minimalparts

minimalparts, 10 days ago to random

In the long term, #degrowth does not mean lower algorithmic quality. It means algorithmic diversity. Machine learning is much more than the one architecture that is currently being sold as ‘intelligence’. Machine learning is wonderful. But it can and should be small, energy-efficient and planet-friendly.

minimalparts, 1 month ago to random

For people interested in federated Web #search: we are preparing for the release of a few test instances of PeARS. We have added a copy of the instances’ FAQ on our website so that people can get a preview of what it all looks like. Feel free to check it out at https://pearsproject.org/federated-faq/.

emilymbender, 1 month ago to random

Bender & Shah 2024 is now officially published at TWeb.

For everyone who is using or promoting LLM driven chatbots as information access systems: this one is for you

https://dl.acm.org/doi/10.1145/3649468

minimalparts, 1 month ago to ai

These last two posts by @emilymbender highlight how #AI #hype has conquered Higher Ed: https://buttondown.email/maiht3k/archive/, and I couldn’t agree more.

I find it properly disconcerting that many academics fell for the hype when clearly, it is purely the result of a very good lobbying effort on the part of Big Tech. So just to get things straight, here’s my personal experience of what happened in the field of Natural Language Processing (#NLP), starting back in 2017 (references at the end of the thread) 1/6

PhieLaidMignon, 1 month ago (edited 1 month ago) to random French

Toujours en 2024, à part Yacy, qui est très insatisfaisant dans le tri des résultats, je ne trouve AUCUN projet de moteur de recherche open source et décentralisé (j'ai dit moteur, pas metamoteur)

edit:

vous allez rire mais j'ai trouvé un moteur de recherche plus performant sur cette requête que google et duckduckgo. C'est... Mastodon.

En cherchant Search Engine sur Mastodon, je suis tombé sur https://github.com/StractOrg/stract

ça semble prometteur comme moteur de recherche opensource. à voir comment le crawling est effectué

minimalparts, 2 months ago to random

Meanwhile, we are cracking on with the federated search code... Sneak peak of one of our first prototype instances 🙂

Another screenshot of a prototype for a federated Web search engine. This shows the results page for the query 'water consumption LLMs'.

minimalparts, 4 months ago to ai

"#AI is going to change everything," they say. True. So much so that I don't recognise myself anymore in search results. Here is the Google snippet for my German LinkedIn page. Sounds good, except that I am not a neurobiologist, nor a translator, that I do not live in Boston and have never written a fantasy novel.

I feel very achieving, though.

minimalparts, 9 months ago to random

What is the meaning of the word 'red'? How do small children learn it? What about AI systems? In a new post full of insights, Casey Kennington explains how natural languages are intrinsically linked to the real world, and how humans naturally ground symbols in their physical environment... in stark contrast with LLMs. To find out how current AI systems differ from human cognition, head over to Denotation IO and read 'Meaning and the world: the symbol grounding problem':

https://denotation.io/blog/meaning-and-world-symbol-grounding-problem/

minimalparts, 1 year ago to random

This week on Denotation IO, 'Infinity is not Everything'.

https://denotation.io/blog/infinity-is-not-everything/

Looking at the linguistic and logical processes behind creativity, and asking to what extent they are implemented by neural net architectures.

TLDR; ⬇️ /1

minimalparts, 1 year ago to ai

#introduction 🙂

I've been a researcher in the field of #computationallinguistics for over 15 years. Interested in #semantics, #cognitivescience, energy-efficient and ethical #AI systems. I've worked at the universities of Cambridge, Potsdam, Stuttgart and Trento, with a brief stint at Pompeu Fabra in Barcelona.

I am one of the core developers of PeARS (https://pearsproject.org/), a slow but friendly open-source project aiming at developing a fully private, decentralised Web search engine.