https://hackerpaper.kaipfister.de/p/2010.06775/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=24858452urn:hackerpapers:mention:2010.06775:24858452Thu, 22 Oct 2020 14:35:42 GMT<p>It&#x27;s a super cool paper that invents &quot;vokenization&quot; to generate a large amount of visually-grounded language datasets and trains visually-grounded models on those. Most language models are trained on pure text data. Although it achieves significant success in recent years, this is not how humans acquire a language. It raises an interesting question &quot;Can language models achieve a high level of language understanding by reading the text input alone?&quot; The answer is probably &quot;no&quot;. To push the boundary of language models, adding other learning signals in the learning process is the key to success.…</p><p><a href="https://hackerpaper.kaipfister.de/p/2010.06775/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CL