Computational Language Understanding Lab at University of Arizona
Tucson, AZ
Joined January 2016
- Tweets45
- Following13
- Followers102
- Likes1
Huge congrats to @asiful109 , @eduardo_nlp , @msurd on winning the @nvidia Academic grant!🥳
Md Asiful Islam is a fourth year PhD student advised by @msurd, working on LLM Safety and Alignment at @LabCLU @uarizona
ALT Asiful at his desk holding NVIDIA DGX Spark hardware device, with computer monitor displaying matching DGX Spark branding in background
Interpreting Answers to Yes-No Questions in Dialogues from Multiple Domains
underline.io/events/458/sche…
Best of Both Worlds: A Pliable and Generalizable Neuro-Symbolic Approach for Relation Classification
underline.io/events/458/sche…
My Heart Skipped a Beat! Recognizing Expressions of Embodied Emotion in Natural Language
underline.io/events/458/sche…
Generating Uncontextualized and Contextualized Questions for Document-Level Event Argument Extraction
underline.io/events/458/sche…
RobustSentEmbed: Robust Sentence Embeddings Using Adversarial Self-Supervised Contrastive Learning
underline.io/events/458/sche…
Improving Toponym Resolution by Predicting Attributes to Constrain Geographical Ontology Entries
underline.io/events/458/sche…
Semi-Structured Chain-of-Thought: Integrating Multiple Sources of Knowledge for Improved Language Model Reasoning
underline.io/events/458/sche…
clulab retweeted
We (w/ @msurd) will be presenting our Spotlight paper at #ICLR2024 in Vienna next week. Drop by for some amazing intellectual exchanges on data contamination.
Paper: arxiv.org/abs/2308.08493
Code: github.com/shahriargolchin/t…
Media: thenewstack.io/how-to-detect…
Congratulations, @msurd !
I'd like to congratulate my colleague, Mihai Surdeanu @msurd, on his promotion to full Professor at U.Arizona! So glad to see Mihai get the recognition that he deserves for his excellence as both a researcher and teacher. (And he's an amazing colleague too.) Congrats, Mihai! 🏆
3/ Zijie (@Zijie_Wang_NLP) explored strategies to interpret answers to yes-no questions in many domains.
@NAACL24 Findings: arxiv.org/abs/2404.16262
This work is a follow-up of his previous work interpreting answers in many languages: aclanthology.org/2023.findin…
2/ Nayem (@nayemu11) and Enfa (@beingenfa) have been busy asking and answering questions to extract event arguments.
@LrecColing 24: arxiv.org/abs/2404.16413
#NAACL24: arxiv.org/abs/2404.04770
This work is part of an interdisciplinary project lead by Steve Corman from @asuhdshc.
MohammadHossein Rezaei (@mhrezaeics) will be an undergraduate visiting research intern this summer with Diyi Yang (@Diyi_Yang) at @Stanford.
By the way, both @robert_nlp and @ShahriarGolchin are on the market! See some of their papers in our earlier tweets.
clulab retweeted
🔗 Preprint: arxiv.org/abs/2404.07544
LLMs like GPT-4 or Claude 3 can perform both linear and non-linear regression surprisingly well when given (input, output) examples in their context, without any parameter update.
(1/n)
And one more paper from @robert_nlp and colleagues: LLMs can do complicated regression tasks using only in-context examples: arxiv.org/pdf/2404.07544.pdf
5/ Fahmida, Asiful, and @robert_nlp introduce a new meta dataset for few-shot relation that expands the few-shot TACRED dataset proposed by @yoavgo and team (LREC-COLING 2024): arxiv.org/abs/2404.04445
