Yixin Nie
@EasonNie
Research Scientist at @AIatMeta @meta | ex Ph.D. at @uncnlp @unccs | ex M.S. at @uchicagocs @uchicago | Past Intern at @AIatMeta @microsoft | #AI
Joined February 2015
Very happy to share our recent work on vision-language transformer representation learning from videos and audios without assuming the prior existence of text. #NeurIPS2022 It has been a fulfilling and enjoyable experience working with @ZinengTang and @jmin__cho.Excited to share our #NeurIPS2022 paper! TVLT: Textless Vision-Language Transformer TVLT learns compact VL transformer representation from only visual👁 & audio👂signals w/o text arxiv.org/abs/2209.14156 github.com/zinengtang/TVLT @ZinengTang* @jmin__cho* @EasonNie* @mohitban47 🧵- Our work on addressing contradiction in open-domain dialogue modeling will be live on #ACL2021NLP (23:00 UTC 7:00pm ET Aug2 Monday Session: 4B). Paper: arxiv.org/abs/2012.13391 GitHub: parl.ai/projects/contr… ACL Video: underline.io/events/167/ses…Happy to share our new paper on addressing contradictions in dialogue modeling. arxiv.org/abs/2012.13391 We introduce DialoguE COntradiction DEtection (DECODE) task and a new dataset with contradictory dialogues to study how well NLU models can capture consistency in dialogues.
- Happy to share our new paper on addressing contradictions in dialogue modeling. arxiv.org/abs/2012.13391 We introduce DialoguE COntradiction DEtection (DECODE) task and a new dataset with contradictory dialogues to study how well NLU models can capture consistency in dialogues.
- ChaosNLI: What Can We Learn from Collective Human Opinions on NLI Data? Come join our #EMNLP2020 Q&A on Nov 18 at 6-7pm EST/11-12pm UTC @XiangZhou14 @mohitban47 @uncnlp EMNLP page: virtual.2020.emnlp.org/paper_main.154… video: slideslive.com/38938929 paper: arxiv.org/abs/2010.03532Excited to introduce *ChaosNLI* to study Collective HumAn OpinionS on NLI data! (at #emnlp2020, with @XiangZhou14 @mohitban47) We recollect 100 labels per example for 4.5K+ examples in SNLI, MNLI, & Abductive-NLI (English) to approximate label distributions. 1/4
