Patrick Y. Wu (@PatrickYWu) header
Patrick Y. Wu (@PatrickYWu) profile picture

assistant professor of CS @AmericanU | computational social science, NLP, AI/ML | previous: @CSMaP_NYU, @UMich, @UChicago

Washington, DC
Joined July 2011
Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)Photo by Patrick Y. Wu (@PatrickYWu)
I was Judge Posner's Chief RA for a year after college. His mentorship still resonates in my thinking and work today, and our weekly conversations about his cat, Pixie, made me a lifelong cat person. My condolences to his family and loved ones.
We mourn the loss of Senior Lecturer Richard Allen Posner, a longtime #UChicagoLaw scholar, former federal appellate judge, and pioneer in the field of law and economics, who died on November 10. Read more about his legacy: buff.ly/e3c3WdL
7
297
We mourn the loss of Senior Lecturer Richard Allen Posner, a longtime #UChicagoLaw scholar, former federal appellate judge, and pioneer in the field of law and economics, who died on November 10. Read more about his legacy: buff.ly/e3c3WdL
98
1,050
392
3,533
639,236
Patrick Y. Wu retweeted
I was really annoyed at the debate about what empirical measure to use for ideology so I coded a lil script to use a LLM judge candidate ideology with pairwise comparisons and then put everyone on a single L-R scale (after Patrick Wu's work). Results are here. Judge for yourself Generally, in the Senate, we are seeing a smaller benefit for moderates than the predictions made by the centrist crowd during the 2025 vintage of this debate. And in the House, DSA candidates are running better than the average Democrat!!
I think this is right. It's not just that Talarico (et al) combines elements of left populism, abundance pragmatism, and cultural moderation. It's also that the definitions of "left" and "moderate" are in flux. Take "left." Talarico's all-of-the-above position on energy, eg, would have been grounds for near excommunication from the climate left during the heights of the Green New Deal era. But many leftists now proudly claim him as one of their own, in part bc of his rhetoric on corporate abuse of power and "not left vs right, but top v bottom." So what we're seeing might be fusionism. But just as importantly, the elements being fused are in flux.
16
44
12
357
69,018
Patrick Y. Wu retweeted
I'm on the job market this year! I work on measuring the societal impacts (overreliance, homogenization etc.) of AI/LLMs and proposing ways to mitigate these. I'm at #COLM this week, and #NeurIPS later this year, and would love to chat more!
6
54
7
184
13,441
Patrick Y. Wu retweeted
LLMs are increasingly used for text annotation, esp. in the social sciences. Often, this involves placing text items on a scale: eg, 1 for liberal and 9 for conservative There are a few ways to accomplish this task. Which work best? Our new #EMNLP2025 paper has some answers 🧵
3
9
21
2,241
In the Global South, WhatsApp is more popular than X or Facebook. New in @The_JOP, we ran a WhatsApp deactivation experiment during Brazil’s 2022 election to explore how the app facilitates the spread of misinformation and affects voters’ attitudes. journals.uchicago.edu/doi/ab…
2
13
3
28
7,942
Can large language models (LLMs) fairly annotate data on contentious topics? Our new paper dives into this question—looking at whether LLM-generated labels reflect diverse viewpoints or skew toward majority perspectives. The results are surprisingly nuanced. 🧵
1
11
24
2,677
6/ This has implications for both research and industry: ⚠️ Fairness evaluations should be context-specific 🤔 Model choice alone will not solve bias 🔍 Human disagreements are part of the complexity—not a flaw
1
1
2
200
7/ 📄 You can find our full paper here: arxiv.org/abs/2503.23243 w/ @m_dot_brown (the project lead!), @ShubhamAtreja, and @libbyh
2
146
We're extending the deadline to Sunday, June 22, 11:59pm AoE! Submit your extended abstracts to the NLP4Democracy Workshop at @COLM_conf !
📢 Excited to announce the first workshop on ✨NLP for Democracy✨ (NLP4Democracy) at @COLM_conf! 🗓️ Submissions (extended abstracts) due on June 19, 11:59pm AoE 🔗 Website: sites.google.com/andrew.cmu.… #COLM2025 #LLMs #NLP #NLProc #ComputationalSocialScience
1
5
5,150
📢 Excited to announce the first workshop on ✨NLP for Democracy✨ (NLP4Democracy) at @COLM_conf! 🗓️ Submissions (extended abstracts) due on June 19, 11:59pm AoE 🔗 Website: sites.google.com/andrew.cmu.… #COLM2025 #LLMs #NLP #NLProc #ComputationalSocialScience
1
1
3
14
12,639
NLP4Democracy will be very interdisciplinary, and we encourage submissions from scholars who don't typically submit to NLP venues. Submissions should be extended abstracts of up to 2 pages (not including tables and figures). 📎 Submission link here: openreview.net/group?id=colm…
1
2
336
If you’re attending #APSA2024 today, feel free to join our panel on Navigating Political Narratives to hear from me and my incredible colleagues @allisonwkoh @siyuliang46 @_paugc @PatrickYWu. Thanks to @cbarrie @arthur_spirling for agreeing to be our discussants!
6
15
2,392
It was great working with @nicolekwu on this paper that's now out at PS: Political Science & Politics! Special thanks to @APSAtweets for helping us field the survey. Check out the summary of the paper below.👇
As departments and universities were rapidly/frantically holding panels in response to generative AI last year, @PatrickYWu and I got curious and partnered with @APSATweets to survey political science educators on AI & pedagogy. bit.ly/3YMl1TM
1
14
1,873
Excited to see this @journalqd ! The latest of several papers with @KirstenWidner and @annagunderson examining what interest groups do on social media 🐊🍾🍻
New publication dropped with @MG_Macdonald and @annagunderson ! journalqd.org/article/view/4… We examine the social media behavior of over 1,500 US interest groups on Facebook and Twitter 1/4
5
18
2,736