Lukas Mosser (@porestar) header
Lukas Mosser (@porestar) profile picture

Geoscience + Machine Learning + Generative AI | Bouldering | Photography

Stavanger, Norway
Joined October 2007
Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)Photo by Lukas Mosser (@porestar)
Lukas Mosser retweeted
Here’s to the wanting. Here’s to the dreaming. Here’s to everyone.
808
1,209
544
14,091
5,989,027
Lukas Mosser retweeted
Google have, dare I say... RELEASED A GOOD MODEL 🥹 Gemini 3.8 Flash provides ~Opus 5 performance at much lower cost, while being super fast
199
200
96
4,418
419,165
This is one of the craziest ideas I've ever seen. He converted a drawing of a bird into a spectrogram (PNG -> Soundwave) then played it to a Starling who sung it back reproducing the PNG. Using the birds brain as a hard drive with 2mbps read write speed. youtube.com/watch?si=HMtVdHB…
346
3,347
902
34,395
3,309,440
Lukas Mosser retweeted
the LLMs have reached boyfriend levels of intelligence
50
404
28
13,346
464,936
Lukas Mosser retweeted
New Feature: Deck Explorer 🔮 This lets you load a tarot deck from a gist, explore it and render major arcana, minor arcana, or a full PDF of everything for printing on hard paper. App: lukasmosser--tarotgpt-run.mo… Example Deck PDF for Printing: drive.google.com/file/d/1JKn…
1
4
264
Can we use Fourier series to model univariate PDFs? (esp. useful in neural compression) We introduce a lightweight, flexible, and end-to-end trainable Fourier basis density model. w/ @saurasingh Johannes Ballé Paper: arxiv.org/abs/2402.15345 Code: github.com/google/codex
1
6
15
1,304
📢Job alert – come and join ESE! We're seeking two lecturers/senior lecturers in #AI, #MachineLearning, and #DataScience. ⏰Deadline: 19 January More info on how to apply⬇️ imperial.ac.uk/jobs/descript…
2
2
728
Replying to @deliprao
p(x_t | x_0, ..., x_{t-1})
1
2
1
21
2,196
I will be hiring through #ELLIS this year: 3 fully-funded PhD positions for troublemakers in #ML #AI who want to design the next gen of #probabilistic #models and #programs that are provably #reliable and #efficient Join april-tools.github.io @InfAtEd Email me! Please share!
3
29
3
94
34,061
Good news: the European Summer School on AI has opened up virtual participation. No need to be in Slovenia next week: essai.si It’s your chance to still register to listen to me talk about “ML beyond static datasets” for 5 days & lots of other interesting courses!
6
23
2,722
Lukas Mosser retweeted
Everything you ever wanted to know about Self-Supervised Learning but were afraid to ask. A giant cookbook of SSL recipes. By a large crowd from Meta-FAIR with various academic collaborators led by @randall_balestr and Mark Ibrahim. arxiv.org/abs/2304.12210
36
633
37
2,899
494,008
It's a Thursday, and a variety of Twitter accounts with GAN-generated faces have paid $8 for a blue checkmark. #BlueCheckShenaniGANS (GAN = "generative adversarial network", the AI technology used by tools like thispersondoesnotexist.com) cc: @ZellaQuixote
97
1,344
377
3,119
How your email finds me
244
22,947
1,093
156,415
📜🚨📜🚨 NN loss landscapes are full of permutation symmetries, ie. swap any 2 units in a hidden layer. What does this mean for SGD? Is this practically useful? For the past 5 yrs these Qs have fascinated me. Today, I am ready to announce "Git Re-Basin"! arxiv.org/abs/2209.04836
59
558
130
2,670
Lukas Mosser retweeted
#IAMG2023 will be in Trondheim, Norway. The conference will not get any better than the contributions from you. Therefore, we invite all of you to suggest scientific sessions and short courses. Deadline for proposals is Oct 31st 2022. iamgconferences.org/iamg2023
4
6
new fan-made NeurIPS 2022 trailer:
50
498
126
2,967
What exactly is ML monitoring? We've published a new entry-level blog on all things it can comprise. When you deploy an ML system in production, it integrates into the business. It's ROI time! To track its performance, you need to look at more than one thing 👇🏼 (1/n)
1
10
1
42
ArXiv arxiv.org/abs/2207.05209: Fourier neural operator for PDEs on arbitrary geometries. Learns to map the irregular input grid to a regular latent space grid, where the FFT is applied. 10^5 times faster than numerical solvers and 2x as accurate as existing ML-based PDE solvers.
3
33
2
180
Lukas Mosser retweeted
🔋THREAD🔌 The inconvenient truth about climate change is that solving it will involve digging, blasting & leaching more minerals from the skin of this planet than ever before. No one much likes to talk abt this. But talk about it we must. ⛏👇
474
2,532
760
7,254
In applied ML, you want to use the most accurate test method possible -- carefully crafting (and regularly updating) a separate test as close as possible to the prod data, and entirely separate from the validation sets used to tune the model
3
19
2
388