4 ms·

> brains seem to edit I'd love to hear from NLP experts on speech vs writing from a machine's perspective. > not a representation of natural communication so

> brains seem to edit I'd love to hear from NLP experts on speech vs writing from a machine's perspective. > not a representation of natural communication so much as an idealised version Written language lacks stresses and intonation. I wonder if the written dialect attempts to fill in this gap, and whether tonal languages idealize their speech differently. > conversational tone can improve formal writing about complex topics Feynman's lectures are a nice instance of this.

syllogism 11y ago
Spoken language is pretty diverse. The data set we've been using for spoken language understanding is a corpus of telephone conversations between strangers about an assigned topic. Naturally, this contains a lot of ums and uhs! But spoken language about technical topics and difficult ideas tends to also be disfluent. For a long time computational linguists were using algorithms that performed okay on written language but poorly on spoken language. Specifically, they used algorithms that weren't linear-time in the length of the input. This meant that the input had to be pre-segmented, so you had to run another model before the parser, and then accept the errors from the previous model. Now that we've got linear-time algorithms that are doing everything jointly, it's not so bad. We just need to get a bit better at using intonation features, and parsing lattice input so that we can deal with the recognition errors. Mostly, it's a relief not to have to deal with the really tortured language often found in journalese. News reports are written really weirdly.
foota 11y ago
I have an interesting anecdote relating to this, one of my best friends told me the other day that she says the word "probably" as probly and even listening closely to what she says with that in mind it's still difficult to hear anything other than "probably"

syllogism 11y ago

Spoken language is pretty diverse. The data set we've been using for spoken language understanding is a corpus of telephone conversations between strangers about an assigned topic. Naturally, this contains a lot of ums and uhs! But spoken language about technical topics and difficult ideas tends to also be disfluent. For a long time computational linguists were using algorithms that performed okay on written language but poorly on spoken language. Specifically, they used algorithms that weren't linear-time in the length of the input. This meant that the input had to be pre-segmented, so you had to run another model before the parser, and then accept the errors from the previous model. Now that we've got linear-time algorithms that are doing everything jointly, it's not so bad. We just need to get a bit better at using intonation features, and parsing lattice input so that we can deal with the recognition errors. Mostly, it's a relief not to have to deal with the really tortured language often found in journalese. News reports are written really weirdly.

foota 11y ago

I have an interesting anecdote relating to this, one of my best friends told me the other day that she says the word "probably" as probly and even listening closely to what she says with that in mind it's still difficult to hear anything other than "probably"