Philip Gillißen · @guerda
202 followers · 2242 posts · Server ruhr.social

@tante haven't embeddings been the core of NLP breakthroughs since ? I mean we discover racist embeddings already back then. I feel that LLM (and transformers per se) are just using more context to utilize embeddings

#word2vec

Last updated 1 year ago

Hacker News · @ycombinator
30 followers · 2464 posts · Server rss-mstdn.studiofreesia.com
celery · @celrydoor
25 followers · 318 posts · Server lingo.lol

I found #409 in 296 guesses!
🟩🟩🟩🟩🟩🟩🟩🟩🟩🟩🟧🟧🟧🟧🟧🟧🟧🟧🟥🟥
cemantle.certitudes.org/pedant
Also, the site has some really cool info on its algorithm. I don't understand it yet but it uses the algorithm to calculate the distance between two words. I wonder if that's also what does.

#pedantle #word2vec #Semantle

Last updated 1 year ago

ai-jobs.net · @aijobs
41 followers · 169 posts · Server mstdn.social
Frederik Elwert · @felwert
834 followers · 1938 posts · Server mstdn.social

The article starts with two theoretical issues: First, why to people have such a strong attachment to religious architecture, and second, what role play emoji, in addition to hashtags, in communicating emotions via social media during singular events.

Methodologically, we used , emoji usage statistics, and to study how emoji were used and how this shifted over time. Here’s what we found: 3/

#word2vec #topicmodelling

Last updated 1 year ago

ITHCWY · @ithcwy
33 followers · 243 posts · Server botsin.space
ITHCWY · @ithcwy
17 followers · 190 posts · Server botsin.space
Andrew Feeney · @andrewfeeney
388 followers · 939 posts · Server phpc.social

We use the Fast Fourier Transform () to do for things like images (e.g. jpeg). Most of the information is dropped and only the most important information is retained, which when reversed provides a noisy but recognisable version of the original image.

If this happens with a simple matrix of multi-dimensional vectors (e.g. a bitmap) could it not also be done with word embeddings like to perform lossy compression on text? Is this a thing?

#fft #algorithm #lossycompression #word2vec

Last updated 2 years ago

Thomas E. Gladwin · @TEG
453 followers · 443 posts · Server mastodon.online

Messing about with using a simulated toy language - nice way to try to grok and check how things are working. tegladwin.com/files/howto/word

#word2vec

Last updated 2 years ago

Stephan Janssen · @stephan007
167 followers · 27 posts · Server techhub.social

Loving my new "intelligent" search feature 🥳

Next step: create my own based on a Devoxx corpus which consists out of 10K technical talks!

#cfp #llm #transformers #pytorch #happyholidays #word2vec

Last updated 2 years ago

Dirk Van den Poel · @dirkvandenpoel
61 followers · 14 posts · Server mastodon.online
Josef Fruehwald · @JoFrhwld
339 followers · 75 posts · Server lingo.lol

there’s typos, and maybe everything isn’t perfectly explained, but i’m pretty proud of these course notes on how works. jofrhwld.github.io/teaching/co

#word2vec

Last updated 2 years ago

Josef Fruehwald · @JoFrhwld
339 followers · 75 posts · Server lingo.lol
Thomas E. Gladwin · @TEG
376 followers · 251 posts · Server mastodon.online

Psychology-slash-AI paper about how semantic similarity, as per a language model (word2vec), is statistically associated with behavioural congruence effects at the item level, based on data from an IAT variant:
tandfonline.com/eprint/HQSV7KQ

#association #bias #nlp #ai #psychology #IAT #word2vec #semantic

Last updated 2 years ago

K. J. R. · @kjr
31 followers · 88 posts · Server qoto.org

A superficial but interesting article about text analytics, with a comparison with deep learning applied to pictures.
Machine Learning algorithms: Working with text data

bit.ly/2rhVIrJ

#word2vec #machinelearning #deeplearning #nlproc

Last updated 6 years ago