Diyi Yang
@Diyi_Yang
Followers
17K
Following
12K
Media
129
Statuses
1K
Assistant Professor @Stanford CS @StanfordNLP @StanfordAILab Computational Social Science & NLP
Joined December 2016
Excited to be on the list of Forbes 30 Under 30 Class of 2021 in the Science category #ForbesUnder30
49
29
932
Very honored and excited to receive the @NSF CAREER Award😀. This fund will be used to develop socially aware language technologies to support mental health and well-being!. HUGE thanks to all my amazing students, collaborators, mentors, and advisors who help make this happen♥️
76
16
866
Very honored to have been selected as a #SloanFellow! Huge thanks to my incredible students and my mentors ♥️.
We have today announced the names of the 2024 Sloan Research Fellows! Congratulations to these 126 outstanding early-career researchers:.
78
21
567
I started as an assistant professor this week at @ICatGT @gtcomputing, after a 4-month wonderful visit at @GoogleAI. My lab focuses on computational social science and natural language processing. Super excited to work with amazing colleagues and students on cool research!.
23
19
425
Very honored to be a Microsoft Research Faculty Fellow. This is a collective honor, and would not have been possible without the support and efforts from my amazing students, mentors, colleagues, and collaborators! @ICatGT @gtcomputing @mlatgt.
Congratulations to the 2021 Microsoft Research Faculty Fellows! This fellowship recognizes innovative, promising new faculty whose exceptional talent for innovation identifies them as emerging leaders in their fields. Learn about their research interests:
61
8
423
Check out our tutorial on Limited Data Learning at #ACL2022, 9:30-1pm Irish time on May 22nd, with @colinraffel @ank_parikh 🚨. We will review current approaches and challenges associated with learning from limited labeled data #NLProc
6
46
309
Learning social skills is out of reach for most people🙁 . How can we make social skill training more accessible?. We introduce 🌟APAM🌟 (AI Partner and AI Mentor) that leverages LLMs for social skill training via realistic practice and tailored feedback!.
Social Skill Training with Large Language Models. People rely on social skills like conflict resolution to communicate effectively and to thrive in both work and personal life. However, practice environments for social skills are typically out of reach for most people.
7
64
303
We're looking for postdocs to lead work on socially aware NLP and computational social science. #NLProc #ComputationalSocialScience Reach out to me via email by Nov 15th if you are interested!.
7
87
289
🤩 Excited to receive the @Amazon Faculty Research Award to work on "abstract conversation summarization at scale" with my PhD student @jiaao_chen. Thanks @AmazonScience for supporting our research 🔥🔥🔥 @ICatGT @mlatgt.
21
5
292
Can #LLMs generate novel research ideas? Check out our one year-long human study with 100+ NLP researchers, done by the amazing @ChengleiSi.
Automating AI research is exciting! But can LLMs actually produce novel, expert-level research ideas?. After a year-long study, we obtained the first statistically significant conclusion: LLM-generated ideas are more novel than ideas written by expert human researchers.
2
34
277
So. I just finished the FIRST in-person teaching of a class in my career 🤖.
6
2
273
I'm looking for PhD students who are interested in Social Computing, Natural Language Processing to join our lab in fall 2020 at @ICatGT @gtcomputing. For more details and applying, please see: . Or talk to me over #CSCW2019😀
7
99
256
Our positive reframing work has been selected as an 👉 ACL 2022 Outstanding Paper 👈 #NLProc #acl2022 . Congratulations to my students and co-authors @cjziems 🎉. We'll be presenting this work in Dublin soon 😀.
You’re baking 🥧 and you ruin it! Do you want to shout “I’ll never make a good pie 🤬” . Self-talk can shape how we think and feel…. Maybe machines can help us look on the brighter side 😇 . We introduce this as a new #NLProc task: #PositiveReframing #ACL2022 @cjziems
13
18
236
We're very excited to release 🌟DiVA — Distilled Voice Assistant 🔊 @WilliamBarrHeld . ✅End-to-end differentiable speech LM; early fusion with Whisper and Llama 3 8B.✅Improves generalization by using distillation rather than supervised loss.✅Trained only using open-access
5
46
233
I’m extremely proud of my #undergrad Yujia Gao - her work on "Understanding the Usage of Online Media for Parenting from Infancy to Preschool At Scale" received a 🎖️Best Paper Honorable Mention Award from SIGCHI @sig_chi 🔥 #proudadvisor Also, congrats to all other winners!!!.
