Hacker News Reader

Comment by famouswaffles | Hacker News Reader

It's weirdly flown under the radar but GPT style models as translators are a lot better than state of the art Machine Translators (Deepl, NLLB, Google etc).

Like above, you can already see the difference with close language pairs where Google etc are already very good.

For pairs like English and Japanese etc, google et al will happily devolve into half gibberish so the difference is even more stark.

I did a number of examples a couple months back with English/Chinese before 4 was released. Even then you can see it's a lot better and 4 is as usual a lot better than 3.5.

https://github.com/ogkalu2/Human-parity-on-machine-translati...

View on HN

It's not under the radar because that's sadly not the case, although it looks like it should work better on the surface. Neural machine translation is just more consistent, doesn't hallucinate and can be easily and cheaply trained over time.

They have some benefits, but as a lot of LLM research has found, they are not production ready. Yet.

>It's not under the radar because that's sadly not the case

It is the case. Haven't seen anyone who uses both who thinks otherwise.

also https://arxiv.org/abs/2301.13294. This benchmarks 3.5 which is quite a bit worse than 4 as it is against Google, NLLB, Deepl

here https://arxiv.org/pdf/2304.02210.pdf, GPT wins overwhelmingly with human evaluations. Seems like the typically evaluation models aren't really cutting it anymore especially BLEU

>Neural machine translation is just more consistent, doesn't hallucinate

It's not more consistent. The 2nd bit is just wrong lol.

One of the biggest complaints of Deepl is the tendency to make stuff up to make translations seem more natural.

Summarization and Translation are the tasks GPT models hallucinate the least.

I wrote a much longer reply but looks like you deleted the downvoted comment and reposted it.

> It's not more consistent.

I think the parent was probably saying for a given input, Google MT provides the same output.

What is the value of temperature/variability in a LLM powered MT model?

I’d assume given the same inputs, you should only be given the best output.

>I think the parent was probably saying for a given input, Google MT provides the same output.

I don't care about being given the exact same output (you're not getting deterministic translations from people either). I care about quality translations. Variability for GPT style translations is much more about word choice and style than wrong or extremely different translations. and if i really wanted to, i could guide both (word choice, style) either with instructions or examples.

Good news! Per your second link, GPT-4 is a stunning improvement up to "borderline passes quality control"!

In my experience, ChatGPT tends to produce more fluent output, but is less likely to closely follow the input. For some high-resource language pairs, complete mistranslations are rare, but for other languages, not so much. Of the ones I can evaluate, Burmese is particularly error-prone:

ChatGPT translates ငါ့ကို ဒုတိယ အခွင့်ရေး တစ်ခု ပေးခဲ့တယ်။ as "I have received a second warning.", which is incorrect. အခွင့် https://en.wiktionary.org/wiki/%E1%80%A1%E1%80%81%E1%80%BD%E... does not mean "warning", even though that is a likely completion of "I have received a second " in English.

Google Translate gives me "Gave me a second chance.", which closely matches the Burmese sentence down to dropping the subject (common in Burmese, rare in English) which makes the translation sound weird.

So any claim that ChatGPT is better/worse at translating really needs to specify the languages involved and what your goal for the translation is. (E.g. the benchmark paper you link seems to focus on the ability to steer the translation by providing additional context.)

I’m not making a claim for chatGPT so much as I am making a claim for GPT style models.

It’s not really a question of high resource vs low resource languages so much as what languages ended up in the training corpus.

1. There’s a lot of transfer learning going on with predict the next token LLMs. A model trained on 500B tokens of English and 50b tokens of French will speak French far better than if it was trained on only 50b tokens of French.

2. You don’t need parallel corpora for every single pair you want to translate to. This means that GPT LLMs only need single text data for the vast majority of languages. Training most NMT models you would need Burmese/English parallel data.

Both of the above combine to mean that not only is quality demonstrably better, amount of data needed is lower too.

GPT's burmese isn't worse because it's low resource. It's because open ai made no specific attempt to include burmese text.

They're not even trying. GPT-3 was 93% English with the 2nd biggest holder less than 2%

>> 2. You don’t need parallel corpora for every single pair you want to translate to.

Neural Machine Translation models don't need those either. That's been the case at least since 2016.

That was one of the big claims by Google, that Google Translate had created its own "interlingua" thanks to which it could translate between languages for which it doesn't have parallel corpora:

Following several months of testing, the researchers behind the AI have seen it be able to blindly translate languages even if it's never studied one of the languages involved in the translation. "An example of this would be translations between Korean and Japanese where Korean⇄Japanese examples were not shown to the system," the Mike Schuster, from Google Brain wrote in a blogpost.

https://www.wired.co.uk/article/google-ai-language-create

It was obvious already back then that this "interlingua" was no other than English, e.g., from the same article:

The team said the system was able to make "reasonable" translations of the languages it had not been taught to translate. In one instance, a research paper published alongside the blog, says the AI was taught Portuguese→English and English→Spanish. It was then able to make translations between Portuguese→Spanish.

But, hey, hype.

>Neural Machine Translation models don't need those either. That's been the case at least since 2016.

Right. But those models still operate on parallel corpora alone.

In LLMs, translation pairs are not the only data improving translations. Context, Cultural Knowledge gleaned from other sources. Fluidity of the target language writing itself.

>But, hey, hype

It's not hype man. GPT-4 is much better for distant language pairs.

