deboraha

neuralreckoning, 11 days ago to science

Thought about hypothesis testing as an approach to doing science. Not sure if new, would be interested if it's already been discussed. Basically, hypothesis testing is inefficient because you can only get 1 bit of information per experiment at most.

In practice, much less on average. If the hypothesis is not rejected you get close to 0 bits, and if it is rejected it's not even 1 bit because there's a chance the experiment is wrong.

One way to think about this is error signals. In machine learning we do much better if we can have a gradient than just a correct/false signal. How do you design science to maximise the information content of the error signal?

In modelling I think you can partly do that by conducting detailed parameters sweeps and model comparisons. More generally, I think you want to maximise the gain in "understanding" the model behaviour, in some sense.

This is very different to using a model to fit existing data (0 bits per study) or make a prediction (at most 1 bit per model+experiment). I think it might be more compatible with thinking of modelling as conceptual play.

I feel like both experimentalists and modellers do this when given the freedom to do so, but when they impose a particular philosophy of hypothesis testing on each other (grant and publication review), this gets lost.

Incidentally this is also exactly the problem with our traditional publication system that only gives you 1 bit of information about a paper (that it was accepted), rather than giving a richer, open system of peer feedback.

#science #philosophy

deboraha, 12 days ago to random

Hey, has anyone here had experience examining a PhD thesis “by publication”? I’m doing an admin task around revisions to a thesis (in Australia we don’t have a thesis defence), and the student/supervisor team have dismissed any comments on chapters which were already published (there were many comments from 3 examiners). Is this common practice? #academicchatter #phd

jimbob, 13 days ago to random

Conferences that insist you use their Powerpoint slide template for your talk :binfire:

djnavarro, 1 month ago to random

Okay. The mere fact that I saw Mark Zuckerberg posting on fediverse today is a pretty good sign that it is only a matter of time before I end up having to see Libs of TikTok or Gays Against Groomers. I had quite enough of that from twitter. I am absolutely not letting that shit back into my life.

Preemptively blocking the entire threads.net domain.

deboraha, 3 months ago to psychology

OK Fedi pals, it’s that time of year again and I’m looking for published papers for my 4th year #psychology #research_methods students to pull apart. The articles need to have some major flaws but not TOO obvious or too extensively blogged about already. The students write a blog post rather than an essay so something they can have a bit of fun with is good. TIA, please #retoot for visibility! #openscience #questionableresearchpractices #highered

MattCrumpLab, 3 months ago to random

Leaving alone the general apartment noise of living in Brooklyn, micing a piano just doesn’t sound the same as sitting in front of one

jonny, 3 months ago to random

My cat just jumped up on my desk and dropped a 2023 5 Peso coin out of her mouth. She is an indoor cat, and I havent been to Mexico since I was a kid. Should I be scared or just glad she's finally trying to pay her way

jonny, 3 months ago to random

test post to get a new post ID past a certain time plz ignore

nixCraft, 7 months ago to random

Bash history https://bash.cyberciti.biz/guide/Recalling_command_history #unix #linux #bash

futurebird, 7 months ago to random

What kind of music comes to mind when you think about ants doing things? I need some good (no lyrics) “ant music” any suggestions?

Drmowinckels, 7 months ago to random

Quarto manuscript projects are going to be my new life saver! This is magic yall! Thank you @minecr for presenting this!

https://quarto.org/docs/manuscripts/

#PositConf2023 #QuartoPub

ct_bergstrom, 8 months ago to random

It’s quite odd. Despite a supposed crisis of overburdened peer reviewers, many editors of the elite journal seem unwilling to make even basic decisions on their own, without coming back to reviewers repeatedly.

This summer I have reviewed several papers where I have said something to the effect of “the authors have addressed my concerns and I encourage publication of the paper in its present form”, only to have the paper come back to me for a third or fourth time for further review.

Why?

deboraha, 8 months ago to random

Ok, weirdest thing, after my previous post I got a bunch of emails (to my gmail) from what look like bots, wanting to participate in our study. Has anyone else experienced this? All first.last123(at)gmail.com format addresses, all with bot-like phrasing, the first batch within minutes of each other. I included a link in my last post so why wouldn’t they just go via the link? What should I do in case any are genuine?

theconversationau, 8 months ago to random

DNA of the Y chromosome has been completely sequenced end-to-end, and it’s just as weird as we expected.