7
3
219
So honored to be part of this list - "30 Influential Women Advancing AI in 2019" with @deviparikh. And yes, we are advancing AI at @gtcomputing @mlatgt @ICatGT.
15
9
210
Very honored to receive this 😃 This would not have been possible without the support and efforts from so many students, mentors, and collaborators!.
Congrats to our own @Diyi_Yang for selection to @IEEEorg's prestigious list. Can't wait to see what the future has in store!.
19
2
211
Today is a big day!🤗 . My student @jiaao_chen has passed his thesis proposal on Efficient and Adaptive Machine Learning for NLP 🎉 🎉🎉 . Thanks for all the committee members @zsoltkira @mark_riedl @alan_ritter @colinraffel
4
9
205
We all interact with #LLMs every day, but how to design better human-LLM interaction?. Come to our #NAACL2024 tutorial: 🌟Human-AI Interaction in the Age of LLMs 🌟 w/ @tongshuangwu Marti Hearst. We will cover 3 aspects from a HCIxNLP perspective: . 🎨🎨Design: Why should they
4
30
200
Check out our paper on using #LLMs in #psychology 👇
Using large language models in psychology: . 💡 LLMs have the potential to advance psychological measurement, experimentation and practice. 💡 LLM generated on-topic, grammatically correct useless information, but not based on research and psychology construct. 💡A critical
2
34
187
In order to thank Anna Rogers @annargrs for her great program chair service at #ACL2023 @aclmeeting in Toronto, Canada decides to name their SkyDome as the Rogers Centre!!!
3
4
187
People like to talk as it's easy and natural. Now that there are Large *Audio* Models 🔊, which model do users like the most?. Introducing Talk Arena🎤: an open platform where users speak to LAMs and receive text responses. Through open interaction, we focus on rankings based on
2
45
183
Congrats to my PhD student Caleb Ziems @cjziems on being awarded the NSF Graduate Research Fellowship 🎉 Looking forward to all the great work from Caleb in the coming years 🔥🔥🔥 #proudadvisor.
9
2
178
Finished my first quarter of teaching #computationalsocialscience 😀 Had a fantastic time engaging with students at 224c 🔥
3
10
179
Thanks @AmazonScience for supporting our work on human-in-the-loop for long text generation!.
Representing 24 universities in 7 countries, the spring 2022 recipients of the #AmazonResearchAwards will receive funds to pursue their proposals across various research areas.
8
6
172
LLMs undergo alignment to make them helpful, harmless, and adapt to human preferences, but whose preferences & any #unintended impacts 🤔. Check out our work “Unintended Impacts of LLM Alignment on Global Representation” @michaelryan207 @WilliamBarrHeld.
Aligned LLMs should be helpful, harmless, and adopt user preferences. But whose preferences are we aligning to and what are unintended effects on global representation?. We find SFT and Preference Tuning steer LLMs towards US English use and opinions. 🧵
4
18
162
See our tutorial slides here 🚀 . Feel free to interact with it via comments 🤖.
We all interact with #LLMs every day, but how to design better human-LLM interaction?. Come to our #NAACL2024 tutorial: 🌟Human-AI Interaction in the Age of LLMs 🌟 w/ @tongshuangwu Marti Hearst. We will cover 3 aspects from a HCIxNLP perspective: . 🎨🎨Design: Why should they
1
30
156
Our recent #AAAI2020 work looks at neutralizing subjective #bias in text. It also releases a parallel corpus of 180K biased/unbiased sentence pairs mined from @Wikipedia #NLProc.
Can #NLProc reduce bias in our news and politics?.Automatically Neutralizing Subjective Bias in Text.Pryzant, @jurafsky, … #AAAI2020.Parallel corpus of 180k biased and neutralized sentences.Models for editing subjective bias out of text #Wikipedia #npov.
2
23
151
Facing difficult conversations? Interested in learning conflict resolution? 💬. Check out 🌟Rehearsal, our #LLM-empowered virtual partner to help you practice conflict resolution skills with tailored feedback #CHI24 #ConflictResolution
Before taking it out on your roommate for leaving dirty dishes out, you probably want to practice your conflict resolution skills first. Expert conflict resolution trainers, however, are EXPENSIVE. What if we practiced with an angry LLM instead? 😈 #CHI24.
