Linguistics Outside of Academia




When I tell my Alexa to announce that dinner is ready, not only does this magical electronic device understand what I'm saying, but it does just that by first ringing a dinner bell on all my connected Amazon devices and then speaking my recording of, "Dinner is ready."

This is no small electronic feat. Behind all the devices we carry that understand spoken language: Siri, Google, Alexa, and more, is a team of Language Engineers using their knowledge of linguistics and human nature to improve the accuracy of our verbal commands.

I read a fascinating interview with Language Engineer, Brent Woo, on the website Superlinguo. I had never thought about linguistics being such an important part of making our artificial intelligence (AI) work the way it does. When asked, "What is your job?" this was Mr. Woo's response:

Language Engineers work on the language designs for voice assistants. When you say something funny like “let’s raise the roof” and the voice assistant can interpret that to raise its volume to the maximum level, that was likely due to an LE language design. LE look into common variations on requests, understand the failures in human-machine communication, and adjust recognition coverage for them. We work on the scale of millions of utterances a day, so this process also has to be automated to some degree.


There is so much science behind how my voice assistants work. I remember watching Star Trek as a kid and thinking the idea of talking to a computer was magical. Captain Picard could simply say, "computer," and the computer would listen and do what he said. It seemed too good to be true. 

But here I am at 50 years old. I've got an Alexa connected in multiple rooms in my house. I can talk to my TV remote and it will do what I tell it to do. I have a "smart plug" so when I use it with my Christmas tree lights I can turn them on and off with just a voice command. I have a computer in my pocket that I can pull out and say, "Siri, remind me to pick up my kid at 2:00pm," and at 2:00pm I will get a reminder.

All of this is possible because of scientists. And some of those scientists just happen to be linguistic experts like Mr. Woo. They have to take their knowledge of the human language and apply it to the commands we humans want done.

When Mr. Woo was asked, "How does your linguistics training help you in your job?" he responded:

A common scenario is product or business stakeholders will request to have utterance X go to feature Y, and an LE will know from their expertise that no sane human customer would utter X to trigger Y. The LE has to present the case why that wouldn’t be a good design, and propose alternatives, and overall help the feature get better language coverage. This involves classic syntax for figuring out sentence structures and variations, and semantics for the annotation schema. Sometimes phonetics comes up when there are speech recognition conflicts or misfires! 

A quick look on hiring websites confirmed that linguistics experts are needed in many fields to help with this kind of AI integration. I found positions available all over the country. Mr. Woo's LinkedIn page had a link advertising for Alexa and their science team:


The field of voice recognition is evolving so incredibly fast. Linguistic experts help make sure our AI works the way we want it to.


WORKS CITED:

Superlinguo, Lauren Gawne. “Linguistics Jobs: Interview with a Language Engineer.” Superlinguo, 4 Sept. 2022, https://www.superlinguo.com/post/694502157231259648/linguistics-jobs-interview-with-a-language. Accessed 23 Sept. 2022

"Meet the Alexa Science Group." YouTube. Uploaded by Inside Amazon Videos, 5 Jul. 2016. https://www.youtube.com/watch?v=SMLuo74LHN8. Accessed 23 Sept. 2022


Comments

Post a Comment

Popular posts from this blog

Like, totally, dude. This assignment is, like, awesome.

Image
In an effort to seek out other sources on linguistics, I took a gander on the page Unravel . If English words had jobs , by Christopher Hoyt, was absolutely hilarious. The letter, written by the word LIKE, and the corresponding answers from management (named MERCURY), had me laughing all the way through. Using comedy, the article talks about how language continues to adapt and change as our culture changes.  👉 One of the tongue-in-cheek points I think Hoyt was trying to make is the average English vocabulary is shrinking. LIKE said that they understood management had to make some "tough calls about which words you would keep on staff." I decided I wanted to see if that was true. My husband and I have had conversations about diminished word use and the increase of emojis and gifs to get points across. A quick Google search led me to an article by David Shariatmadari: Why it's time to stop worrying about the decline of the English language . Right away, this article seemed...
Read more

Words matter.

Image
This meme came across my Facebook feed. It's the kind of social justice thinking that sparks my interest and fuels many of my passions. However, the last sentence really struck me.  Words matter. I've always known and understood that words matter. However, now that I'm in Introduction to Language Studies, I find myself thinking about words even more than I did before. This meme led me to do a search within the field of linguistics to learn more about the connections between language and social behavior. The first Google search I did took me to the Linguistic Society of America  (LSA). I continue to be intrigued about linguistics as a profession. I've already explored the use of linguistics in the field of artificial intelligence. Click here to go to that blog post. On the LSA web page, I was able to learn more about the field of sociolinguistics . Walt Wolfram wrote a very interesting article. He says: Language is one of the most powerful emblems of social behavior. In...
Read more

Learning about the Gullah Geechee language

Image
 Learning about the Gullah Geechee language I stumbled upon Sunn m'Cheaux and his work while killing time on TikTok. I wanted to make sure my For You Page was diverse and included creators of color so I followed him. I found how he connected language to racial issues fascinating. Prior to TikTok, I had no idea how much I seriously enjoyed those parts of linguistics. The short video lessons opened my eyes to concepts I had never had presented to me before. For example, Sunn m'Cheaux made a post (see video attached at the top) on July 10, 2022 about the use of the word "throwed" in Usher's hit song "Confessions" (see video attached below). He explains that "cross-over" (white) audiences misheard the lyrics because "throwed" was not a formally recognized word at the time the song came out.  Watch this   These are my confessions Just when I thought I said all I can say My chick on the side said she got one on the way These are my confessi...
Read more