@albertcardona@mathstodon.xyz
@albertcardona@mathstodon.xyz avatar

albertcardona

@albertcardona@mathstodon.xyz

How does the brain work? Someday, we'll figure it out.
Group Leader, MRC LMB, and Professor, University of Cambridge, UK.
#neuroscience #Drosophila #TrakEM2 #FijiSc #CATMAID #connectomics #connectome #vEM #iNaturalist #entomology
Born at 335 ppm.
Brains, signal processing, software and entomology: there will be bugs.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

albertcardona, to random
@albertcardona@mathstodon.xyz avatar

Nothing is forever.

"Once a vibrant platform for artists, #DeviantArt is now buckling under the weight of bots and greed—and spurning the creative community that made it great."

https://slate.com/technology/2024/05/deviantart-what-happened-ai-decline-lawsuit-stability.html

My account is 22 years old, not updated since 2010. Amazed I can still login. The whole site looks like it could have been a Mastodon-like network; it clearly was a social web, before that name even existed.

Now looks like a wasteland. All posts and comments have been automatically tagged as protected from "AI scrappers" – given how much of present-day galleries are just AI-generated, that's kind of funny.

OmaymaS, to ai
@OmaymaS@dair-community.social avatar

"What do you mean by progress when you talk about AI?" and progress for whom?

I asked the techno-optimist guy at an AI Hype Manel!

  • Does progress mean getting bigger or better models?

  • What about the impact on environment, water resources, destruction of communities, mining raw materials in Africa?

He first didn't get my Q. Then he said he believed in the "utilitarian view" & developing intelligence is very important.

Just parroting the AI hype people!

albertcardona,
@albertcardona@mathstodon.xyz avatar

@OmaymaS

An AI hype guy was telling me today how AI replaces thousands of apps. I asked him (it's always a he) how many applications did he use in his laptop: 6. Seem unfazed. I asked then what applications he found useful, and after some discussion, it came down to drawing up posters, websites, and the like, for advertising, overcoming his lack of skills in the art department. I.e., spam, without an expectation of engaging with the recipients. Broadcasting. Indeed chat bots and generative AI are a boon to the spam industry.

It's hard to speak to the convinced. Guess when ChatGPT starts charging lots per subscription they'll rethink their starry-eyed stance.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@OmaymaS

Indeed. And on that, any gains in efficiency of ANNs implementation or GPU tech are squandered by ever growing neural networks. The scale currently is horrendous; the electricity and water usage is outlandish. One wonders, was the choice of clean water or "AI", what would the fanboys choose then. At the moment they choose "AI" for themselves and water shortages for others.

tdverstynen, to random
@tdverstynen@neuromatch.social avatar

There are two paths to success in modern academic science.

  1. Be the gatekeeper for access to funds/resources and dictate conditions favorable to your public image.

  2. Do good science with as few resources as possible and get lucky.

Can you guess which is more common?

albertcardona,
@albertcardona@mathstodon.xyz avatar

@tdverstynen

I guess we could discuss what "success" means, and what "get lucky" means.

The first choice reads horrible, as in, as formulated, only an individual not actually interested in science but on the accolades would do that.

The second choice is, to me, academic research as usual. It's amazing how much a laptop, experience, ingenuity, and a network of friends and collaborators can accomplish, even without a budget. And the journey and the nurturing personal relationships are the reward; the end goal is always yet another metaphorical lush valley, jungle, or sea floor to explore.

albertcardona, to random
@albertcardona@mathstodon.xyz avatar

Not a mosquito but a Cecidomyiinae, likely of the genus Lasioptera sp: a kind of gall midget or gall gnat.

https://www.inaturalist.org/observations/81412856

From a lucky shot in focus back in 2021. Thanks to the community for the ID. With Cecidomyiidae being a family likely more speciose than even beetles [1], the chances of identifying this animal are slim.

[1] Hebert et al. 2016 https://royalsocietypublishing.org/doi/full/10.1098/rstb.2015.0333

xtaldave, to strucbio
@xtaldave@xtaldave.net avatar

:geordi_no: Alphafold3. Not FOSS. Not FAIR. Public version has important features redacted.

:geordi_ye: OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization | Nature Methods

Faster. FOSS. FAIR.

@strucbio

https://www.nature.com/articles/s41592-024-02272-z

albertcardona,
@albertcardona@mathstodon.xyz avatar

@xtaldave @strucbio

OpenFold is a big deal. And a great showcase for free open source software, enabling scientists to improve on each other's work in the best scientific tradition.

