Axel Mosig

youtu.be/lT6WU-1S96U?...

13.01.2026 18:40 — 👍 0    🔁 0    💬 0    📌 0

Zen Browser Beautifully designed, privacy-focused, and packed with features.

zen-browser.app

13.01.2026 18:26 — 👍 0    🔁 0    💬 1    📌 0

😅 Switching to Zen browser (vertical tabs+different contexts) was a life changing experience for me and reduced my open tab count by at least 80%

13.01.2026 18:26 — 👍 1    🔁 0    💬 2    📌 0

Everyone should watch this

06.10.2025 12:47 — 👍 15    🔁 8    💬 2    📌 0

In this sense, "putting on the evolutionary pair of glasses" is nothing else but to account for the inevitable theory-ladenness of observation. There seems in fact no good reason to assume why machine learning should be exempt from theory-ladenness, which is one way to read the NFL theorems.

14.09.2025 13:50 — 👍 0    🔁 0    💬 0    📌 0

A #philscy perspective on our work: As machine learning is underdetermined (following the no-free-lunch theorems), the training process has little chance to infer an internal representation that is phylogenetically plausible - unless we introduce explicit inductive bias.

14.09.2025 13:45 — 👍 0    🔁 0    💬 1    📌 0

...in particular, we can consider the splits diagrams of all quartets of taxa and train the neural network so that the splits diagrams for all quartets degenerate to trees.

14.09.2025 13:40 — 👍 0    🔁 0    💬 0    📌 0

If we pass a (knwon and well-established) phylogenetic tree to the training process of the neural network, we can utilize mathematical structures behind evolutionary distances...

14.09.2025 13:39 — 👍 0    🔁 0    💬 1    📌 0

As has been observed by other authors before, the internal representations of neural networks are usually incompatible with known phylogenies. So, how could a neural network learn to infer features that are evolutionarily plausible?

14.09.2025 13:37 — 👍 1    🔁 0    💬 1    📌 0

A quartet-based approach for inferring phylogenetically informative features from genomic and phenomic data Neural networks are widely used in bioinformatics to extract features from morphological, structural, and sequence data of different taxa. A key quest…

Link to the full paper: www.sciencedirect.com/science/arti...

14.09.2025 13:32 — 👍 0    🔁 0    💬 0    📌 0

Equipping Artificial Intelligence with the Lense of Evolution Random patterns have so far obscured AI’s view of evolutionary relationships. A team in Bochum is teaching them how to go about this.

We found AI to be a bit short-sighted towards evolution, so we went out to grind some computational lenses. The resulting pair of evolutionary glasses is finally available in the latest issue of CSBJ

news.rub.de/english/pres...

14.09.2025 13:30 — 👍 0    🔁 0    💬 3    📌 0

Sieht sehr böhmisch aus!

14.09.2025 13:23 — 👍 1    🔁 0    💬 1    📌 0

Bhattacharya is the New Lysenko Starving the NIH through ineptitude, bitterness, and thirst for revenge

New post about Lysenko and Bhattacharya from @angierasmussen.bsky.social

open.substack.com/pub/rasmusse...

18.08.2025 17:11 — 👍 239    🔁 89    💬 10    📌 8

In Henk de Regt’s book on Understanding Scientific Understanding, skill takes center stage for understanding as a primary goal of science.

13.08.2025 19:25 — 👍 2    🔁 0    💬 1    📌 0

We are witnessing the destruction of science in America | Paul Darren Bieniasz If we stay on this administration’s course, future life-saving medicines may never be invented

"Science in the US is entering a phase resembling the Lysenko era in the Soviet Union."
www.theguardian.com/commentisfre...

10.04.2025 14:05 — 👍 487    🔁 220    💬 14    📌 13

Almost like it was written for the world these days... the "Final Speech" by Charly Chaplin's hairdresser and dictator-by-coincidence Hynkel. youtu.be/J7GY1Xg6X20?...

16.03.2025 17:02 — 👍 0    🔁 0    💬 0    📌 0

Large Language Model is Secretly a Protein Sequence Optimizer

Figure 1

Figure 2

Figure 3

Large Language Model is Secretly a Protein Sequence Optimizer [new]
LLMs, surprisingly, can optimize protein sequences via directed evolution, even without explicit protein training data, finding high fitness variants.

17.01.2025 04:40 — 👍 1    🔁 1    💬 0    📌 1

I donated. It would be great to have a social-media protocol that wasn't subject to capture, and it does seem possible.

freeourfeeds.com

14.01.2025 22:45 — 👍 124    🔁 27    💬 4    📌 1

…and most importantly: Calculators could actually *do* arithmetic, and did not imitate the process of doing arithmetic.

12.01.2025 14:14 — 👍 0    🔁 0    💬 0    📌 0

Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: A cross-sectional study Eric Oermann and colleagues ask whether a DL-based model for pneumonia detection performs well in external validation and consider the effects of hospital system–specific biases.

There is also the "classical" paper by Zech et al. which identifies potential shortcuts in pneumonia detection (see e.g. Fig. 2)

journals.plos.org/plosmedicine...

05.01.2025 20:23 — 👍 1    🔁 0    💬 0    📌 0

The risk of shortcutting in deep learning algorithms for medical imaging research - Scientific Reports Scientific Reports - The risk of shortcutting in deep learning algorithms for medical imaging research

AI predicts beer consumption from knee x-rays www.nature.com/articles/s41...

05.01.2025 18:52 — 👍 1    🔁 0    💬 1    📌 0

"Researchers trained medical AI models on over 25,000 knee X-rays. They then essentially worked backwards, tasking the deep learning programs to find commonalities that predicted nonsensical traits, such as which knee-owners clearly drank beer or ate refried beans—which is patently absurd."

12.12.2024 18:50 — 👍 120    🔁 49    💬 6    📌 8

cdn.bsky.app/img/feed_thu...

Here's what we've been missing in the meantime. Better tell the owner of the "Hammer of Justice" not to hit his own thumb.

19.11.2024 10:47 — 👍 0    🔁 0    💬 0    📌 0

Meanwhile, in the less free corners of the internet…

18.11.2024 17:00 — 👍 1    🔁 0    💬 0    📌 0