Saadia Gabriel
@GabrielSaadia
Followers
761
Following
139
Media
4
Statuses
165
NYU Faculty Fellow and incoming UCLA Assistant Professor. In her free time, interested in generation and ethics of AI.
Cambridge, MA
Joined August 2019
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
taeil
• 1006905 Tweets
#湊あくあ卒業ライブ
• 830149 Tweets
ホロライブ
• 211374 Tweets
あくあ色
• 178574 Tweets
Chiesa
• 127614 Tweets
伝説のアイドル
• 71670 Tweets
WORLD TOUR
• 66869 Tweets
#TaraftarİstifaBekliyor
• 66436 Tweets
BBドバイ
• 48351 Tweets
最高のアイドル
• 47919 Tweets
コロナ対策
• 41269 Tweets
最高のライブ
• 33470 Tweets
あくたん6年間
• 25459 Tweets
西鉄バス
• 24599 Tweets
テイルさん
• 19026 Tweets
カグラバチ
• 18538 Tweets
ドーピング
• 17295 Tweets
おつあくあ
• 13626 Tweets
#AskNani
• 12470 Tweets
スパチャ
• 11777 Tweets
サイヤ人級
• 11688 Tweets
ふつうの軽音部
• 10469 Tweets
June brought some small changes to my life: defending my thesis, moving across the country to start a postdoc
@MIT_CSAIL
, finally meeting the research team at Spotify and speaking at CHIL on potential harms/benefits of LLMs for healthcare! What will come next in July…
13
9
203
Breaking news - I will be presenting work on modeling implications of misinformation (Misinfo Reaction Frames) in-person at ACL later today! We focus on impact of Covid-19 and climate change misinformation, releasing a dataset of reader reactions to news headlines.👇
1
6
68
I made the list! Thank you to all the wonderful mentors who have supported over the years (especially
@YejinChoinka
,
@franziroesner
,
@nlpnoah
,
@hmd_palangi
,
@real_asli
,
@LukeZettlemoyer
,
@HannaHajishirzi
, Yulia Tsvetkov and
@MarzyehGhassemi
).
NOW ANNOUNCING: 2024
#ForbesUnder30
The entrepreneurs on the 2024 Forbes 30 Under 30 list are fighting climate change with carbon credit marketplaces, reinventing the 911 call, building banks and brewing up eco-friendly plastics.
125
18
66
8
9
48
Very excited to announce I'll be presenting a new ACL Findings paper at the GEM workshop next month! We investigate effectiveness of factuality metrics for abstractive summarization, including dialogue summarization. (1/2)
1
4
38
I may be one of the last NLPers on Earth to have discovered this, but it made integrated gradients in pytorch so much more straightforward and saved me time to do other fun things over the weekend like making a batch of homemade apple cider chai 🙌🏾
NLP community: Interpreting text models with Captum – an open source, extensible library for model interpretability built on PyTorch. Sentiment Analysis and interpreting BERT Models in the tutorials.
2
80
328
0
2
30
It's great to always remember when reviewing that there are other human beings on the other end of this process and every word you write will have an impact on them. An extra hour spent reading a paper carefully can make a huge difference in someone's life. End of Saturday rant.
0
2
27
I'm jumping back on Twitter for an important announcement! We can't wait for your submissions:
We are excited to announce the
#NeurIPS2023
call for tutorials! Deadline is June 21, and should include a discussion panel.
@MarzyehGhassemi
,
@GabrielSaadia
and I look forward to your submissions. Please help spread the word!
0
2
18
0
0
6
If you're going to NeurIPS, definitely stop by on Dec 11th!
Introducing the 2023 NeurIPS Tutorials!
(Reposting with corrected link)
Curious about the Tutorials offered this year at NeurIPS? Check out our blog post below:
0
18
68
1
1
7
Joint work from my MSR internship with amazing co-authors
@real_asli
,
@rahuljha
,
@YejinChoinka
and
@JianfengGao0217
. Looking forward to chatting with folks at ACL about evaluation of generation.
Paper:
#acl2021nlp
(2/2)
0
1
6
So glad to see this message. To say the last year has been tough is a huge understatement and even though I like to sugarcoat things in my life, I think particularly in grad school it can be far too easy to forget to slow down every once in awhile and take care of yourself.
The facade of normalcy is a trap. We will look around and see things looking normal and wonder why we don’t feel normal, if we are broken. We are broken. But also know that there will be a lot of people feeling the same way underneath. 6/7
2
29
259
1
0
5
@emilymbender
Definitely not a journalist (: , but this comes up in "The Oxygen of Amplification":
0
0
4
@anmarasovic
And last but not least, I'll recommend this paper from Regina Barzilay's group ()
1
0
3
@niloofar_mire
@emnlpmeeting
😭 this happened to me too. Get lots of rest
@niloofar_mire
and I hope you feel better soon!!
0
0
1
@anmarasovic
Actually, one other paper off the top of my head that is pretty important/insightful (: (), I’m forgetting plenty.
0
0
3
I appreciate the optimism of the last line. Motivational words to power me through the next two weeks' writing.
Powerful words by the great writer of the 20th century, James Baldwin, in a 1984 interview published in Paris Review.
0
6
30
1
0
2
@ybisk
@strubell
@NVIDIAAI
@Raspberry_Pi
@nvidia
This is so cool! I had a class in undergrad that used hardware and that was the main reason a number of us ended up becoming CS majors.
0
0
2
@anmarasovic
Highly recommend Kate Starbird (, ), Franzi Roesner () and Amy Zhang’s work as well ().
0
0
2
Looking through
#EMNLP2021
reviews (hooray for prompt feedback). While peer reviewing is great for constructive criticism, I get very concerned seeing how many points raised by reviewers are already addressed in the paper or seem like they're directed at another paper.
0
0
2
0
0
1
@anmarasovic
Isabelle Augenstein also has an impressive body of work on fact-checking, like MultiFC (). Fever is a go-to paper on fact-checking as well (, )
1
0
1
Classic
0
0
1
@universeinanegg
That’d be exciting to see, I wonder how much of the bottleneck is vision models having similar weaknesses to language-only models (e.g. focusing on the wrong things)?
1
0
1
@anmarasovic
I second the comment about Preslav Nakov (:, he has extensive work in this area, particularly propaganda ().
0
0
1
0
0
1
To find out more, you can stop by one of these poster sessions:
In-Person: Today (5/23) at 5pm IST in poster session PS3-1
Virtual: 5/24 at 7pm IST in poster session VPS2
Virtual: 5/25 at 7am IST in poster session VPS3
Paper:
1
1
1
Understanding implications of misinformation is a challenging topic, and I am grateful to wonderful co-authors
@YejinChoinka
,
@MaartenSap
,
@eunsolc
,
@franziroesner
, Skyler Hallinan and Pemi Nguyen for taking it on with me.
#acl2022nlp
0
0
1