Source code: https://github.com/aqlaboratory/openfold

Open data for training: https://registry.opendata.aws/openfold/

lhgmk2, to macrophotography
@lhgmk2@mas.to avatar

Spiders can share too. #macrophotography

albertcardona,
@albertcardona@mathstodon.xyz avatar

@lhgmk2

That's remarkable.

albertcardona, to Java
@albertcardona@mathstodon.xyz avatar

In java, is there a simple, idiomatic, standard-library built-in way to iterate a Stream n elements at a time? E.g., in pairs, using a BiConsumer or similar two-argument Function?

albertcardona,
@albertcardona@mathstodon.xyz avatar

@PrehensileTrunksRUs

Thank you! Indeed Stream gatherers are exactly it – enabling via windowFixed and windowSliding to group consecutive stream elements.

aeryn_thrace, to random
@aeryn_thrace@mastodon.social avatar

Disquieting thought:

Accelerating scientific development depends on substantially slowing the rate of research paper publication.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@aeryn_thrace

Can only be seen as specious if someone authors more papers than is humanely possible to read and write in a year.

An average of 3 papers as senior author per year is already a lot of work.

peterdutoit, to random
@peterdutoit@mastodon.green avatar

"Key numbers on Porto Alegre & surrounding state per Folha de S.Paulo

  • 95 dead
  • 131 missing
  • 50k homeless
  • 160k evacuees
  • 17 hospitals closed
  • Airport closed indefinitely
  • 450k homes without power
  • 85% of city's 1.3 mln residents without working water supply"—Brain Winter
albertcardona,
@albertcardona@mathstodon.xyz avatar

@peterdutoit

It is said the future is already here, just not homogeneously distributed.

What happened in Porto Alegre will happen where we are now, sooner or later, in our lifetimes, repeatedly.

albertcardona, to uk Catalan
@albertcardona@mathstodon.xyz avatar

Auroras over Cambridge, UK. Surreal!

#aurora #UK

albertcardona, to random
@albertcardona@mathstodon.xyz avatar

Did somebody say ... nectar?

Harmonia axyridis, the invasive (in the UK) Asian ladybug.
https://www.inaturalist.org/observations/214878124

Daojoan, to random
@Daojoan@mastodon.social avatar

This is par for the course in what I call the Exploiter Economy:

Addiction-optimized platforms that strip away our free time and bombard us with ads. Products designed to extract maximum value from users while returning the bare minimum. And always, the creators get the short end of the stick while Big Tech laughs all the way to the bank.

https://joanwestenberg.com/blog/its-an-exploiter-economy-not-a-creator-economy

albertcardona,
@albertcardona@mathstodon.xyz avatar

@Daojoan

Well put. One should always perform due diligence before committing one's time and effort into a platform. If it isn't a non-profit, it won't end well.

albertcardona, (edited ) to academia
@albertcardona@mathstodon.xyz avatar

Dr. Núria Cortadellas and hear colleague Almudena Fernández taught me electron microscopy – so grateful for all the time they shared with me and the opportunities granted.

I am very happy that my alma matter, the University of Barcelona, named a lecture hall in her honour (in 2022! How did I miss that?):

http://www.ccit.ub.edu/EN/2022/new20220325.html

albertcardona, to random
@albertcardona@mathstodon.xyz avatar

Online course on "Scientific Image Editing and Figure Creation" using open source software and .

By BioVoxxel via Zoom, on:
Thu 27 Jun 2024 09:00 - Fri 28 Jun 2024 15:30 CEST

Register at: https://www.tickettailor.com/events/biovoxxel/1198561

Details: https://www.biovoxxel.de/workshops/scientific-image-editing-and-figure-creation/

steveroyle, to random
@steveroyle@biologists.social avatar

It seems I have found a new way to procrastinate on a grant application. Using #Rstats to make my Gantt chart!

albertcardona,
@albertcardona@mathstodon.xyz avatar

@steveroyle

I recently got reviews back for a grant application where all my figures were left out by accident. The reviews were very good, and none mentioned the lack of figures or Gantt chart.

As a reviewer myself, I have never paid any attention whatsoever to the Gantt chart when one was included, as I consider it at best wishful thinking. Seems that many others are of the same mind. Just saying :)

albertcardona,
@albertcardona@mathstodon.xyz avatar

@steveroyle

Nope. To be fair I was referring to them in the text as the preliminary data – there was just one figure –, not with numbers or panel lettering.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@steveroyle

I am so over checkbox ticking exercises. I'd rather support a scientist writing an honest grant than one trying to dazzle me with what ultimately amounts to busywork.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@steveroyle

😀 Plans and meeting the enemy and all that.

albertcardona, (edited ) to Neuroscience
@albertcardona@mathstodon.xyz avatar

"A petavoxel fragment of human cerebral cortex reconstructed at nanoscale resolution" by Shapson-Coe et al. 2024 (Lichtman lab).

