Hey there! I'm the first author and I'm very shocked to see this on HN. To answer your question -- the distribution change that the models should be robust to is the multilingual lexicon, as discussed with the Singlish example. We know that a speaker's linguistic background can influence how they communicate in Creole. So with the Singlish example, a sentence with the same meaning can be really different if its a person who also speaks Mandarin, versus someone who speaks Malay. Maybe in a training set, we'll see that a dataset is actually biased towards examples containing Mandarin, and it then becomes important to be robust towards the lesser represented Malay then, for example.
Singlish and Nigerian Pidgin are a really good examples of Creoles where we would expect robustness to multilingual vocabulary to be important, as these Creoles are are "linguae francae" within their respective countries (although, these two languages have VERY different levels of acceptance within their respective societies). Looking back on this work, Haitian Creole makes less sense, as Haiti is largely a monolingual country, with only a few also speaking French. But with the history of the language, the vocabulary is still mixed between French origin and like Akan/Igbo/Yoruba etc.
Hi! Its a bit late, but I'm following up on this thread. If you're interested, feel free to reach out to me and we can coordinate to start writing some snail mail! Perhaps we can also coordinate with the other commenter above? You can reach me at hclent1 the at symbol gmail :-)
Hey, I'm following up on this thread, even though its a bit late, because I'm really interested in this idea of snailmail and such ideas for topics of discussion. Is it OK for me to e-mail you about this?
I'm not currently reading like I want to, but I think having a snail mail reading discussion thing would be awesome. Maybe we can loop in the other user who also replied to this comment thread?
I'm also curious about this! I would love to know how to discover things in Discord. I'm only a part of 1 or 2 Discord channels, but I learned about them from podcasts that advertised them.
I love this!! I still try to send snail mail as often as possible -- everyone love recieving it. But the problem is that I rarely ever get a reply back and... well, its a bit frustrating to send letters and postcards without reply. So finding a community like this, where I could not only enjoy writing/recieving letters, but also talk about interesting things! 10/10.
Singlish and Nigerian Pidgin are a really good examples of Creoles where we would expect robustness to multilingual vocabulary to be important, as these Creoles are are "linguae francae" within their respective countries (although, these two languages have VERY different levels of acceptance within their respective societies). Looking back on this work, Haitian Creole makes less sense, as Haiti is largely a monolingual country, with only a few also speaking French. But with the history of the language, the vocabulary is still mixed between French origin and like Akan/Igbo/Yoruba etc.
Hope this helps, thanks for taking a look :-)