBeers x @Absalon1720 event next week: @germansave, who will talk about the evolution and predictability of human lives 👀 🎟️Get a free ticket and join the fun🍻: absaloncph.dk/event/databeer…
Want to hear more details about the #life2vec, #LLMs, and its role in understanding life trajectories? ... and also see me being anxious and awkward for a couple of hours 😅 Drop by my PhD defence on 13th March at @DTU_Compute More info here: compute.dtu.dk/english/kalend…
When it comes to #timeseries one can not expect to train LTSM (large time series model) on fake (aka “synthetic” data) and obtain good performance. An interesting finding is that the zero-shot performance of these models is significantly influenced by the source dataset,…
United States Trends
- 1. Serrano 110 B posts
- 2. #TysonPaul 67,6 B posts
- 3. #NetflixFight 35,7 B posts
- 4. Ramos 62,9 B posts
- 5. #buffering 5.627 posts
- 6. #netflixcrash 2.943 posts
- 7. Rosie Perez 7.736 posts
- 8. My Netflix 45,4 B posts
- 9. ROBBED 48,7 B posts
- 10. Jerry Jones 6.569 posts
- 11. #SmackDown 65,9 B posts
- 12. WTF Netflix 10,6 B posts
- 13. Cedric 1.062 posts
- 14. Christmas Day 13,5 B posts
- 15. Love is Blind 4.181 posts
- 16. The Netflix 201 B posts
- 17. $NFLX 4.324 posts
- 18. Streameast 2.270 posts
- 19. Peter 150 B posts
- 20. Goyat 24,4 B posts
Something went wrong.
Something went wrong.