So proud to have hooded my first five PhDs today: @tuvllms, @kalpeshk2011, @simeng_ssun, @mrdrozdov, and Nader Akoury. Now, they're either training LLMs at Google, Nvidia, and Databricks, or staying in academia at Virginia Tech and Cornell. Excited to watch their careers blossom!
Mohit Iyyer
1,125 posts
- We're releasing QuAC, a large QA dataset for studying information-seeking dialog! Download the data, read our #emnlp2018 paper/datasheet, and submit systems at quac.ai w/ Eunsol Choi @hhexiy @yatskar @YejinChoinka @percyliang @scottyih @LukeZettlemoyer #NLProc
- Really excited to return to @umdcs / @ClipUmd in January... this time, as a faculty member! It's been an amazing six years @UMass_NLP, and I'm incredibly thankful to my collaborators & mentors. Prospective PhDs & postdocs: please email/DM me, or let's chat next week at EMNLP!.@UofMaryland's CS dept is welcoming eight new tenure-track faculty members! With expertise in cutting-edge areas, they'll boost our academic & research capabilities. Welcome aboard! ๐ข๐ Read more: go.umd.edu/Hires24
- I'm looking for a postdoc at UMass! lots of freedom to define research agenda. topics of particular interest include NLP/DH applied to literary text and language generation. applicants from outside CS welcome. DM/email or find me at #acl2019nlp if interested. #nlproc
- Big congratulations to @kalpeshk2011!! It's been amazing to work with such a gifted and creative student over the past five years. Would not be surprised if Bard's long-form generation capabilities improve substantially after he joins the team :DHappy to share that I've defended my PhD thesis! A big thank you to my advisor @MohitIyyer and my committee @andrewmccallum, @HamedZamani and @YejinChoinka for all their support and feedback! Next step: I am super excited to be joining @Google Bard as a Research Scientist!
- As another NFL Sunday begins, we're happy to release the FOOTBALL dataset (github.com/jmerullo/footbโฆ) to study commentator bias: 1,455 transcripts of NFL/NCAA football broadcasts from 1960-2019, each tagged with player mentions linked to metadata e.g. race and position #emnlp2019
- Huge congrats @tuvuumass, who just became my first graduated PhD student!! He'll be starting his own group soon @VT_CS, so prospective PhD applicants interested in topics like multitask/multimodal transfer learning, or param-efficient LLM adaptation: def apply to work with him!I successfully defended my Ph.D. thesis. A special thank you to the members of my thesis committee: my wonderful advisor @MohitIyyer, @MajiSubhransu, @HamedZamani, @lmthang, and @colinraffel for their insightful feedback and advice on my research and career plans.
- Proud of @UMassCS undergrads Jack Merullo and Luke Yeh for doing a great job presenting our #emnlp2019 work on characterizing racial bias in football commentary! This was their first time attending an #NLProc conference. Also FYI: Jack is currently applying to PhD programs!
- Today I'm presenting @shufan_wang_'s #emnlp2023 work on evaluating text generated by retrieval-augmented LMs. TLDR: a significant perplexity drop does not imply better generations! For more come find me at the 2pm poster session in the East Foyer.
- Lots of recent work focuses on ๐๐ฎ๐ญ๐จ๐ฆ๐๐ญ๐ข๐ detection of LLM-generated text. But how well do ๐ก๐ฎ๐ฆ๐๐ง๐ฌ fare? TLDR: ppl who frequently use ChatGPT for writing tasks are elite at spotting AI text! See our paper for more (and congrats to @jennajrussell on her first paper!!)People often claim they know when ChatGPT wrote something, but are they as accurate as they think? Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy ๐ฏ
- UMass is hiring in NLP (and also vision / other "data science" related fields)! See our job posting at careers.umass.edu/amherst/en-us/โฆ and please feel free to contact @andrewmccallum, @brendan642, or me with any questions :) i'll also be at #emnlp2019 if you want to chat in person
- My student @kalpeshk2011 giving his first conference talk ever at #emnlp2018 on analyzing and reproducing results of sentiment models on complex syntactic constructions! tldr: people should average over more random seeds and use ELMo :) paper at arxiv.org/abs/1808.07733
- Excited that our work on characterizing racial bias in football commentary was covered by @TheUndefeated! Thanks to @big_data_kane for thoroughly explaining and contextualizing our research to a non-academic audience. We're continuing with this work, so look out for more soon!1/n I wrote about new research (by @MohitIyyer + colleagues) that uses artificial intelligence + big data to study racial bias in sports broadcasting for @TheUndefeated. theundefeated.com/features/artifโฆ
- Happening tomorrow: our Workshop on Narrative Understanding at #emnlp2024! Please stop by if you're interested in story understanding & generation :)#EMNLP2024 ๐ด is over but thereโs more to look forward to. The ๐ Narrative Understanding Workshop is happening on Fri, Nov 15 in Pearson. We have an amazing lineup of ๐ฃ๏ธ speakers and a host of interesting papers ๐ in the poster session. Make sure to swing by! #wnu2024 #NLProc















