Teemu Sarapisto

Teemu Sarapisto

@tsarpf.bsky.social

CS/ML PhD research in high-dimensional time-series @helsinki.fi Before: 7y of C++/JS/VR/AR/CV/ML at Varjo, Yle, Reaktor, Automattic ... After dark: modular synthesizers & electronic music

61 followers143 following30 posts

Teemu Sarapisto
@tsarpf.bsky.social

Infograph of my previous paper. Generated via Nano Banana Pro. These models are getting quite useful! I think it has genuinely good explanations. Maybe I would've opted for a bit less "AI does this and that", and there are some small mistakes, but not too much to distract from the main message.

Teemu Sarapisto
@tsarpf.bsky.social

Infograph of my previous paper. Generated via Nano Banana Pro. These models are getting quite useful! I think it has genuinely good explanations. Maybe I would've opted for a bit less "AI does this and that", and there are some small mistakes, but not too much to distract from the main message.

Teemu Sarapisto
@tsarpf.bsky.social

Just received the first volume, and damn, clearly a ton of effort was put into it! There's a ~90 page intro to "foundations of complexity science" (which is also sold separately). (The super interesting) papers each have a 5-10 page intro with historical context, and are full of annotations ❤️

Santa Fe Institute @sfiscience.bsky.social

Today, the SFI Press published Vol. 4 of Foundational Papers in Complexity Science. This concluding book follows Volumes 1 & 2 in May and Volume 3 in September & contains papers published between 1989 and 2000 — when complex-systems science had become a fledgling field of study in its own right.

Teemu Sarapisto
@tsarpf.bsky.social

Ok, heh, well, partially the reason for the \infty in there is because the simulator got stuck in a non-chaotic loop due to integration errors. Grabbing the samples before the looping just gives a more uniform distribution

Teemu Sarapisto
@tsarpf.bsky.social

Two very interesting papers on identifiably (provably) recovering "true"/unique latent states + dynamics / time-evolution from high-dimensional data in ICLR2025: 1. VAE + nonlinear ICA-style aux variables openreview.net/forum?id=d16... 2. Contrastive learning openreview.net/forum?id=ONf...

openreview.netSelf-supervised contrastive learning performs non-linear system...Self-supervised learning (SSL) approaches have brought tremendous success across many tasks and domains. It has been argued that these successes can be attributed to a link between SSL and...
Teemu Sarapisto
@tsarpf.bsky.social

Basic idea of #DeepSeek R1, the new RL LLM model: Start from a pretrained LLM, sample K responses for tasks (e.g., LeetCode problems with N tests). Update weights to favor better-than-average responses, as the true correct code is unknown. Also, reward correct output format. #LLM #genAI #RL

Teemu Sarapisto
@tsarpf.bsky.social

Interested to see if a proper academic bsky will happen! The amount of great papers and memes I found on Twitter made me stick to it this long, but definitely something new would be nice.

You are up to date.