Ganesh

[hardmaru @hardmaru]

As neural networks scale up, the most valuable people in deep learning field today are the people who can make systems 10-100x more efficient.

[Muhammad AbdulMageed @mageed]

I love making complex content accessible to diverse audiences, which is why I'm excited to lecture about #deeplearning in U. of Toronto Quant-TIDE workshop later today (Quantitative Training for Intersectional Diversity in Engineering; https://t.co/WjufDitEEG). 1/n

[thamar | @thamar_solorio]

Tech companies about to take oodles of graduate students as interns: Thank you for providing our students the opportunity to work on exciting NLP projects, and access to your valuable infrastructure. 🧵

[Adi Haviv @adihaviv]

What if I told you that transformer language models don’t need positional embeddings?

It turns out that transformer language models *without any positional encoding* still learn positional information.

https://t.co/spLcUjvqqf
@ori__ram @OfirPress @peter_izsak @omerlevy_
🧵1/5

[Sergey Tulyakov @SergeyTulyakov]

Got a research internship offer? Congrats! Here are some tips to make it a success

[Nathan Lambert @natolambert]

The real challenge of doing a phd in cs these days is developing the presence of mind to accept that only your challenges matter. There's nothing to gain by stressing over other people having "more" papers. Figure out how you solve problems and what is exciting, nothing more.

[Muhammad AbdulMageed @mageed]

We built speech recognition models that better serve six different groups of non-native Eng. speakers—Arabic, Hindi, Korean, Mandarin, Spanish, and Vietnamese L1. More: https://t.co/56f6yM6jTn #speech #DeepLearning #NLProc @UBCiSchool @UBCLinguistics @CAIDA_UBC @UBC_CS

[(((ل()(ل() 'yoav))))👾 @yoavgo]

believe me, i find the "what-if-we-made-it-even-larger??" research trend as intellectually boring as the rest of you. but i seriously don't understand why generally smart people are driven to luddism by it. you don't have to work on it, you know.

[Muhammad AbdulMageed @mageed]

Deep learning is transforming how humans interact with other humans and with machines. Here’s a Medium post I put together explaining, in simple terms, how it actually learns about language. #DeepLearning #AI #NLProc https://t.co/1jstpWrQPD

[Muhammad AbdulMageed @mageed]

We're accepting applications for a talented Research Associate in Natural Language Processing and Deep Learning to join my group at #UBC. Details here. https://t.co/RWk4bVBYZj #nlp #DeepLearning #MachineLearning @UBCLinguistics @UBC_CS @UBCiSchool @UBCLangScis @CAIDA_UBC

[Muhammad AbdulMageed @mageed]

1/2 Five papers will appear at #EACL2021 (2 main, 3 workshops). With my students @ChiyuZhang0851 @arelmadany @e_nagoudi @MKhalifaaaa @genuinelygeek, Rannie Lin, Dinesh Pabbi and colleagues @nyhabash @hbouamor, Shady Elbasiouny, and others

[Muhammad AbdulMageed @mageed]

I’m very excited about this upcoming talk feast, and so much looking forward to engaging discussions with colleagues from diverse fields of the type UBC research culture nurtures. Join us!! #AI #DeepLearning #LanguageMatters #Cantopop

[RSC // SRC @src_rsc]

Social media and the #pandemic: What we’re learning from the Twitterverse /via @globeandmail Written by @mageed @UBC https://t.co/7AlMnAIHga

[UBC Linguistics @UBCLinguistics]

"Social media and the pandemic: What we’re learning from the Twitterverse" a new piece by @UBCLinguistics faculty member Dr. Abdul-Mageed (@mageed)
https://t.co/F4xO5nX27w https://t.co/sqgkert2ng

[Kate Saenko @kate_saenko_]

We need a lot more research on how to defend against AI-generated misinformation. I cover some recent academic efforts in my talk, but governments and companies must do a lot more
https://t.co/E3WoqVqFcX

[Muhammad AbdulMageed @mageed]

🙌🙌🏽Happy New Year!!
I'm thrilled to share new paper w/@arelmadany & @e_nagoudi: "ARBERT & MARBERT: Deep Bidirectional Transformers for Arabic". Two powerful language models; new benchmark for Arabic NLU (ArBench) covering 5 different tasks; new SOTAs. 1/n #NLProc #ArabicNLP

[Ganesh @buzzganesh]

Stoked to have my first acceptance during PhD at
@coling2020

Automatic Detection of Machine Generated Text: A Critical Survey
w @mageed, @laksvs

We
- survey detectors for English
- categorize false +ve errors of RoBERTa (SoTA) & more
- discuss future directions

arXiving soon

[Dhivya Chinnappa @dhivyachinnappa]

In the coming days, I will live-tweet about papers from ACL 2020 focusing on the area ‘Computational Social Science and Social Media’ in Tamil and English. Follow #acl2020ta for Tamil ACL tweets and #acl2020en for English ACL tweets.

#acl2020nlp

[Ganesh @buzzganesh]

For those techies who are puzzled on where to work/study, here comes a site which gives good insights on why Canada should be considered:
https://t.co/ZiQ7xRtK9J
(from my friend, @yottabytt)
Retweet! :)

[Kota Yamaguchi @kotymg]

Styling with Attention to Details. (arXiv:1807.01182v1 [https://t.co/NV4xajuiSu]) https://t.co/IHq4nHy2P2 #research #feedly