Josh Glaser

@joshuaiglaser

Machine learning for neuroscience. Postdoctoral research scientist at Columbia.

Vrijeme pridruživanja: ožujak 2018.

Tweetovi

Blokirali ste korisnika/cu @joshuaiglaser

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

  1. proslijedio/la je Tweet
    31. sij

    Cosyne schedules are up! Be sure to come to our workshop (organized w/ and Matt Whiteway) "Interpretable computational neuroscience: What are we modeling and what does it have to do with the brain?"

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    23. sij

    I'm very excited to share that my work with and on S1 representation of reaching is now published in eLife! Thanks to some insightful reviewer comments, I think it's become a much stronger paper now. Begin thread... (1/6)

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    16. sij

    Excited to share what and I have been working on recently! We unify and generalize statistical models of neural dynamics during decision-making using switching state-space models

    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    29. lis 2019.

    So let's say you record 100 neurons from each of 5 animals in response to 2 different stimuli... or count dendritic spines on 100 neurons in each of 10 animals, with half of animals subjected to 1 of 2 drug conditions... Then you've got hierarchical data. (THREAD)

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet

    During natural scene search, the frontal eye field represents the prior probability of a saccade direction. Paper finally out with , Wood, Segraves :

    Poništi
  6. proslijedio/la je Tweet
    9. lis 2019.

    Check out our new preprint on using multi-area recurrent neural networks to better understand decision-making. This is joint work with first author and co-senior author .

    Poništi
  7. 1. lis 2019.

    We've turned our "Machine learning for neural decoding" paper into a tutorial. Check it out! w/ And our code package:

    Poništi
  8. proslijedio/la je Tweet
    21. ruj 2019.

    "Due to random chance, some percentage of models will outperform other ones, even if they are all just as good as each other. Maths doesn’t care if it was one team that tested 100 models, or 100 teams."

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet

    Neuroscience needs interpretable models of neural population activity. Great new review relating this to latent variable models.

    Poništi
  10. proslijedio/la je Tweet

    Animals will respond differently to the same inputs when they are hungry/sated, attentive/bored, etc. But how do we know what state the animal is in? In this preprint with , we show how you can do that automatically! 👇

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    21. svi 2019.

    We have a new preprint on bioRxiv Area 2 neural coding in monkeys it describes I want you to read it But so I could tweet it I wrote these six limericks to serve as a guide! (1/6)

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    2. svi 2019.

    Hierarchical Recurrent State Space Models Reveal Discrete and Continuous Dynamics of Neural Activity in C. elegans This has been a big part of my postdoc and I'm glad to finally share it! Huge thanks to my coauthors/mentors Annika, Dave, Manuel, and Liam.

    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet

    Omitted variable bias is what makes confounding problematic. I love 's analysis of the effect of the omitted variable bias on GLMs:

    Poništi
  14. proslijedio/la je Tweet

    Call for Applications: NeuroNex Junior Scientist Workshop on Advanced Neural Data Analysis, Aug 26-29 2019. Details here:

    Poništi
  15. proslijedio/la je Tweet
    14. ožu 2019.

    Check out our new review on the importance of analyzing population activity in neuroscience! "Towards the neural population doctrine."

    Poništi
  16. 11. ožu 2019.

    A final version of our review on the roles of supervised ML in neuroscience is out! w/

    Poništi
  17. 5. sij 2019.

    In fact, a greater proportion of EMG was explained by a nonlinear gain than a linear mapping. This can help to explain how motor cortex controls the wide range of forces encountered in the real world. 3/3

    Prikaži ovu nit
    Poništi
  18. 5. sij 2019.

    In a single task, muscle EMGs can be predicted relatively well using a linear mapping from the activity of motor cortex neurons. However, when using multiple tasks that span a wide range of forces, we found that this mapping was highly nonlinear. 2/3

    Prikaži ovu nit
    Poništi
  19. 5. sij 2019.
    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    5. pro 2018.

    New preprint by brilliant in my group. A new state-space model with multi-scale hierarchical dynamics prior. Collaboration with and Monica Bugallo 1/6

    Prikaži ovu nit
    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:

    ·