@fauxneticien Profile picture

Nay San

@fauxneticien

Linguist & data nerd. I mainly come here to lurk, and 💙 memes and things on #rstats 📈.

Similar User
Journal of Sociolinguistics photo

@JournalofSocio1

Sasha Wilmoth photo

@SashaWilmoth

John Mansfield photo

@johnbasil

Debbie Loakes photo

@debbie_dloa

Dr Celeste R Louro 🌿 photo

@CelesteRLouro

Dr Viola Wiegand photo

@violawiegand

Dr Marc Olivier-Loiseau photo

@loiseaulivier

Jayden Macklin-Cordes photo

@JaydenC

Wendy Bennett photo

@WendyBennett21

rolsi journal photo

@rolsi_journal

Erich Round photo

@ErichRound

DCAL photo

@DCAL_UCL

Elwys De Stefani photo

@elwysdestefani

Pinned

For fieldwork recordings with a mix of a target language and a more widely-used metalanguage for questions and commentary, we can automatically isolate and transcribe the latter—making possible text searches on untranscribed recordings! How? See👇 and arxiv.org/pdf/2204.07272… 1/7

Tweet Image 1

Nay San Reposted

The Jimmie Barker corpus (now part of the AIATSIS Collection) documents the longitudinal speech data of Muruwari Elder, Jimmie Barker, between 1968 and 1972. Access the data article by Mount et al. (2024) here: doi.org/10.1080/072686… #AJL #LanguageCorpora

Tweet Image 1

Nay San Reposted

Posts that share experiences of racism are disproportionately flagged for removal by humans and algorithms, with witnessing such content moderation influencing Black American's feelings of isolation, finds @leecinoo @krisgligoric @ria_kalluri @keeksanch @fauxneticien @jurafsky

Tweet Image 1
Tweet Image 2
Tweet Image 3
Tweet Image 4

Nay San Reposted

Big Tech promises a future with universal translation, where "people will be able to make more authentic connections in their native languages". This is the picture: human communication hostage to the machine; biases and hallucinations so hard to detect cross-culturally. 🧵1/8

Tweet Image 1

Nay San Reposted

📢 Excited to announce the expanded 🌟Responsible Foundation Model Development Cheatsheet🌟 ➡️ A Survey & Review of Tools🛠️& Resources🧮🗝️📚. ➡️ We ask what1⃣what responsible practices developers can adopt, &2⃣ what tools are missing, misused, or under-used in the AI dev…

Tweet Image 1

Did the gown and hood part of PhD things today, celebrating with some folks who helped me get there! @jurafsky @tolulope_ai @BarteldsMartijn

Tweet Image 1
Tweet Image 2

Nay San Reposted

Thrilled to introduce Mist, Rime's family of next-gen TTS models. Trained on a massive dataset of conversational speech, Mist is a powerful building block for real-time voice applications. rime.ai/blog/introduci…

Tweet Image 1

I contributed the speech modality here but if anything's been missed please do help add more! github.com/allenai/fm-che…

New Resource: Foundation Model Development Cheatsheet for best practices We compiled 250+ resources & tools for: 🔭 sourcing data 🔍 documenting & audits 🌴 environmental impact ☢️ risks & harms eval 🌍 release & monitoring With experts from @AiEleuther, @allen_ai,…



Nay San Reposted

🥳🥳🥳 Great news! I feel very grateful to be the recipient of an @NWOFunding Rubicon grant! After my @univgroningen @GroNlp PhD, I will continue my research as a postdoc at @stanfordnlp with @jurafsky I couldn't be more happy!

Tweet Image 1

Nay San Reposted

I’ll be at #ACL2023! Looking forward to presenting our paper ‘Making More of Little Data: Improving Low-Resource Automatic Speech Recognition Using Data Augmentation’ in the ‘Speech and Multimodality’ session! See you in Toronto! 🇨🇦 arxiv.org/abs/2305.10951


Nay San Reposted

Extremely excited that the paper of my @stanfordnlp visit has been accepted to #ACL2023! 🎉 Big thanks to my collaborators @fauxneticien, Bradley McDonnell, @jurafsky and @martijnwieling! Please find a short summary of our paper 👇


Nay San Reposted

i'm announcing the company we've been building. it's called rime. here's a tiny demo of what truly generative text-to-speech can accomplish.


Nay San Reposted

Great to hear from previous research assistant @fauxneticien - thank you to the many research assistants the Centre has been so lucky to have had as part of the Centre

Tweet Image 1

Nay San Reposted

Communities are eager to make the arduous process of language documentation easier, and models from the 20+ years of research in low-resource NLP are poised to help. But many language documentation projects forgo the use of any NLP models. Why? (1/8) aclanthology.org/2022.computel-…


Nay San Reposted

🎉🚨 New website alert! 🚨🎉 ALA Labs is our new website at @atlaslivingaust for people who want to learn to make #dataviz or conduct analyses in R! You can find how-to articles on how to do cool stuff that you can reproduce yourself! #RStats labs.ala.org.au

Tweet Image 1

Loading...

Something went wrong.


Something went wrong.