Amrita Bhattacharjee

🚨 Thrilled to announce our EMNLP 2024 paper on defending against LLM-driven social engineering attacks, co-authored with @TharinduKTS, @Amrita_Bh, and more! Check it out: arxiv.org/pdf/2406.12263 🚀 #EMNLP2024 #NLP

Introducing Mistral NeMo 12B NIM, an advanced open language model for chatbots, multilingual tasks, coding and summarization. Created by @MistralAI and NVIDIA and runs on a single NVIDIA L40S and RTX GPUs. nvda.ws/4d7cVsV #AI

garak: A Framework for Security Probing Large Language Models "We argue that it is time to rethink what constitutes ``LLM security'', and pursue a holistic approach to LLM security evaluation, where exploration and discovery of issues are central. To this end, this paper introduces garak (Generative AI Red-teaming and Assessment Kit), a framework which can be used to discover and identify vulnerabilities in a target LLM or dialog system" arxiv.org/abs/2406.11036

So honored to receive the Outstanding Paper Award for our ConDA work! Thank you @aaclmeeting ! @liuhuan @SCAI_ASU @ASUEngineering

Richard Feynman said “What I cannot create, I do not understand”💡 Generative Models CAN create, i.e. generate, but do they understand? Our 📣new work📣 finds that the answer might unintuitively be NO🚫 We call this the 💥Generative AI Paradox💥 paper: arxiv.org/abs/2311.00059

Join us this Wednesday in beautiful Bali! 🌴✨ Don't miss out on this amazing opportunity! See you there! #AACL2023 #Bali afnlp.org/conferences/ij…

We've released annotated slides for a talk titled "Evaluating LLMs is a minefield". Current ways of evaluating chatbots/LLMs don't work well, especially for questions about societal impact. There are no quick fixes. More research is needed. w/ @sayashk 🧵cs.princeton.edu/~arvindn/talks…

Our paper on Contrastive Domain Adaptation for AI-Generated Text Detection got accepted to IJCNLP-AACL 2023! 🎉 Preprint: arxiv.org/abs/2309.03992 @liuhuan #NLP #nlproc

@kayo_yin @psanfernandes @_emliu @andre_t_martins @gneubig Congratulations! 🎉

We (w/ @psanfernandes @_emliu @andre_t_martins @gneubig) won best resource paper #ACL2023NLP!!! 🥳🥳 Huge thank you to the ACL reviewers and best paper committee for this amazing recognition 🙏

In “#ReinforcementLearning Methods for #Wordle: A POMDP/Adaptive Control Approach” by @sbhambr1, @Amrita_Bh, and @DBertsekas, we describe the on-line solution for Wordle using rollout, which is within 0.4% of the optimal & solves in seconds! @SCAI_ASU #AI arxiv.org/abs/2211.10298

Glad to share that our review paper on #disinformation has been recognized as a #TopCitedArticle in Wiley's WIREs Data Mining and Knowledge Discovery 2021. W/ @Amrita_Bh @Faisal_Alatawi @tahoranazer @kaize0409 @Mansoureh_K @liuhuan wires.onlinelibrary.wiley.com/doi/abs/10.100…

You've probably seen results showing impressive few-shot performance of very large language models (LLMs). Do those results mean that LLMs can reason? Well, maybe, but maybe not. Few-shot performance is highly correlated with pretraining term frequency. arxiv.org/abs/2202.07206

We (@liuhuan, @kaize0409, @LiJundong, @nvchawla) will give a tutorial about "Graph Minimally-Supervised Learning" at WSDM 2022 on Feb 21st! If you have registered for the conference, welcome to join our session (2:30 pm - 6:00 pm MT)! #WSDM2022

Fascinating! #neuroscience lnkd.in/fNfTYnX

We're using AI to create more accurate forecasting models to predict when and where floods will occur - proud to partner w/ India’s Central Water Commission as the first region to roll out early flood warnings in Search #GoogleAI #Searchis20 blog.google/products/searc…

Women 👏 in 👏 tech 👏. Come on down- the bathroom lines are still non-existent 🚽