OpenAI Embeddings API - (Interview Recap and Background)

[@connor-shorten]

Please check out the full podcast: kzbin.info/www/bejne/q3fbl41mq7Gmpbc! Thank you for watching, I really hope this content is useful for understanding the latest advances in Deep Learning for Search!

[@alexandrei1176]

Where you able to ask him about Niels Reimer findings?

[@connor-shorten]

We discuss the general topic of evaluation of embeddings at 12:35 in the podcast -- for your information (copying and pasting this from Yannic Kilcher's video and discussion with Arvind) Copy-Paste from Yannic Kilcher COMMENTS DIRECTLY FROM THE AUTHOR (thanks a lot for reaching out Arvind :) ): 1. The FIQA results you share also have code to reproduce the results in the paper using the API: twitter.com/arvind_io/status/... There's no discrepancy AFAIK. 2. We leave out 6 not 7 BEIR datasets. Results on msmarco, nq and triviaqa are in a separate table (Table 5 in the paper). NQ is part of BEIR too and we didn't want to repeat it. Finally, the 6 datasets we leave out are not readily available and it is common to leave them out in prior work too. For examples, see SPLADE v2 (arxiv.org/pdf/2109.10086.pdf) also evaluates on the same 12 BEIR datasets. 3. Finally, I'm now working on time travel so that I can cite papers from the future :) END COMMENTS FROM THE AUTHOR

[@juangarcia5876]

Are there any videos where you discuss usecases for image domains?

[@connor-shorten]

Yes! We recorded a podcast with Karen Beckers about search for eCommerce that will be published very soon on the SeMI KZbin channel -- if you want more advice for your particular ideas, happy to continue the chat in the Weaviate slack channel!

[@lastfatality]

Do u have any videos for beginners?

[@connor-shorten]

I recommend Andrew Ng's course, I might add content like this to Henry AI Labs in the future if I see more interest in it.