#10 AI Success!: Translation Gets Faster and BetterMachine translation, properly used, can help us communicate better
Once again, our Walter Bradley Center director, Robert J. Marks, is back with Jonathan Bartlett and Eric Holloway —this time to to discuss real advances in AI in 2020.
Readers may recall that we offered a fun series during the holidays about some of the oopses and ums and ers in the discipline (typically hyped by uncritical sources).
Now it’s time to celebrate the real achievements! Let’s start with how machine translation, properly used, can help us communicate better.
Our story begins at 03:11. Here’s a partial transcript. (Show Notes and Additional Resources follow, along with a link to the complete transcript.)
Robert J. Marks: Okay, let’s get started with the countdown of AI smash hits of the last year or so. Number 10, AI text translations. If you’ve ever taken a foreign language, you know that you just can’t take a word in German and translate it to English, keep them in the same order and it be a proper translation.
Note: Bad translations can cause a lot of grief: “The HSBC bank has had to spend millions of dollars to rebrand a campaign that was improperly translated: The phrase “assume nothing” was translated in various countries to mean “do nothing,” which obviously changed the entire meaning—and not in a very good way.
In 2009, the bank had to relaunch the campaign to rectify the problem—all caused by a simple bad translation. – “When Translations Go Wrong” at Espresso Translations
Robert J. Marks: So something has happened in AI text translation, which has really helped us. What’s going on here, Jon?
Jonathan Bartlett: I’ve worked with some companies that would hire a translator and the translator would go and re-type everything from one language to another. And that’d take a lot of time. Now, what they can do now is they can automatically translate these applications from one language to another. Now we don’t expect it to be a hundred percent so you do an automatic translation and then you have your translators go and simply correct where the AI has gone wrong. And so what’s really made AI tech translation work so well—the expectation that it’s not going to be perfect, but that we’re going to have a second pass. The goal of the AI is not to replace the person but to make the person more powerful.
Robert J. Marks: I talked to a person that uses Google translate a lot: and I said, “What’s your means of success?” She says, “Well, I take Google translate and I transform for example, English into French. And then I transform the French back into English, using Google translate and see if I get roughly the same thing, because in translation, it’s like the old game of telephone where you whisper a phrase to a friend who whispers it to somebody else and you go down the line and at the end comes out just totally ridiculous of what it was initially set to.”
Note: Back translation is not a new idea. It is an old idea that has been digitized. For example, when translating the Bible from the original Greek into modern Inuit, back translation is used. It wasn’t originally a big AI thing; it was just a way of being sure that we were communicating what the text intended to convey: During a During a translation project of the Bible into modern Inuit, “I happened to overhear one of the priests turn to his bishop and say, ‘It’s hard sometimes to work with this text. It feels alive.’” If he conveyed that, he succeeded.
Keep watching for our nerds’ thoughts on other big AI advances in recent years! #9 is coming up soon!
- 02:10 | Introducing Jonathan Bartlett
- 02:39 | Introducing Dr. Eric Holloway
- 03:11 | #10: Text translation (Microsoft – https://docs.microsoft.com/en-us/ai-builder/prebuilt-text-translation, Apple – https://apps.apple.com/us/app/translate-translator-ai/id1375535400, DeepL – https://www.deepl.com/en/home)
- 09:19 | #9: “We hit the road with Comma.ai’s assisted-driving tech at CES 2020” (Road Show by CNET), comma two at the Comma.ai shop
- 13:31 | #8: “Daimler, Waymo, and GM Make Big Gains in Level 4 Self-Driving” (Mind Matters News)
- 19:37 | #7: “Hacking AI: Exposing Vulnerabilities in Machine Learning” (Datanami)
- 22:35 | #6: “After Thursday’s Dogfight, It’s Clear: DARPA Gets AI Right” (Mind Matters News)
- Jonathan Bartlett at Discovery.org
- Eric Holloway at Discovery.org
- #10: Text translation (Microsoft – https://docs.microsoft.com/en-us/ai-builder/prebuilt-text-translation, Apple – https://apps.apple.com/us/app/translate-translator-ai/id1375535400, DeepL – https://www.deepl.com/en/home)
- #9: “We hit the road with Comma.ai’s assisted-driving tech at CES 2020” (Road Show by CNET), comma two at the Comma.ai shop
- #8: “Daimler, Waymo, and GM Make Big Gains in Level 4 Self-Driving” (Mind Matters News)
- #7: “Hacking AI: Exposing Vulnerabilities in Machine Learning” (Datanami)
- #6: “After Thursday’s Dogfight, It’s Clear: DARPA Gets AI Right” (Mind Matters News)