Yue Dong @ NeurIPS 2023 (@YueDongCS) / X

Yue Dong @ NeurIPS 2023

349 posts

Yue Dong @ NeurIPS 2023

@YueDongCS

Assistant Prof @UCRiverside. PhD from @Mila_Quebec @McGillU. Trustworthy NLP+AI safety & Summarization! Former intern @GoogleAI @MSFTResearch @allen_ai

California, US

yuedong.us

Joined September 2010

Yue Dong @ NeurIPS 2023
@YueDongCS
Apr 26, 2022
Life update - I'm thrilled to share that I'll be joining @UCRiverside as an Assistant Professor in January 2023, following my PhD studies @MILAMontreal @McGillU. I'm excited to work with my fantastic colleagues and build an NLP group @UCR_CSE! yuedongcs.github.io #NLProc
Yue Dong @ NeurIPS 2023
@YueDongCS
Aug 5, 2023
My group has 2 PhD positions for Fall 2024 (email me if interested). I am open to discussing collaborations on Trustworthy NLP&AI safety; or just fun stuff in Southern California #SoCal Check out #KDD2023 SoCal Data Science Day for more of #SoCal :
Yue Dong @ NeurIPS 2023
@YueDongCS
Aug 5, 2023
Thought about driving from #UCRiverside to #LongBeach to attend #KDD2023 daily🚗 Give it up as a round trip on weekends ~2h but weekdays >4h 🔁 #LosAngeles to me is for fun, not for work🥲 Look forward to meeting familiar and new faces at #KDD2023 😊
sites.google.com
SoCal Data Science Day
Register for this special day using the link below!
55K
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 8, 2023
There are two PhD positions in my group for 2024. I am open to discuss more if you are interested in working on trustworthy NLP & AI safety.
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 8, 2023
On my way to #ACL2023NLP. It’s the first in-person conference since I become an assistant professor during the pandemic. The past year we witness a huge shift in text generation research and look forward to hearing your opinions #NLProc #ACL2023 #ACL2023nlp
61K
Yue Dong @ NeurIPS 2023
@YueDongCS
Dec 9, 2023
On my way to #NeurIPS2023. Excited to chat with old friends and meeting new ones. I am also recruiting 2-3 PhD students for Fall 24, focusing on trustworthy NLP and LLM safety, please shoot me a dm if you’d like to chat in person. More see my website:
yuedong.us
Yue Dong 董悦
Personal website of Yue Dong.
24K
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 10, 2023
Greatly enjoyed T6 by @danqi_chen's group @ #ACL2023NLP. One interesting insight: with the success of #LLMs, retrieval-based models shift towards retrieval-based prompting / in-context learning, though they still underperform full FT models. More here: acl2023-retrieval-lm.github.io
22K
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 29, 2023
Kudos to the LLM watermark team for the outstanding #ICML2023 paper - well deserved! Our follow-up work highlights the challenges of applying watermarks to conditional text generation (CTG) - severe performance hit & detection paradox. paper:
arxiv.org
Watermarking Conditional Text Generation for AI Detection:...
To mitigate potential risks associated with language models, recent AI detection research proposes incorporating watermarks into machine-generated text through random vocabulary restrictions and...
14K
Yue Dong @ NeurIPS 2023
@YueDongCS
May 6, 2021
Super excited to share that I have two papers accepted at @aclmeeting, ranging from neural text degeneration to text summarization. Appreciation to congratulations to all my amazing collaborators! @_csBhagav @YejinChoinka @GXiming @ABosselut @memray0 @ericxyuan @KhushbooThaker
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 9, 2022
I'm excited for #NAACL2022! I'll be giving a remote presentation at the tutorial on "Text Generation with Text-Editing Models". The hybrid tutorial will take place on Sunday, July 10 from 9:00am to 12:30pm (PDT). Details in text-editing.github.io
Yue Dong @ NeurIPS 2023
@YueDongCS
Jun 5, 2023
Proud to graduate with McGill friends in Montreal @koustuvsinha @bogdan_mazoure @AltakroriM
6.3K
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 9, 2022
Hope to see you tomorrow at the #NAACL2022 Text Generation with Text-Editing Models tutorial! It is held both physically in Ballroom Columbia A and virtually via Zoom from 9am to 12:30pm (PDT), with Q&A sessions from 8-8:45am and 12:30-1pm. @ericmalmi, @varphi, @alseveryn
Eric Malmi
@ericmalmi
Jul 2, 2022
A lot of NLP is moving to larger models, casting everything as generation (looking at you, T5 & co), but there's a countertrend: fast Text-Editing models, casting generation as seq labeling. Learn more at our #NAACL2022 tutorial: text-editing.github.io
Yue Dong @ NeurIPS 2023
@YueDongCS
Nov 19, 2023
🎉 Honored to receive the Best Paper Award at #SoCalNLP2023 with @Erf_Shayegani & @nael_ag (3/150+ submissions)! Our paper proposes a compositional attack across text-image modality for jailbreak that only requires white-box access to the vision encoder. 🧵1/4
8.7K
Yue Dong @ NeurIPS 2023
@YueDongCS
Jul 8, 2023
On my way to #ACL2023NLP. It’s the first in-person conference since I become an assistant professor during the pandemic. The past year we witness a huge shift in text generation research and look forward to hearing your opinions #NLProc #ACL2023 #ACL2023nlp
63K
Yue Dong @ NeurIPS 2023
@YueDongCS
Dec 9, 2023
Our paper on LLM watermarks for conditional text generation has been accepted to #AAAI2024. We discovered a trade-off between task performance and detection of watermarking LLMs for CTG task, and proposed a semantic watermark for CTG. Paper details:
arxiv.org
Watermarking Conditional Text Generation for AI Detection:...
To mitigate potential risks associated with language models, recent AI detection research proposes incorporating watermarks into machine-generated text through random vocabulary restrictions and...
5.5K
Yue Dong @ NeurIPS 2023
@YueDongCS
Feb 14, 2024
We reveal #LLMs safety alignment varies by NLP tasks: weak in summarization, risking in-context attacks. Safety-sensitive documents #ChatGPT refuse to translate can be translated if we ask for a summary first #NLProc Demo: fyyfu.github.io/SafetyAlignNLP/ Paper: arxiv.org/abs/2312.06924
fyyfu.github.io
TWITTER BANNER TITLE META TAG
TWITTER BANNER DESCRIPTION META TAG
4K