6
21
144
What’s the best way to present your poster? Put it near the bar 🤓 check out our #NSF project on 🔥social AI data infrastructure🔥
4
4
129
working on #naacl2021 while attending #emnlp2020 and reading #acl2021 cfp. one day as a #nlproc-er 😂.
5
0
124
Our work "MixText: Linguistically-Informed Interpolation of Hidden Space for Semi-Supervised Text Classification" got accepted at ACL 2020 #ACL2020 #NLProc. 👉 BIG congrats to @jiaao_chen for getting his first work published as a PhD student at @ICatGT!.
1
5
124
Reduce Bias in Text? Find it out in our podcast with @robotsmarts 📢.
Can #AI be used to remove subjective bias in text like news articles? What problems arise when we alter or sanitize the message? How does AI itself perpetuate bias? We explore costs/benefits in the latest Interaction Hour w/ @robotsmarts and @Diyi_Yang!. 🔗
3
21
122
We're looking for a postdoc fellow to work on human-AI interaction around civil discourse 👉 visit our job posting and apply by Feb 15: #NLProc.
0
49
117
Don't miss out! 👉 Prof. Chris Manning @chrmanning is giving a keynote talk on 🌟Academic NLP research in the Age of LLMs: Nothing but blue skies🌟 at 2pm today #EMNLP2023 #NLProc #LLM
2
16
116
Some pic highlights of poster session for our ✨"NLP for Computational Social Science" course 😀 . Thanks everyone in the class for a great quarter 🙏 . Huge thanks to our TAs @liang_weixin & Jing Huang 🔥
5
3
112
Can we continually learn a sequence of NLP tasks by adaptively composing both newly added modules and old modules? Yes, we can 🤖. By strategically adding and reusing modules, we avoid forgetting and transfer knowledge using fewer parameters! #ACL2022 #NLProc @StevenyzZhang
3
19
111
Super excited 🤩 about this taxonomy of *SEVEN* social factors of NLP. Preprint coming soon!.
✨NEW PAPER ALERT ✨. "The Importance of Modeling Social Factors of Language" by @dirk_hovy and @diyi_yang outlines the challenges (and possibilities) in making #NLProc human-like. #NAACL2021
1
8
106
"Tweeting results in significantly more article citations over time" #AcademicTwitter
Research used randomized controlled trial to show "tweeting improves citations". What improves the long-term impact of a paper?.
0
28
101
Hello #acl2020nlp, I will be hosting a Birds-of-a-Feather meetup on 👉Computational Social Science👈 tomorrow at 11am PDT (July 6th). Please come and say hi 😀.
5
11
102
Super excited that (1) our #CHI2019 paper "Seekers, Providers, Welcomers, and Storytellers: Modeling Social Roles in Online Health Communities" received a Best Paper Honorable Mention Award - huge thanks to my coauthors #ComputationalSocialScience.
3
6
99
👇 Check out our work on a road map for using #LLMs as #ComputationalSocialScience tools!.
Can Large Language Models (#ChatGPT) transform Computational Social Science?. Our recent work shows how they might (in partnership w/ experts). We evaluate on 24 #CSS tasks + draw a roadmap 🚗🗺️ to guide #LLM-augmented social science 🚀. Paper: 🧵 thread
1
13
98
Super excited to work together with Yelena Mejova @yelenamejova and our @icwsm team! . Looking forward to a fantastic #ICWSM2022 conference #Atlanta 🔥.
We are happy to announce the @icwsm 2022 General Chairs! @Diyi_Yang and @yelenamejova . Conference dates: June 6-9, Georgia Tech Hotel and Conference Center.
1
3
95
You’re baking 🥧 and you ruin it! Do you want to shout “I’ll never make a good pie 🤬” . Self-talk can shape how we think and feel…. Maybe machines can help us look on the brighter side 😇 . We introduce this as a new #NLProc task: #PositiveReframing #ACL2022 @cjziems
3
15
93
Had a great conversation with Dr. @rastogi_ananya on building natural language processing systems that are human-centered 🌟 @NatComputSci #ALD #AdaLovelaceDay Check it out!.
