UtahNLP

🏆ACL Reviewer and (Senior) Area Chair Awards 🔗2023.aclweb.org/program/best_r… 👏Outstanding SACs @caro__lawrence, @anmarasovic & @ChenhaoTan 🧵 (1/6)

I'm going back for a PhD! 👨🏼‍🎓 Excited to announce that I just accepted an offer to join the PhD program at @KahlertSoC at @UUtah working with the @UtahNLP group 🏔️ SO privileged to have such an opportunity so near family and within a stone's throw of the wasatch mountains 😇🚵👨‍👩‍👧

👏🔥our paper received Distinguished Paper award !

Best Papers (1/3) 📢S2: Language Grounding (Poster) ... 📌 Do Androids Laugh at Electric Sheep? Humor 'Understanding' Benchmarks from The NY Caption Contest 🔍Delve into the "understanding" of cartoons and the complexities of humor. 🤖 🔗aclanthology.org/2023.acl-long.… 🧵(4/5)

I’ll be presenting our work “Verifying Annotation Agreement without Multiple Experts: A Case Study with Gujarati SNACS” as a virtual poster on Tuesday at #ACL2023 and in-person at LAW on Thursday. Joint work with @viveksrikumar Paper: tinyurl.com/3989kvnp (1/3)

Excited to share our #ACL2023NLP (#NLProc) work on "Evaluating Inter-Bilingual Semantic Parsing for Indian Languages" appearing in @5thnlp4convai workshop. We propose IE-SEMPARSE, an Inter-bilingual Seq2seq Semantic parsing dataset for 11 Indian languages. - 1/n

Excited to share our #ACL2023NLP (#NLProc) paper on "Information Synchronization Across Multilingual Semi-Structured Tables"! 📚🔍 We explore the problem of Information Synchronization across multilingual tables and construct a large-scale dataset InfoSync. #NLPResearch -1/n

Check "Leveraging Data Recasting to Enhance Tabular Reasoning" at #EMNLP2022 @emnlpmeeting . Joint work with @aashna_jena (@iiit_hyderabad -> @GoogleAI), @manshri (@iiit_hyderabad ), @eisenjulian (@GoogleAI . Explored data recasting between tabular tasks to enhance reasoning.-1/n

Check our work at #EMNLP2022, see thread 🧵⬇️ 1. Multilingual NLI for Indic Languages twitter.com/keviv9/status/… 2. Data Augmentation for Tabular NLI twitter.com/keviv9/status/… 3. Data Recasting for Tabular NLI twitter.com/keviv9/status/… Catch @keviv9, @aashna_jena at @emnlpmeeting 1/n

Check for more details here: Page: autotnli.github.io Paper: vgupta123.github.io/docs/autotnli.… Code: github.com/Dibyakanti/Aut… Data: shorturl.at/jvMO2 Video: youtu.be/jExTxUMHZ7E Poster: shorturl.at/ntT45 more updates underway huge kudos to @DibyakantiK -3/n

Check "Realistic Data Augmentation Framework for Enhancing Tabular Reasoning" at #EMNLP2022 @emnlpmeeting. Joint work with @DibyakantiK (@IITGuwahati, @iitgai)@SoUmmYaah (@IITKgp), @imsure318(@TechAtBloomberg). We explored semi-automatic way to generate tabular NLI dataset.-1/n

We study the effects of alternative fine-tuning procedures, languages, multi-linguality, and mixed-language input on pre-trained language models. Page: indicxnli.github.io Paper: arxiv.org/pdf/2204.08776… Data: huggingface.co/datasets/Divya… Code: github.com/divyanshuaggar… -2/n (n/n)

Check "IndicXNLI: Evaluating Multilingual NLI for Indian Languages" at #EMNLP2022 @emnlpmeeting. Joint work with @divyanshuaggarw (@dtu_delhi -> AmExAI @AmericanExpress), @anoopk (@MSFTResearch). IndicXNLI is an NLI dataset for 11 Indian languages. @UUtah @UtahSoC @UtahNLP -1/n

Our @AppFolio Best Paper Award winner was "CONDAQA: A Contrastive Reading Comprehension Dataset for Reasoning About Negation" from Abhilasha Ravichander et al from CMU/AllenAI! Paper link: arxiv.org/abs/2211.00295 Thank you to sponsor @AppFolio for supporting our event! (3/4)

@aaclmeeting @areddys53 @IITGuwahati @iitgai @imsure318 @TechAtBloomberg @UUtah @UtahSoC @UtahNLP Code: github.com/abhilashreddys… Paper: vgupta123.github.io/docs/pettable.… Poster: drive.google.com/file/d/1ntE2PP… PPT: drive.google.com/file/d/1SJKS1y… -3(n)/n

@aaclmeeting @areddys53 @IITGuwahati @iitgai @imsure318 @TechAtBloomberg @UUtah @UtahSoC @UtahNLP We use Pattern-Exploiting Training to augment pre-trained language models' knowledge & reasoning abilities. Upgraded model better understands facts and tabular reasoning and prefer better against character and word-level pertrubation. Project Page: infoadapet.github.io -2/n

Check "Enhancing Tabular Reasoning with Pattern Exploiting Training" at #AACL2022 @aaclmeeting. Joint work with @areddys53 (@IITGuwahati, @iitgai); @imsure318 (@TechAtBloomberg). We exploit prompt based learning, i.e, PET, to extract knowledge for tables @UUtah @UtahSoC @UtahNLP

.@Lasha1608 did some amazing work here, I think this is one of the best datasets I've been a part of creating.

Lots of effort went into this one. I'm very excited to see it released and grateful that I had a chance to work closely with @Lasha1608 and @nlpmattg! Looking forward to hear/see what you all think 😌

🚨Help NLP models process negation🚨 Introducing ℂ𝕆ℕ𝔻𝔸ℚ𝔸, a *contrastive* reading comprehension dataset that requires reasoning about negation w/ @nlpmattg & @anmarasovic @ai2_allennlp, at #EMNLP2022 📝Paper arxiv.org/abs/2211.00295 🚀Data github.com/AbhilashaRavic… [1/8]