Hui Wei

@feiliu_nlp Have a great time in Vienna 🥳

Day 1 at #ACL2025 in Vienna done! Presented both an oral (Session 3, 2-3:30pm) and a poster (Session 5, 6-7:30pm). So nice catching up with old friends and meeting new ones. Grateful for all the conversations. Excited for the rest of the conference! #ACL2025NLP

@feiliu_nlp Thanks for presenting our PlanGenLLM paper 👏👏

@chhaviyadav_ @kamalikac Congrats, Chhavi!🥳

Successfully defended my thesis today! Extremely grateful to my amazing advisor, @kamalikac, whose unwavering support, insights, teachings and motherly love made all this possible.. Even the darkest of days seemed lighter & hopeful after spending just 5 mins with her! It has been my absolute privilege to be her student and become a mini Kamalika along the way. I only hope I can make her proud! Also a big thanks to my committee members, #SanjoyDasgupta, @BergKirkpatrick & #TaraJavidi! A bit about my research -- my research has been about designing Trustworthy AI solutions but with real-world incentive structures in mind. While AI holds great promise for societal impact, it also poses serious risks—from reinforcing bias to compromising privacy. The Trustworthy AI literature offers numerous solutions for desirable properties such as fairness, explainability, privacy and so on. Yet, many of these solutions fall short when deployed in the real-world since the real-world is rife with misaligned incentives between stakeholders (model developers, data providers, customers etc.) in the AI pipeline. My research aims to develop end-to-end Trustworthy AI solutions that account for these realities. I (1) examine how existing methods fail under incentive misalignment and (2) design end-to-end robust solutions using cryptographic approaches which account for said incentives. Together, my research calls for a rethinking of what it means to be truly Trustworthy in a world shaped by conflicting incentives. Defense Slides 👇

Email: shfang@augusta.edu Person Website: shiwei-fang.github.io School Website: augusta.edu/ccs/

If you are interested in joining his research group, please feel free to contact him via email with your CV and a brief description of your research interests or specific projects you would like to explore.

[PhD Opportunities in Low-Power, Low-Cost, and Ubiquitous Systems Research!!] My friend, Shiwei Fang (shiwei-fang.github.io), Assistant Professor at Augusta University in Augusta, Georgia, is hiring PhD students in his lab!

@HealthAI_Lab @CHILconference @nutritionorg @ADA_DiabetesPro Thanks for such an awesome talk on CHIL!

Check out our ICLR paper about contrastive learning on time series data! The work is led by amazing @maxxu05 who will present tomorrow (05/07/2024) at Poster Session 2 @ 4:30 PM as Poster #156!

@IDeznabi @manningcics Congratulations, Iman!

🎉 Excited to share that I've been awarded the "Thesis Writing Fellowship" for Spring 2024 from @manningcics ! This marks a significant milestone in my academic journey, fueling the final phases of my PhD journey. Grateful for the recognition and support.

What is supposed to do when you find official implementations by the paper authors are different from what they claim in the paper? Trust the paper and change the codes, or just trust the codes and ignore the paper?

@narges_razavian Thanks for the awesome guidance and being such an excellent advisor, Narges!

This work was done by my amazing MSc student (now a PhD student at UMass) @HuiWei15 and our awesome collaborator Arjun V Masurkar. Great Job, Hui!

We have a paper on accuracy of dementia subtype diagnosis (Alzh vs Lewy body) by clinicians (vs autopsy). The accuracy is terrible. Among those diagnosed with AD in clinic, 32% have Lewy bodies (undiagnosed). Mixed AD+LBD sensitivity is just 3%. frontiersin.org/articles/10.33… 1/5

@narges_razavian Congratulations Narges and enjoy your “vacation”!

Has somebody used advanced learning rate scheduler (e.g. cyclic learning rate, one cycle learning rate) with adaptive learning rate optimizer (e.g. RMSprop, Adam)? If so, how is the performance? Does the scheduler break the internal tracked learning rate in those optimizer? 🧐🧐

@feiwang03 Congratulations 🎉

Thrilled to be inducted into #FAMIA class of 2021

@JustinDomke Thanks for giving us such an amazing course! Learnt a lot from it!

Just finished teaching a graphical models course. This was definitely my favorite result. ("Look at probabilities and utilities working together in harmony!") But somehow, the proof is too easy to be satisfying...