We talked to @Diyi_Yang about her work on understanding human communication in a social context, building NLP systems that are human-centered, and the challenges that female researchers face in the field. 👉 🧵3/6
0
12
93
Thanks @Samsung 😀 .I'm very grateful to all my students, mentors, and collaborators 🤩 AND so excited to receive this with my friend @judyfhoffman 🎉
Trick or treat? With Halloween underway in the West, two of @GeorgiaTech's #AI experts were honored at the Samsung AI Forum in Korea. @Diyi_Yang and @judyfhoffman were named Samsung AI Researcher of the Year – GT had 2 of the 5 honorees. Congrats to both!
1
2
92
Super excited about GT NLP seminar this fall #NLProc . If you are from GT or a nearby university, check out this:
0
5
92
We will host a Computational Social Science workshop and datathon for an interdisciplinary group of scholars from the great Atlanta area 🤩. If you are interested in #computationalsocialscience, please consider apply and RT 🔥. APPLY HERE 👉
1
26
89
Check out our framework for enabling and evaluating *human-agent collaboration* 👐. by our amazing Yijia @EchoShao8899 .
LM agents today primarily aim to automate tasks. Can we turn them into collaborative teammates?. Introducing Collaborative Gym (Co-Gym), a framework for enabling & evaluating human-agent collaboration! I now get used to agents proactively seeking confirmation or my deep thinking.
1
20
91
1
16
76
Congrats to my Ph.D. student @WilliamBarrHeld for receiving an NSF Graduate Research Fellowship Honorable Mention 🎉🎉🎉 #ProudAdvisor.
1
4
86
Thanks @salesforce @SFResearch for supporting our work!.
Thank you to everyone who submitted a proposal to our third annual Salesforce AI Research Grant. We’re proud to announce our 2020 round of winners. Congratulations!! .@bluevincent @Diyi_Yang @mutembesa @danqi_chen . Read More:
4
0
85
It's bittersweet news. @ICatGT @gtcomputing has been a wonderful, collegial, and supportive department, and I have thoroughly enjoyed working here. I am extremely grateful for the support I have received from my mentors, students, and colleagues, as well as @GeorgiaTech ♥️.
3
1
82
Happy New Year Y’all 🎉 And Happy Research🤣.
Met Prof. @Diyi_Yang at Georgia tech today. She was very insistent on research progress updates for the week even during holidays. 😂😁
1
1
83
Congratulations to the team 😀 Looking forward to meeting everyone at New Orleans in person in May ✈️.
#chi2022 Best paper honorable mention!! OMG?! thanks again my wonderful collaborators! esp. @tongshuangwu @YueJiang_nj @VictoriaLinML @Diyi_Yang ! See y’all at New Orleans!
1
0
80
Interested in Social NLP 🔥🔥🔥? Check out this #nlphighlights with @pdasigi and @nlpmattg. Very fun conversation!.
#nlphighlights 119: Social NLP with Diyi Yang. We had a great chat with @Diyi_Yang about an overview of the field and some examples of the kinds of questions researchers in the field seek to answer, including some of Diyi's exciting recent work!.
0
13
73
Asking people to rank ~1K outputs takes effort 🤯. What if we can use a few demonstrations to elicit user preferences🤔 . Check out DITTO 👇 . Using ideas from online imitation learning, DITTO generates online comparison data by treating users' demonstrations as preferred over
LLMs sound homogeneous *because* feedback modalities like rankings, principles, and pairs cater to group-level preferences. Asking an individual to rank ~1K outputs or provide accurate principles takes effort. What if we relied on a few demos to elicit annotator preferences?
0
12
73
Looking forward to working with this great team!.
NAACL2022 Board election results:-Luciana Benotti (Chair).-Heng Ji (Secretary, re-elected).-Members:.Eduardo Blanco.Anna Rumshisky.Thamar Solorio (re-elected).Diyi Yang.Thanks to all candidates, the nominating committee, and all NAACL members for voting!.
1
1
73
💯 Working at @ICatGT @gtcomputing .✅ No SNOW.✅ Good places to live .✅ Modern offices .✅ Southern food.✅ Strong students.✅ Great faculty.
The Georgia Tech School of Interactive Computing is hiring tenure-track faculty. Priority areas: Machine Learning, Robotics, and Data Visualization. But all strong candidates in any area will be considered.
