Sami Hamdan

PhD candidate in applied machine learning. Data Science, Open Science/Source, Neuroscience

[Sami Hamdan @SaHamPy]

Created a dashboard to explore KFold Cross-Validation using sklearn, @bokeh, @HoloViews and Panel. What happens if you increase n of folds, sample size or the complexity of the used model? Find it out by running it yourself: https://t.co/x54EeJog6s

[Rodrigo 🐍🚀 @mathsppblog]

Python 🐍 is a really elegant language.

And yet, it can also give birth to monstrosities like the one below.

I recommend you take a sit before taking a look at the code...

The program below prints `"Hello, world!"` in Python 3.9+!

[Sami Hamdan @SaHamPy]

For context: these circles were introduced to help ppl keep their distance from other ppl while COVID. Source in German: https://t.co/dqKLxzpiqN

[Sami Hamdan @SaHamPy]

Suspicious circles in Düsseldorf Germany. Could it be a sign of a recent alien visit? @ProbsNotAliens

[Sami Hamdan @SaHamPy]

@NirSSB @ProbsNotAliens Also wirklich ein Alien 🤔

[Sami Hamdan @SaHamPy]

@ravikantkolluru @dontusethiscode Is there a recording of the talk ? Sry cannot find it and sounds interesting.

[Maarten van Smeden @MaartenvSmeden]

Good description of leakage, but very often an even better solution is not to split your data into train-test sets at all. Cross validation and internal-external validation should be the starting point (deviate only when necessary)

[Sami Hamdan @SaHamPy]

@willmcgugan @ryrobyrne This as repl + good matplotlib support would be a total game changer for my nvim workflow

[Software Freedom Conservancy @conservancy]

Today is the day to #GiveUpGitHub

[Simon Eickhoff @INM7_ISN]

Shammi More presenting a tour-de-force on evaluating BrainAge approaches @OHBM

From CV performance to generalization, reliability and pathology

Workflow choices have a substantial influence and may impact downstream analyses.

Visit posters 39 & 81 to learn more

[elvis @omarsar0]

This looks like a very useful tool to generate diagrams with Python code.

Would love to see support for creating ML related diagrams, like neural networks.

https://t.co/Sk7y8F1FJp

[Simon Eickhoff @INM7_ISN]

For an innovative bridge between histology and function and how they shape hemispheric asymmetry, visit poster 741

Bin Wan presents great new work involving the @BigBrainProject

[Simon Eickhoff @INM7_ISN]

Differences between women 👩‍💼and men 👨 and how they are they modulated menstrual cycle 📆

Check out poster 738 and talk to Svenja Küchenhoff about the ‚little differences‘

[Simon Eickhoff @INM7_ISN]

Little brain, little critters

If you are interested in the cerebellum, poster 772 is a must-see. Neville Margliese investigated cerebellum vs cerebrum (co-) evolution

[Simon Eickhoff @INM7_ISN]

Lots of great work on BrainAge here in Glasgow!

But are these methods actually stable? How much do pipeline parameters matter?

If you dare to find out, visit Shammi More at poster 39

[Simon Eickhoff @INM7_ISN]

Is this right? Find out at poster 697

Patrick Friedrich showing fascinating data on hemispheric asymmetry in humans 🧑‍🦱and chimps 🐒

[Simon Eickhoff @INM7_ISN]

At poster 81 Sami Hamdan presents JuLearn, a user-friendly environment for machine-learning on neuroimaging. Particularly interesting for those getting into the field!

Sami will be a person to discuss confounds in individual prediction, too.

[Simon Eickhoff @INM7_ISN]

🔥 Lots of exciting posters by our wonderful students up @OHBM #OHBM2022 today 🔥

Is TFCE thresholding a good alternative to cluster-level inference in meta-analyses? Be sure to meet Lennart at poster 488 and chat with him on new developments in ALE

More posters ⤵️

[Sami Hamdan @SaHamPy]

Today, I will present julearn an easy to use #MachineLearning library from @fzj_inm7 at @ohbm #OHBM2022 especially usefully for #neuroscience. You can find me there at 12:45-01:45 PM Poster: WTh081.
We also provide a new YouTube Tutorial: https://t.co/FKuhQYMNAO

[Sami Hamdan @SaHamPy]

I hate that "the brain" and some neurotransmitters like dopamine became these buzzword for pseudo science and bad science communication.