Laurynas Rekasius

@LaurynasRek

Data Scientist / lifelong education believer. Follow me if you want to know about best free education resources

Denver, CO
Vrijeme pridruživanja: ožujak 2019.

Tweetovi

Blokirali ste korisnika/cu @LaurynasRek

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

  1. 30. lis 2019.

    Interesting read about steps in search for but useful for everyone else as well. I think it is always helpful to remind yourself about key aspects in this journey.

    Poništi
  2. 23. lis 2019.
    Poništi
  3. 21. lis 2019.

    Interesting read about self-directed learning. My attention caught characters of people who do that: -Highly reflective -self efficient -supportive -higher responsibility -inquisitive It also has some tips on how to bevome self-directed learner

    Poništi
  4. proslijedio/la je Tweet
    15. lis 2019.

    We've trained an AI system to solve the Rubik's Cube with a human-like robot hand. This is an unprecedented level of dexterity for a robot, and is hard even for humans to do. The system trains in an imperfect simulation and quickly adapts to reality:

    Prikaži ovu nit
    Poništi
  5. 4. lis 2019.

    D19: continue to learn how to automate ppt creation with R and making the output tables look very clean. Super tedious works when you need to take care of each number individually 🙂Libraries I use: officer, magrittr, ggpubr, data.table

    Poništi
  6. 3. lis 2019.

    D18: learned about RAdam optimizer and had and interesting discussion about current state and is it possible that we will have another AI winter. Great time with fellows at meetup

    Poništi
  7. 3. lis 2019.
    Poništi
  8. 1. lis 2019.

    Super scary and exciting at the same time😃 What China is doing would be impossible anywhere else in the world. This is good and bad at the same time. As we protect our privacy and rights we will be also left behind in revolution

    Poništi
  9. 30. ruj 2019.

    D17: continued work on fraud competition. Did some feature engineering, moved modelling to a new system and did my last submission. Let’s see how the final result will look tomorrow

    Poništi
  10. 29. ruj 2019.

    D16: Today I learned that I have no idea how to properly manage resource usage when modeling big datasets (subsampling didn’t work)And that’s especially a problem on enviroments like . It crashes constantly 😂😂 any useful resources on this subject?

    Poništi
  11. 28. ruj 2019.

    D15: second day of going up and down on public leaderboard, but in the end I am not sure how much my result has improved 🤔 Doing oversampling should have helped with generalozation as well as variable removal, but score does not show that..

    Poništi
  12. 27. ruj 2019.

    D14: finally entered first ever competition! Doing fraud detection challenge. Baseline result submitted and luckily I am not the last one on the leaderboard 😃 only couple of days until competition will close, so let’s where I end after this weekend 🙂

    Poništi
  13. 27. ruj 2019.

    D13: learned more about driverless ai tool. Looks like a thing definitely worth trying to speed the process of solving more typical problems like churn prediction, credit worthiness etc.

    Poništi
  14. 26. ruj 2019.

    If Facebook owns reality, it'll be impossible to opt out. It's super scary, but also so incredible that I can't wait this to happen! via

    Poništi
  15. 26. ruj 2019.
    Poništi
  16. 25. ruj 2019.

    D12: learning about platform. Let’s what we are going to learn on this practical lecture 🙂

    Poništi
  17. 24. ruj 2019.
    Poništi
  18. 24. ruj 2019.

    Algorithms are modern day gods and mathematicians together with computer scientist are like priests who try to explain god’s will to ordinadry people. - from the book - Weapons of Math Destruction. I am already hooked 🙂 will share more insights

    Poništi
  19. 22. ruj 2019.

    D11: learned about importance of ethics in and possible issues that biased data could bring and how diverse teams could help to address this type of problem upfront. Will read more about this-“Weapons of Math Destruction”

    Poništi
  20. 21. ruj 2019.

    D10: leaned about embedding in and regularization. It was also interesting to know that public leaderboard could be very misleading and that’s why you should learn how to create very good validation sets on your own and trust in them.

    Poništi

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:

    ·