4
4
71
What to do if you get negative results in research, after spending three years on it? Will it be a contribution to the scientific community?.
12
4
69
For people who are joining this online, we will also post a recorded version of this tutorial 📷 that you can check out later. Internet seems not working for us 🙃#EMNLP2023.
Kicking off the #emnlp2023 #hai tutorial, with @Diyi_Yang giving an overview and @SebastinSanty talking about good/bad designs -- Glad to see a full room of people who beat jet lag!
1
4
71
🔥🎙️ Dec 6, 9:30-12:30, 📍Leo 3& 4. Come to our tutorial on human-AI interactions! 🤖
Ever surprised by users’ unexpected behaviors? Wondered why they prefer one app over another that’s powered by the same model? Come seek for answers in our #EMNLP 2023 tutorial: “Designing, Evaluating, and Learning from Human-AI Interactions”! w/ @Diyi_Yang @SebastinSanty 1/2
1
14
69
*Rehearsal* got mentioned as an example of using AI to help conflict resolution🚀 Thanks @chris_bail for the shout out 😀
GenAI could make online conversations more civil.
3
9
69
Great chatting with Chris on social #nlp 😀 Check out the podcast🎙️.
New podcast episode with @Diyi_Yang – Moving to Stanford and @StanfordNLP, linguistic and social variation, interventional studies, and shared stories and lessons learned from an ACL Young Rising Star:
0
2
67
One great aspect of giving talks at other universities is chatting with students, getting to know their work, and brainstorming together. Really enjoy this virtual visit to @ClipUmd 😀.
2
2
68
rejection, rejection, rejection . .
Academic life is characterized by rejection. Rejection is more common than adulation, admiration, respect, or deference. If you can’t stand being rejected, then you won’t enjoy trying to make a living as an academic. Most careers never face constant peer review.
0
3
64
Congrats to our undergraduate @Yihe_Oscar_Liu on being selected for Honorable Mention for the CRA Outstanding Undergraduate Researcher Award 🎉🎉. 🚨👉Oscar is applying to graduate school this year, co-advised by @asbruckman and myself - keep an eye out for Oscar’s application.
0
1
66
🚨 PrivacyLens evaluates #LLMs' privacy awareness using contextual integrity theory! . 🔍 It goes beyond static QA to detect privacy leaks 🔐 during LLMs agents' task execution process! . Overall, a multi-level approach to evaluate #privacy awareness of LLMs in action ✨.
Language models today are (1) widely used in personalized contexts and (2) to build systems that interface with tools. Do they respect privacy when helping with daily tasks like emailing? Introducing PrivacyLens to evaluate if LMs know privacy norms in action at inference time!
1
8
65
20-min walking to my classroom for in-person teaching in #Hotlanta around 3pm is . a good amount of exercise 🤖.
2
0
64
SOTA hate speech models fail on Latent Hatred 🙃🤖.
Implicit hate can easily go undetected. In our #EMNLP2021 paper (w/ @mai_elsherief, @Diyi_Yang, @DMuchlinski, @munmun10), we present "Latent Hatred: A Benchmark for Understanding Implicit Hate Speech" ⚠️(may contain upsetting examples). [1/6]
3
6
61
Prospective Ph.D. Applicants: 👉 please check out this Graduate Application Support from @gtcomputing . Please help RT 🤖.
@gtcomputing is helping underrepresented Phd applicants in CS/HCC/ML/robotics and related areas by matching them with current grad students who provide feedback on their applications. Applicants are invited to fill out this form 📝by Nov 15:
0
17
61
A great visit at Bloomberg today!.
Our CTO #DataScience Speaker Series welcomes @Diyi_Yang - who's joining @gtcomputing 🐝 - to @Bloomberg HQ in NYC. She'll speak with our #AI engineers about ways to use #NLProc, #ML & social science to understand & measure human communication to build better online communities.
2
2
58
Excited 🤩 Looking forward to chatting more about #computationalsocialscience #NLProc.
Super Stoked to have @Diyi_Yang from SALT Lab give a talk at Columbia NLP seminar this week :) The first time I read her work on computational humor, I was fascinated. Recently we have had the privilege to read a lot of her excellent work on Persuasion, Bias in NLP etc. #NLProc
0
3
58
Like Sketching? 🤩 Using #Sketch2Code, VLMs can help aid the conversion of rudimentary sketches into webpage prototypes!.
