Did a little rewilding work in Wicklow over the weekend

We are hungary for voting reform 😉

So bullish on the EU right now

Lots of exciting work happening in this space :)

That is a form of plagiarism

ESA ESA ESA!

What kind of use cases do you use satellite data for?

Day 18 of thesis write-up. My room has never been this clean before.

I don't have much experience with R, but I know a good package is IML: cran.r-project.org/web/packages...

However, this one focuses on permutation-based XAI methods, not gradient-based ones as you have suggested.

Not exactly. You don't need to integrate a method into the package. You can use a "post-hoc" method instead. These are separate methods applied to a model after it has been trained to explain it's predcitions.

A great package for neural networks (if you are using PyTorch) is captum: captum.ai

Day 17 of thesis write-up. Lost a bit of momentum as I was a bit sick the last few days. Getting back into it.

Is "thanks vibe coding" the new "thanks Obama"?

This looks like an amazing role. Any chance you could move the entire lab to Dublin, Ireland? 😅

We are hiring a Software Engineer who can help us turn our methods into easily usable tools for:

* Explainable AI
* Cell Tracking
* Interactive Learning

...and some more: careers.humantechnopole.it/job/Research...

Feel free to reach out directly if you are interested and/or have questions!

Will give this a read :) what kind of XAI methods did you explore?

Or the immediate destruction via a dual-vector foil

Also, XAI/ Explainable ML are different names for the same field. Some even call it interpretable machine learning (IML).

I see :) the architecture alone wouldn't be considered XAI. For all neural networks, you can find the derivatives of the ouput w.r.t. the input. Yet, raw gradients can be misleading, and we still need XAI methods like Grad-CAM or integrated gradients to help make sense of those derivatives.

Oh, great, the AI says we can eat this mushroom. Only a 4% chance that it will kill us 😅

Using AI to cure cancer or fight climate change makes as much sense as using neuroscience to figure out AI. It's all yak shaving. If your goal is to "solve X" just go solve X, don't try to solve Y then use Y to solve X.

To become an adjective, an author needs:
1) to be a great author
2) a surname that lends itself to an -ian suffix

SKY Perfect JSAT's 100-Satellite Plan: Japan's Largest Satellite Operator Bets Big on Earth Observation Asia's largest satellite operator SKY Perfect JSAT has unveiled plans to build a 100-satellite constellation combining Earth observation and communications. With a $230M investment in Planet Labs' Pel...

Like this? kantenna.com/topic/sky-pe...

This is pure escapism. It would be better to focus resources on fixing problems on Earth than trying to leave it. Although space exploration and earth observation can still be part of the solution.

Started reading this collection of short stories by H.P. Lovecraft

I marked projects for one of my supervisor’s courses. The best thing I did was put a bunch of papers I knew very well (including one of my own) in the recommended reading. It was immediately obvious if they used AI so summarise the papers.

Initial conditions:
m1=19.4 m2=41.6 m3=46.7 (solar masses)
v1x=-6.677 v1y=-6.772 v2x=-3.379 v2y=2.709 v3x=-5.869 v3y=4.405 (km/s)
x1=-3.0 y1=17.0 x2=-18.0 y2=21.0 x3=9.0 y3=2.0 (AU from center)
Music: Dune (2021) Medley – Zimmer

Artemis II mission captures first pictures of the far side of the Moon.

You should try warming them up in the microwave first

This angle is even more beautiful