The reconstruction at its current state is already useful and very interesting. Here is to hoping the authors will put in more time and resources to further polish it.

Paper: https://www.science.org/doi/10.1126/science.adk4858

Preprint (2021): https://www.biorxiv.org/content/10.1101/2021.05.29.446289.abstract

Browsable data: https://h01-release.storage.googleapis.com/landing.html

Viren Jain's (Google) press release: https://research.google/blog/ten-years-of-neuroscience-at-google-yields-maps-of-human-brain/

#neuroscience #connectomics

albertcardona,
@albertcardona@mathstodon.xyz avatar

A remarkable finding from Shapson-Coe et al. 2024 paper on human brain #connectomics: the presence of canalized connections in the human brain cortex. Canalized in the Kauffman boolean networks sense [1], which here means: among the many synaptic inputs that any one neuron integrates, some are far stronger (by number of synapses) than the rest.

This is a pattern that we described in the #Drosophila larval nervous system (Ohyama et al. 2015 https://www.nature.com/articles/nature14297 ) and that has been reported as well for the mouse hippocampus (Bartol et al. 2015 https://elifesciences.org/articles/10778 ) and cerebellum (Nguyen et al. 2023 https://www.nature.com/articles/s41586-022-05471-w ).

[1] Canalisation as a term was introduced by Waddington in 1942 in the context of genetics to mean "some phenotypic traits are very robust to small perturbations" https://en.wikipedia.org/wiki/Canalisation_(genetics)

#neuroscience #connectomics

kittylyst, to random
@kittylyst@mastodon.social avatar

As you might know, I have no academic training in computing of any kind. My training was in Mathematics and Theoretical Physics.

It is, therefore, often a surprise when I encounter a CS paper (often a vintage one) and realize just how beautifully it is written and what it expresses.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@kittylyst

Second that. Happened more times than I can recall. Wagner and Fischer’s edit distance paper stands up there in the pantheon of clearly and eloquently written papers:

“The string-to-string correction problem”, 1974 https://dl.acm.org/doi/pdf/10.1145/321796.321811

albertcardona, to Trains
@albertcardona@mathstodon.xyz avatar

In Europe, flying is cheaper than taking the train.

It's an embarrassment, and a major problem: we have to stop flying for silly short distances. Realise that the overheads of flying (reaching the airport, awaiting 2 hours, the flight, the unloading, reaching the destination) largely cancel out any time gains of flying. And the carbon costs are utterly untenable. Not to speak of the modern, dire conditions of the whole flying "experience".

Another embarrassment is that train connections can't be guaranteed when across countries or companies. They aren't even coordinated. As if those who commission and set the schedules didn't travel by train themselves, at least not internationally. In considering how tiny most European countries are, it's frankly bizarre.

There are so many destinations one could travel by train to, yet in practice, it's not sensible. A disgrace.

The upside is that it can be fixed.

albertcardona,
@albertcardona@mathstodon.xyz avatar

@paulthenerd

My rule of thumb would be, if there's a train route that can make the journey in less than 8 hours, there's no reason to fly at all. Flying will take about as much time when including the overheads, and most hours will be wasted rather than comfortably sitting on a train.

Distance is not a factor; time is. E.g., London-Bergen, in Norway, would be quite difficult to replace by train or boat. Whereas London-Barcelona, which is about the same distance, is easy: there already is a route with Eurostar + TGV via Paris.

London-Bergen: 1044 km https://www.distance.to/London/Bergen

London-Barcelona: 1138 km https://www.distance.to/London/Barcelona

albertcardona,
@albertcardona@mathstodon.xyz avatar

@brembs

Sorry to hear. Back in the late 2000s I travelled quite a bit across Germany by train and the experience was always delightful.

What you describe is too close to the current state of British railway, which is so unreliable that, well, one can't rely on it for anything time-sensitive. It is only infrastructure if the system operates so smoothly one grows to rely on it completely. The UK's railway is more akin to a fair joyride: something one goes into for the thrill of it, knowingly.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • provamag3
  • kavyap
  • DreamBathrooms
  • tacticalgear
  • magazineikmin
  • vwfavf
  • Youngstown
  • ngwrru68w68
  • ethstaker
  • slotface
  • rosin
  • mdbf
  • thenastyranch
  • PowerRangers
  • anitta
  • modclub
  • Durango
  • cubers
  • osvaldo12
  • GTA5RPClips
  • everett
  • khanakhh
  • InstantRegret
  • Leos
  • tester
  • normalnudes
  • cisconetworking
  • megavids
  • All magazines
    •