Sketch2Code: Evaluating Vision-Language Models for Interactive Web Design Prototyping.@RyanLi0802 @Diyi_Yang . Paper: Project Page: Demo:
1
8
56
So proud of my student @CamilleAHarris! Come talk to her at #CSCW2023 🔥🔥🔥.
I'm thrilled to share that our #CSCW2023 paper has been awarded Recognition for Contribution to Diversity and Inclusion 🥳! Our work explores the experiences and challenges of Black TikTok creators and affordances that support inclusive online platforms:
0
5
58
Looking forward to it!.
This Thursday, @Diyi_Yang from @gtcomputing will give a talk at our weekly #MachineLearning and Friends Lunchtime seminar series, on how to model human communication to build better social systems. #MLFL
1
1
56
Check out ToTTo, our new ''controlled'' table-to-text dataset with 140k examples, can help generate output that are both natural and faithful to the input. #NLProc 👉
#NLProc Today, we are releasing a dataset called ToTTo, containing approx. 140k examples of table and text pairs, intended for a “controlled” generation task. Data: preprint: 1/.
1
3
55
Summarizing conversations via discourse and action graphs. Talk to us at #NAACL2021 session 4D 😀.
Our recent paper @Diyi_Yang @ICatGT @mlatgt, “Structure-Aware Abstractive Conversation Summarization via Discourse and Action Graphs” will be presented in NAACL 2021 #naacl2021 #NLProc. Preprint: Code:
0
1
54
Pop-ups for AI agents can lead to 87% attack success rate 🤖💥
Humans sometimes get distracted by pop-ups… but for AI agents, it’s worse!. Pop-ups explicitly designed for agents can make them click 87% of the time, majorly derailing their tasks. @taoyds @Diyi_Yang .
1
6
51
Our SALT lab meeting today 🤩 Happy end of the semester to everyone 🔥. with Sublime Doughnuts 🍩 and a beautiful lime pie 🥧made by @katexlh ♥️
2
1
51
Congrats to my awesome collaborators for winning Best Paper Honorable Mentions at #CSCW2021 🎉🎉🎉. Check out the below summaries 🤖 .
Platform bans of offensive influencers have been in the news a lot recently. While much conversation has focused on the ethics of deplatforming, the preceding questions about its effectiveness and what happens in its aftermath have remained under-explored.
0
1
48
Due to #visa issue, I'm unable to attend #chi2019. Huge thanks to @cmuhcii Jodi Forlizzi who will help present our paper about #socialroles on Mon. 2pm, May 6th. @josephseering will present our work on #selfdisclosure on Thur. 11am, May 9th. Please come and support them 🙂👍.
0
5
47
We're working on 🌟customizing LLMs🌟 Interested in trying out our early version and sharing your feedback? Fill out this form!
Tired of your language model "delving" into things? Or maybe you like delving! We're working on a new interaction & method to customize language models, and we're looking for participants! If you're interested, please fill out the Google Form below.
0
4
48
Who wants to miss #stories 😉 Join us at the Story Shared and Lesson Learned Workshop at #emnlp2022 Happening now . at Captial 12B/Zoom
0
6
47
Check out our Georgia Tech #NLP Seminar 🔥 with an amazing line-up of speakers!!.
Ecstatic with the fall schedule .@The_Amool. and I have put together for the Georgia Tech Natural Language Processing seminar series!. Starts next Friday, and runs bi-weekly from 12-1 PM EST. to sign up for the mailing list and add it to your calendars!
1
3
47
Check out 2nd Annual Data Science for Social Good Workshop (#DS4SG) at Georgia Tech @gtcomputing @GeorgiaTechISyE . Who should attend: #undergraduates considering graduate school in related fields. Application due 10/15:
2
13
45
Summarizing conversations using conversation structures #NLProc ✍️✍️✍️ 🔥🔥🔥.
Our recent paper @Diyi_Yang @ICatGT @mlatgt , “Multi-View Sequence-to-Sequence Models with Conversational Structure for Abstractive Dialogue Summarization” got accepted by EMNLP 2020 #emnlp2020 #NLProc . Preprint: Code:
2
0
45
5
2
40