Andreas Loukas

@loukasa_tweet

Research fellow @ EPFL, fascinated by graphs and machine learning

Lausanne, Switzerland
andreasloukas.blog
Vrijeme pridruživanja: srpanj 2012.
5 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @loukasa_tweet

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @loukasa_tweet

  1. Prikvačeni tweet
    27. pro 2019.

    New work "what graph neural networks cannot learn: depth vs width" (ICLR20) studies the expressive power of GNN: It provides sufficient conditions for universality and shows that many classical problems are impossible when depth x width < c n^p. Blogpost:

    1 reply 15 proslijeđenih tweetova 51 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    2. velj

    50 cognitive biases, quite a few because of energy-intense human brains being so efficient at being more (energy-)efficient, e.g. Google effect, lazy causal inference, self-exceptionalism.

    140 proslijeđenih tweetova 552 korisnika označavaju da im se sviđa
    Poništi
  3. proslijedio/la je Tweet
    31. sij

    This is a nice package for making pyplot animations more intuitive: All you do is call "camera.snap()" every time you re-do the plot.

    141 proslijeđeni tweet 562 korisnika označavaju da im se sviđa
    Poništi
  4. proslijedio/la je Tweet
    21. sij

    We are organizing an Workshop on Geometric and Relational Deep Learning! Registration invites will be shared soon. Interested in participating? Consider submitting an abstract or get in touch: w/

    75 proslijeđenih tweetova 258 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. 15. sij

    Apparently, one of my friends has strong feelings towards my research (I found this on my whiteboard today :P)

    14 korisnika označava da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    13. sij

    This year's symposium on Conformal and Probabilistic Prediction will be in my hometown, Verona. Biased as I might be, I think it's an incredibly beautiful city. Importantly, CfP Abstracts: April 27th, Papers: May 4th. Make sure you submit your nice work!

    1 reply 2 proslijeđena tweeta 5 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    10. sij

    Very happy to share our latest work accepted at : we prove that a Self-Attention layer can express any CNN layer. 1/5 📄Paper: 🍿Interactive website : 🖥Code: 📝Blog:

    5 replies 348 proslijeđenih tweetova 1.152 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. 8. sij

    As a novice in Bayesian inference, I appreciated this explanation of "Deep Neural Networks as Gaussian Processes". The relation between the GP and Bayesian approaches was also illuminating.

    4 korisnika označavaju da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    7. sij

    Timely workshop! Deep learning scales up PDEs for complex systems (Schrodinger, Navier-Stokes/Euler, dynamic systems, cosmology) and PDEs theory improves DL (ResNets are discretized ODEs/transport equation that can be regularized for adversarial robustness).

    ML/AI and physical sciences Twitter - We are organizing a workshop on Differential Equations and Deep Learning at Maarten V. de Hoop -Rice Ron Fedkiw -Stanford Wilfrid Gangbo -UCLA -NTU Tom Goldstein -UMD Ricky T. Q. Chen -Toronto/Vector Gavin Portwood -LANL
    10 proslijeđenih tweetova 35 korisnika označava da im se sviđa
    Poništi
  10. proslijedio/la je Tweet
    13. pro 2019.

    Finally I can replace ‘watch -d -n .5 nvidia-smi’!!! Great work by ⁦⁦⁩ et al from ⁩ “GPU Dashboards in Jupyter Lab”

    84 proslijeđena tweeta 289 korisnika označava da im se sviđa
    Poništi
  11. proslijedio/la je Tweet
    3. sij

    Friends in the US traveling to : consider vaccinating against yellow fever in advance! CDC says "availability in the United States is currently limited. Travelers may need to schedule appointments well in advance and visit a clinic some distance away. " 1/2

    1 reply 12 proslijeđenih tweetova 37 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. 3. sij

    Preprint:

    1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    15. stu 2019.

    Just finished translation of my article regarding large graph visualization

    3 proslijeđena tweeta 3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    23. pro 2019.

    Visualizing MNIST classification. Neat.

    19 replies 1.797 proslijeđenih tweetova 5.729 korisnika označava da im se sviđa
    Poništi
  15. 30. pro 2019.

    This is something we need as a community.

    New community-driven effort for graph benchmarks: Hopefully we can soon finally stop using tiny datasets like Cora!
    2 korisnika označavaju da im se sviđa
    Poništi
  16. proslijedio/la je Tweet
    30. tra 2019.

    A common misconception is that the risk of overfitting increases with the number of parameters in the model. In reality, a single parameter suffices to fit most datasets: Implementation available at:

    520 proslijeđenih tweetova 1.473 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    22. pro 2019.

    Proud of my lab: 3 papers at ICLR but they realize they are lucky because there are so many good papers that got turned down. This roulette approach to science is putting immense pressure on young researchers.

    1 reply 9 proslijeđenih tweetova 116 korisnika označava da im se sviđa
    Poništi
  18. proslijedio/la je Tweet
    19. pro 2019.

    Decisions released 🎉 Congratulations to accepted papers; to those who we could not accommodate, we wish you success in your ongoing research. See our blog for the first of our reflections. See you soon in Ethiopia. 🇪🇹🌍

    8 replies 65 proslijeđenih tweetova 340 korisnika označava da im se sviđa
    Poništi
  19. 20. stu 2019.

    [Special Issue: Efficient Graph Algorithms in Machine Learning] Consider sending your original contributions on ! Deadline: 26 June 2020. (retweets appreciated)

    8 proslijeđenih tweetova 3 korisnika označavaju da im se sviđa
    Poništi
  20. proslijedio/la je Tweet
    19. stu 2019.

    The deadline for special sessions at in Amsterdam is coming soon ! More information at

    @eurasip @ToonvW @eusipco2019 @patrickanaylor
    6 proslijeđenih tweetova 4 korisnika označavaju da im se sviđa
    Poništi

@loukasa_tweet još nije objavio nijedan Tweet.

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·