I used GPT-4 and this was the result:

Please translate this:

ငါ့ကို ဒုတိယ အခွင့်ရေး တစ်ခု ပေးခဲ့တယ်။

The sentence "ငါ့ကို ဒုတိယ အခွင့်ရေး တစ်ခု ပေးခဲ့တယ်။" translates to "They gave me a second chance." in English.

Your reply is not passing my sniff test. Your hype and bias are showing.

It may present more fluent text, but if it doesn't know it's strayed from the source text and you can't tell either (because you don't understand source language) then you'll end up with error-laden pseudo translations. At least with NMT you know the errors are consistent.

I don't know who you know who thinks GPT is ahead, but nobody in the very well funded translation industry has a GPT powered translation engine for the key reason that it's not ready for production. For a human to post edit MT, we're mainly talking fixing broken vocab. You'd never present raw MT to a client. It needs editing. Heavily. To think LLM translation doesn't need editing is either coming from someone not in the industry, or blinded by hype. And the kind of editing issues are more insidious, like those found in voice dictated texts. Homophones aren't flagged by QA software because they are real words. Just like LLMs make real sentences, except when they don't, but good luck detecting that, and editing out the additional meaning it has decided to inject.

Have you tried to run the GPT-4 API on a segmented xliff at all? If the segmentation is bad, and full of tags, GPT4 breaks completely. It tries to close sentences that run across segments, it can't handle tags in-line (the chatgpt interface can, but you can't use that at scale).

It can do some impressive work, don't get me wrong, but I'm not sure how hands-on you've really been if you think it's a solved problem.

Production translation is a non-trivial output. The entire industry hasn't released an LLM solution yet for translation (excepting the rewording mini features). What makes you think you know more than those on the ground? Or have you developed something that's still in stealth?

EDIT: Oh wow, all of your 107 submissions to HN from the 6 months your account has existed, have been about AI and LLMs. I guess I got the hype part right. As for industry knowledge the jury is still out, but this could well be the classic HN "I understand tech so obviously I understand everything" play. Keep us posted!

> It may present more fluent text, but if it doesn't know it's strayed from the source text and you can't tell either (because you don't understand source language) then you'll end up with error-laden pseudo translations. At least with NMT you know the errors are consistent.

To be fair, this is an infamous failure mode of neural mtl too, and a big part of what makes the discourse around GPT so ... evocative of the discourse in 2017.

>It may present more fluent text, but if it doesn't know it's strayed from the source text and you can't tell either (because you don't understand source language) then you'll end up with error-laden pseudo translations.

I understand Korean. I've tested it there too and my evaluations haven't changed. It's better. A lot better.

That aside, nothing you've said here doesn't apply to NMT.

I'm going to be Frank here. You don't what you're talking about when it comes NMT and it makes this conversation very frustrating. You keep mentioning cons that apply to NMT as well.

Literally almost all SOTA NMT models (NLLB for instance) today are Transformers or some transformer variant (In fact this was the original proposition in the famous Attention Paper). They're just transformers trained for a translation primary objective.

Those cons of GPT, NMT inherit them too. They hallucinate, they can make pseudo translations etc.

>EDIT: Oh wow, all of your 107 submissions to HN from the 6 months your account has existed, have been about AI and LLMs.

So ? That's what I like to do. Is this some attempt at an ad hominem?

The 2nd paper I linked is about document level translations.

>The entire industry hasn't released an LLM solution yet for translation (excepting the rewording mini features). What makes you think you know more than those on the ground?

Training and inference for GPT scale models is extremely expensive.

>> Keep us posted!

Dude, don't be an asshole. You got useful experience. Help others with it instead of putting them down.

It's fine, people can disagree robustly sometimes. I'd still buy him a beer, or whatever he wanted to drink, if we were in person. It's good to talk openly, right?

Talk openly and argue robustly by all means but why make things personal? That just adds noise to an already noisy conversation. We should try to maximise the amount of knowledge in our discussions, not bring it down.

But I guess that would indeed sound much better around beers.

>but nobody in the very well funded translation industry has a GPT powered translation engine for the key reason that it's not ready for production.

Translator for 20 years talking about using GPT for Japanese-English.

https://www.youtube.com/watch?v=5KKDCp3OaMo&t

https://www.youtube.com/watch?v=8JUepj7wIl0

Indeed. That translator, Tom Gally, is also a Professor at the University of Tokyo and a regular on this site: https://news.ycombinator.com/user?id=tkgally

I've tried using ChatGPT to translate Latin texts from the Renaissance. I know enough Latin (intermediate, but can figure it out with a dictionary) to check it, and it was very, very impressive. What blew me away was not just the fluency of the translation but the fact that I could drop in highly imperfect OCR'd text from Google Books, and it didn't have any trouble making sense of garbled passages. This ability makes it a really distinct advance on Google Translate and the like, at least for my purposes.

Also, another underrated feature: I asked it to summarize each page in a single sentence, while also picking out the passages most relevant to my research question. It did a great job.

That is a truly exceptional use case, and I'm impressed too.

I should have specified above that I'm referring to the practicalities of professional translation workflows as they currently exist for things like high volume flows in dozens of formats, and translation memory leveraging etc.

"Neural machine translation... doesn't hallucinate..." I've looked at neural MT output from Arabic and Russian, and there was lots of hallucination.

Yeah, GPT-4 is so good Iceland is using it for language preservation of Icelandic as it translates text so well, letting them enrich the language with new works.

Like you, I’ve definitely often though “Come ON!” about the increasingly archaic Google Translate in light of DeepL etc. It has really stagnated over the years.

Reply on news.ycombinator.com