mlpro

7 ms·

1.
by mlpro 8mo ago
Novel Ideas are never cheap, lol.
2.
by mlpro 9mo ago
Lol. trying to copy the Universal Weight Subspace paper's naming to get famous.
3.
by mlpro 9mo ago
Lol, yeah.
4.
by mlpro 9mo ago
Oh, look - a new 3D model with a new idea - more data.
5.
by mlpro 9mo ago
I don't understand.
6.
by mlpro 9mo ago
Waymo should do a bit more research in reliability and explainability of their AI models.
7.
by mlpro 9mo ago
Read the paper end to end today. I think its the most outrageous ideas of 2025 - at least amongst the papers I've read. So counterintuitive initially and yet so intuitive. Personally, kinda hate the implications. But, a paper like this
8.
by mlpro 9mo ago
They are not trained on the same data. Even a skim of the paper shows very disjoint data. The LLMs are finetuned on very disjoint data. I checked some are on Chinese and other are for Math. The pretrained model provides a good initializatio
9.
by mlpro 9mo ago
I think its very surprising, although I would like the paper to show more experiments (they already have a lot, i know). The ViT models are never really trained from scratch - they are always finetuned as they require large amounts of data
10.
by mlpro 9mo ago
Why would they be similar if they are trained on very different data? Also, trained from scratch models are also analyzed, imo.
11.
by mlpro 9mo ago
It's about weights/parameters, not representations.
12.
by mlpro 9mo ago
The analysis is on image classification, LLMs, Diffusion models, etc.
13.
by mlpro 9mo ago
It does seem to be working for novel tasks.
14.
by mlpro 9mo ago
Not really. If the models are trained on different dataset - like one ViT trained on satellite images and another on medical X-rays - one would expect their parameters, which were randomly initialized to be completely different or even orth