Will we finally be able to understand how it works?

@ProfJennyGraves from @Latrobe shares the findings here:

https://theconversation.com/the-weird-male-y-chromosome-has-finally-been-fully-sequenced-can-we-now-understand-how-it-works-and-how-it-evolved-212112?utm_medium=Social&utm_source=Twitter#Echobox=1692859946

elduvelle_neuro, 9 months ago to random

Welcome to all! ❤️​🐀​
Potential new followers: can we try something out?

• if you're a #Neuroscientist (of any status, yes, undergraduate students can be neuroscientists), could you answer with your neuro 'categories': Experimentalist, Computational, Philosopher, Fly, Human, Rodent, Whale, Tetrodes, Calcium imaging, fMRI, EEG, etc. This is so that I (and possibly others) can put you in a List! I'll share my lists once they have grown enough 😀​

• if you're not a Neuroscientist, this account will probably not follow you, but you are totally free to follow it, and also to tell us about you and any questions you might have about neuroscience!

• if you want to learn about neuroscience, I will try to post beginner-level explanations with the hashtag #NeuroForNewbies. Just an experiment...

futurebird, 9 months ago (edited 9 months ago) to random

You can tell this has happened with Roger before before.
You can tell it will happen again.

(via tiktok: https://www.tiktok.com/@svphoenixstudio/video/7249157372789296430)

theconversationau, 9 months ago to random

The Seine will be used for Olympic events in 2024 and remain open for public swimming, ending a century-long ban.

It’s part of a global movement to make city rivers healthy and swimmable again, write Gary Osmond (@UQ_news) + @rebeccajolive (@rmit). https://theconversation.com/olympic-swimming-in-the-seine-highlights-efforts-to-clean-up-city-rivers-worldwide-210714?utm_medium=Social&utm_source=Twitter#Echobox=1691036598-1

bartlettje, 9 months ago to random

Really interesting article on the darker history of Galton, Pearson, and Fisher. Provides some suggestions for how you can broach it in statistics curricula and link to modern problems like biases in AI.

Also links to subjective biases as I never knew how shitty some of their eugenics research was.

https://doi.org/10.1080/26939169.2023.2224407

AndyPerfors, 9 months ago to random

I don't understand how anybody can handle five straight days of conferencing, even if they are an extrovert.

I'm not an extrovert, but I'm definitely less introverted than some; I had a really great time; and I took breaks when I felt I needed to. Yet I'm utterly, utterly wiped out now.

(This isn't really a complaint, just an observation. Talking about science to awesome people is wonderful but so is silence 😊)

taylorlorenz, 9 months ago to random

Twitter was able to change the logo of their iOS app but not the name, since Apple requires app names to be at least 2 characters.

They’re also still referring to the app as Twitter in the update notes (via
@sferik)

Screenshot showing name has to be more than 2 letters

incrediblemelk, 9 months ago to FiberArts

After reading @pluralistic's latest piece on planned obsolescence as technofeudalism, I stumbled across this Tumblr post about how a rusty old 100-year-old Singer sewing machine can be restored to FEARSOME, SUPERIOR working order compared to any newish machine, by swapping in standard-sized parts that attach with standard flat-head screws

Meanwhile the modern machine's plastic parts wear out quickly and very little is attached with standard screws. It also depends on software that's now glitching

https://www.tumblr.com/viridianriver/721034993348067328/sewing-machines-planned-obsolescence-ive-got

#PlannedObsolescence #RightToRepair #sewing #SewingMachine #Antique #Vintage #AntiCapitalism

parismarx, 9 months ago to tech

In 2000, Elon Musk tried to rebrand PayPal as X.com, but was forced out by the board and replaced as CEO by Peter Thiel.

Now he’s doing the same to Twitter, but is surrounded by sycophants who can’t tell him it’s a bad idea.

Read more: https://www.disconnect.blog/p/elon-musk-wants-to-relive-his-start

#tech #twitter #x #elonmusk #socialmedia

Elon Musk tweeting the new X logo to replace Twitter.