Soroosh Afyouni

Views very obviously mine! Previously @Cambridge_Uni, @bdi_oxford and @IASWarwick

[Emad @EMostaque]

If Claude is good enough for Nobel Prize winners it is good enough for you

https://t.co/P5pM4MGSqQ

[Soroosh Afyouni @sorafy]

The usefulness of useless knowledge: How the balance between public/private R&D investment has shifted, what this means for high-risk scientific inquiry and how private capital help rebalance the scientific portfolio.
https://t.co/JrHX3t8haY

[Signal Processing @DSP_fact]

A FFT is a linear transform, so it can be represented by matrix multiplication.

[Eiko Fried @EikoFried]

Two days ago, a lawfirm filed a federal antitrust lawsuit against 6 commercial publishers (incl Elsevier & Wiley) in the federal district court in New York.

They allege a 3-part scheme on part of publishers. 🧵

https://t.co/BswKVg5l5G

[Onno Kampman @KampmanOnno]

Our work on modeling time-varying functional connectivity (TVFC) with Wishart processes is out now in Imaging Neuroscience! @Joe_Ziminski @sorafy @markvanderwilk @CambridgeABL

[Onno Kampman @KampmanOnno]

We have made the code for TVFC estimation available in a pip-installable Python package. This includes not only the Wishart process, but also MGARCH and sliding windows with cross-validated window lengths.

https://t.co/S0KijvtemU

[Marvin Schmitt 👉🦋 @MarvinSchmittML]

Search Google Scholar for "casual inference" (not causal inference).

[Benjamin Sun @ben_b_sun]

It has been a joy working with Brad Gibson and @ksuhre on this perspective piece, now online in @molcellprot, glimpsing into the “Promises and challenges of populational proteomics in health and disease”.
#proteomics #genomics

https://t.co/y1Ffj0H2cW

[Gabriel Peyré @gabrielpeyre]

The stochastic block model is the simplest random graph with communities having high interconnectivities. Erdős-Rényi model corresponds to a single community. https://t.co/7rNpRk2GXl https://t.co/FVlEaT1E6S https://t.co/99LyoAuQ5C

[François Fleuret @francoisfleuret]

Wat

[Soroosh Afyouni @sorafy]

Asking for a friend: what must-read do you recommend (review paper, seminal work, reading list etc.) for staying updated in (statistical) neuroimaging and/or connectome for someone who's been away from the field for the past 2/3 years?

RT highly appreciated!

[Peyman Milanfar @docmilanfar]

what happened to wavelets

[Soroosh Afyouni @sorafy]

@chrisalbon Get a reMarkable!

[Daniël Lakens @lakens]

Psychologists have tried using Bayesian methods to test hypotheses for a decade. I honestly believe it is time to give up on it. There are too many problems and no benefits compared to frequentist methods. This is not even a hot take. You just need to not be biased.

[Christoph Molnar 🦋 christophmolnar.bsky.social @ChristophMolnar]

Neural networks with only linear layers would result in linear models. Because a linear combination of linear functions is again linear. Nothing would be gained by adding more layers.

[Soroosh Afyouni @sorafy]

@RezvanFarahi scientific papers?

[Segun Fatumo @SFatumo]

5 years is a VERY LONG time for a paper to be under review.

That was the case for our recently published paper in Nature.

Submitted in 2018 and published in 2023

What do you do as an early career researcher while waiting for a very important paper to be published ?

1/4

RT

[Soroosh Afyouni @sorafy]

The Matrix Cookbook having only 3K citation is the biggest scientific injustice!

[Sabina Leonelli @SabinaLeonelli]

My new short book “Philosophy of Open Science” is now published and can be downloaded for free here: https://t.co/cKBMwgFXde #openscience #hps #sts #Egenis @UniExeterIDSAI @UofE_Research @turinginst @StudiesOpen @AdaLovelaceInst @ERC_Research