Borysław Paulewicz

@BRSLWP

a theoretical cognitive psychologist - SWPS University, faculty in Katowice - c-lab, Jagiellonian University, Kraków

Vrijeme pridruživanja: srpanj 2013.

Tweetovi

Blokirali ste korisnika/cu @BRSLWP

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

  1. proslijedio/la je Tweet

    This tweet was partly inspired by a coffee machine chat with the head of another department when he was about to retire. He said he looked forward to having time to read and think. I think most people wouldn't expect scientists feel they need to retire to have time for this.

    Poništi
  2. prije 7 sati

    If you are willing to use the term "correlational study" then you probably do not understand the difference between observation and random assignment.

    Poništi
  3. proslijedio/la je Tweet

    I never babytalk to my hamster. I use entire Nature articles and proper terminology, including not confusing perfusion and diffusion Academics at faculty meetings are shocked by her aptitude. She isn’t even 14 months and her intellect far exceeds that of most professors.

    Poništi
  4. 29. sij

    I just had a pizza. I have also defined a generalized link function for ordered criteria/thresholds, the log_distance and log_ratio link functions are special cases of this glorious (or trivial, depending on your point of view) generalized link function:

    Poništi
  5. 29. sij

    This looks great! Also, is ANOVA still a thing?

    Poništi
  6. 29. sij

    I guess this may be of interest to you, : -)

    Prikaži ovu nit
    Poništi
  7. 29. sij

    SDT NEWS: The bhsdtr package now implements two new simple and intuitive link functions (log_distance nad log_ratio) for the criteria that preserve the ordering and allow for individual criteria effects (see the github page for more details):

    Prikaži ovu nit
    Poništi
  8. 29. sij

    An over the top paper I wish I had the guts/expertise to write: a weak definition of purposefulness from the RL framework+theory of tasks from AI+category-theoretic mapping between mechanistic and teleological descriptions->a general definition of behavior and cognitive functions

    Poništi
  9. proslijedio/la je Tweet
    27. sij

    Bad news for meta-analyses: Comparing 15 meta-analyses to multi-lab replication attempts of the same effects, on average, meta-analytic effect sizes are almost three times as large as replication effect sizes.

    Poništi
  10. 28. sij

    A generalized parametrization!

    Prikaži ovu nit
    Poništi
  11. 28. sij

    Sorry, should have said Selker, can den Bergh, Criss, and Wagenmakers' parsimonious model. See this great paper:

    Prikaži ovu nit
    Poništi
  12. 28. sij

    SDT unconstrained criteria vector (for hierarchical Bayesian SDT). A range of meaningfully simplified models (e.g. parsimonious model) can be obtained just by fixing some elements of tau at 0.

    Prikaži ovu nit
    Poništi
  13. 27. sij

    Another (much simpler) idea for an SDT criteria link function: the main criterion is unconstrained, the rest of the criteria are represented as log of distances between the adjacent criteria. Ordering can be preserved and individual criteria (random / fixed) effects are possible!

    Poništi
  14. proslijedio/la je Tweet
    23. sij

    Explainability is ML's Achilles heel. This paper presents formal criteria & measures for various aspects of explainability. Its use of do-calculus gives readers immediate view of the data and assumptions needed to meet each of the criteria.

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    22. sij

    The latest Sci-Hub working domain(Last check time:Wed, 22 Jan 2020 17:13:01 GMT) Please retweet to your friends,help us become stronger!

    Poništi
  16. 21. sij

    I believe you may find this interesting :-)

    Prikaži ovu nit
    Poništi
  17. 21. sij

    We also take a (somewhat lighter) swing at Macmillan and Creelman (Detection Theory: A Users Guide) and McNicol (A Primer of Signal Detection Theory)

    Prikaži ovu nit
    Poništi
  18. 21. sij

    A novel parametrization of SDT with ratings + a general hierarchical linear regression structure (+R package [handles meta-d']). We argue that this one is correct and brms-cumulative, Morey, Pratte, & Rouder, and HMeta-d' are not (xcpt in special cases).

    Prikaži ovu nit
    Poništi
  19. 21. sij

    If you like the Bayes Factor then you may find this (very) useful!

    Poništi
  20. 20. sij

    Staircase introduces a well defined set of causal relations. Here is why it does not deconfound metacognition.

    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:

    ·