piyush पीयूष ପୀୟୂଷ

PhD student, percussionist

[bioRxiv Cell Biology @biorxiv_cellbio]

Scaffolded hiPSC liver organoids recapitulating bile duct tubulogenesis and periportal architecture. https://t.co/t2usW0cSXG #biorxiv_cellbio

[Quanta Magazine @QuantaMagazine]

Large language models work well because they’re so large. But this power comes at a cost.
https://t.co/TMwgV7x6te

[CNRS Provence et Corse @CNRS_dr12]

#TalentsCNRS 🥉 | Découvrez la vidéo de Charlotte Perrin qui est lauréate de la médaille de bronze du @CNRS en 2024 !
@I2mMarseille @CNRS_INSMI @univamu @Centrale_Med
https://t.co/V1TphRo02Y

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

chomsky could never lmao /s

[madison @dearmadisonblue]

I'm still blown away by formal language theory basically being invented in 500BCE India. Panini created a generative grammar for Sanskrit with a formal metalanguage (embedded in Sanskrit!), even introducing the concept of bound variables and substitution

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

So the Bayesian approach is great for the actual smoothing, but transformers are remarkable for pruning the hypothesis-set. Can we hybridise to use the best of both worlds? Stay tuned :)

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

This year, we published my first PhD paper on the usability of transformers for tracking problems @eusipco! Find the open access version here and/or continue reading the thread :) @PhilippeRoudot
https://t.co/lcwfqC74uw

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

We thus see the emergence of two regimes, one where we have a lower no. of hypotheses (where the Bayesian approach is unmatched) and another with a higher no. of hypotheses (where transformers take the lead).

[Sauers @Sauers_]

The total compute cost was around $1,600,250, more than the entire prize

[Neil Renic @NC_Renic]

Writing Editing

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

Officially joined @hacktoberfest 2024 after years. Looking forward to one month of open source!
https://t.co/7httxrS6t6

[allen institute @AllenInstitute]

Tight junction protein ZO-1 is our #CellLineOfTheWeek!

Our human iPS cell lines & plasmids – fluorescently-tagged to illuminate structures – are openly available to accelerate discovery: https://t.co/QCetZaSS0V

📦 Distributed by @Coriell_Science & @Addgene

#OpenScience

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

@palash2504 yeah i get it... here's hoping you get to code soon!

[piyush पीयूष ପୀୟୂଷ @peeyoushh]

@palash2504 oh sorry to hear that. hope you're working on having proper boundaries in place :/ all the best