Nick Burns

@nickdaleburns

Data scientist, avid squash player, coffee drinker and dabbler in NLP and RL.

Vrijeme pridruživanja: ožujak 2018.
21 fotografija ili videozapis Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @nickdaleburns

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

  1. proslijedio/la je Tweet
    prije 9 sati

    Now that and Charles derived these beautiful gradients... This year is gonna be all about Bayesian inference for HMMs using Hamiltonian Monte Carlo. And all the practical ways that things can go wrong when the data and model don’t agree.

    New paper from myself, Charles Margossian, and , ! We derive a discrete analogy to the adjoint method for ODEs and use it to work out the gradient of the hidden Markov model marginal likelihood in a new way.
    Prikaži ovu nit
    1 reply 4 proslijeđena tweeta 19 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. 1. velj

    I generally steer away from political posts - but this visualisation looks like it has been very carefully and purposefully designed to mislead. This is not-so-subtle subconscious manipulation.

    To quote "If writing the numbers on a misleading graph was enough to stop it being misleading, there would be no point having data graphics"
    1 proslijeđeni tweet 1 korisnik označava da mu se sviđa
    Poništi
  3. 31. sij

    So excited about Statistical Rethinking V2. McElreath's first edition changed my whole perspective. Cant wait!

    Nice post from Bob Carpenter about advantages of marginalizing out discrete parameters. Forthcoming 2nd edition of my book has a new section on how to do this in code.
    1 korisnik označava da mu se sviđa
    Poništi
  4. proslijedio/la je Tweet
    30. sij

    Wait WHAT? scales::show_col() let's you see what the palette looks like?? How did I not know this!! at

    35 proslijeđenih tweetova 166 korisnika označava da im se sviđa
    Poništi
  5. 31. sij

    "GLAMOUR of graphics". Awesome slides 🤾

    Slides from my talk are up on my site!
    2 korisnika označavaju da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    29. sij

    I've always been fascinated by Bayesian Nonparametrics. I struggled to grasp those ideas directly from papers. Today, by chance, I found the best (imo) single reference for anyone interested: A gentle introduction by and Michael Jordan

    45 proslijeđenih tweetova 218 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    29. sij

    Many of the most useful benefits to adopting Bayesian inference come from bespoke modeling that doesn’t depend on asymptotics to be valid. Ironically this is possible because Bayes is _easier_ to implement than frequentist methods.

    One of the reasons that I think and other frameworks are useful is because they force you to think about your model really hard -- every step of the way is defined by yourself. Even if you're not a Bayesian, an equivalent framework for frequentist would go a long way.
    Prikaži ovu nit
    13 proslijeđenih tweetova 57 korisnika označava da im se sviđa
    Poništi
  8. proslijedio/la je Tweet
    29. sij

    Ep. 8 of 'Learn Bayes Stats' is online! How do you apply Bayesian tools in the online ad industry or when you’re a software engineer or computer scientist? tells you how to use Bayesian thinking every day! See U in your favorite podcatcher or @

    5 proslijeđenih tweetova 19 korisnika označava da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    29. sij

    The Internet of Things will continue to be a terrible idea until this sort of nonsense stops.

    9 proslijeđenih tweetova 24 korisnika označavaju da im se sviđa
    Poništi
  10. proslijedio/la je Tweet
    29. sij

    I am working on an ebook, tentatively called Bite Size Bayes, that introduces Bayesian statistics gradually, for people with no prior stats. Here's Python notebook 2, if you want to check it out. R version coming soon!

    38 proslijeđenih tweetova 150 korisnika označava da im se sviđa
    Poništi
  11. proslijedio/la je Tweet
    28. sij

    Are GWAS studies of IQ/educational attainment problematic?

    7 proslijeđenih tweetova 13 korisnika označava da im se sviđa
    Poništi
  12. proslijedio/la je Tweet
    28. sij

    Causal inference in AI: Expressing potential outcomes in a graphical-modeling framework that can be fit using Stan

    1 reply 14 proslijeđenih tweetova 81 korisnik označava da mu se sviđa
    Poništi
  13. proslijedio/la je Tweet
    27. sij

    Relevance Vector Machine (RVM) is like an SVM but with probabilistic classification. NASA is using it for "Remaining Useful Life Estimation".

    1 reply 5 proslijeđenih tweetova 45 korisnika označava da im se sviđa
    Poništi
  14. proslijedio/la je Tweet
    26. sij

    In the tradition of xxxToVec - here is Time2Vec - pretty interesting - represent time as a learned embedding - might try this with some eeg data.

    If you are building ML models that consume time as a feature, check out our latest work where we show how learning a vector representation for time may improve model performance:
    58 proslijeđenih tweetova 202 korisnika označavaju da im se sviđa
    Poništi
  15. proslijedio/la je Tweet
    24. sij

    You can code? Great. Congrats on having a superpower. Now what the fuck are you going to do with it?

    17 replies 15 proslijeđenih tweetova 171 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    24. sij

    Wow, generating Probability program from code! Nice work !

    1 reply 5 proslijeđenih tweetova 47 korisnika označava da im se sviđa
    Poništi
  17. proslijedio/la je Tweet
    23. sij

    [From the archives:] How Graphs Enhance – a SF GraphTour talk with

    8 proslijeđenih tweetova 8 korisnika označava da im se sviđa
    Poništi
  18. 23. sij

    Nice!

    0:13
    : Is your QQ-plot generated from a normally distributed process? Draw random data from a norm dist of the same length as your data set over & over. This will give you a good idea of how a normal QQ-plot should look. Try it with `ggpubr::ggqqplot()` for easy plots
    1 proslijeđeni tweet 1 korisnik označava da mu se sviđa
    Poništi
  19. 23. sij

    Amazing!

    Following the tradition, I am going to share all the course material for "STAT 453: Introduction to Deep Learning and Generative Models" I am teaching this semester :)
    Poništi
  20. proslijedio/la je Tweet
    23. sij

    Improve predictions with domain knowledge: Ch 6 from Bayesian Hackers in Probability, newly updated. Done with my first pass of all chapters, now back to Ch. 1 to make more TF2 friendly.

    i
    1 proslijeđeni tweet 8 korisnika označava da im se sviđa
    Poništi

@nickdaleburns 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:

    ·