As many of you might already know, I had this dream to become a faculty. Unfortunately, that did not happen, so instead I decided to create my own lab kind of a thing.
I am trying to start working with some students (PhD, masters, undergrad are all welcome) and mentor them.
I'll be presenting our work "Evaluating Differentially Private Synthetic Data Generation in High-Stakes Domains" at EMNLP 2024 (Findings)!
Paper: https://t.co/FFqEQFV0DD
🚨New preprint🚨
We propose FLIRT: Feedback Loop In-context Red Teaming that uses in-context learning in a feedback loop to red team a target model. FLIRT updates its in-context examples based on the output generated by the target model and feedback received from a classifier.
🥳 🥳 I am honored to receive the 2022 @Google Ph.D. Fellowship in #NLProc! I want to take this opportunity to thank my advisor @mohitban47 for all his support and guidance, my labmates at @uncnlp @unccs, prior mentors (@mishumausam & internship advisors), and past collaborators.
🎉Excited to call for applications to ACL Year-Round Mentorship 2022! Welcome all #NLProc students worldwide to apply as mentees https://t.co/jmUJwcS07N
*We are inspired by mentoring sessions at @aclmeeting@emnlpmeeting@naaclmeeting and try to extend it to all year round🥰! /1
Checking out the @OpenAI#dalle2 system card & seeing examples of homogenous people? Check out the project me & others did years ago to exactly address the issue. We even use the same "construction workers" as an example 🥰 (Fig 1)
https://t.co/ltLDU4m3RO
https://t.co/Zfrt5XE2Wl
Glad to be part of this red teaming effort & help evaluate potential risks & limitations of powerful text-to-image generation models such as DALLE-2 — check out this detailed “system card”: https://t.co/fsef4Zm2dW
cc. @AbhayZala7@mohitban47@Miles_Brundage@_lamaahmad@jmhessel
📢7th ACL mentoring session: How to become a professor📢
⏲️April 6, 17-18 UTC / 1-2pm EST⏲️
Mentors:
▶️Yezhou Yang (Arizona State University) @Yezhou_Yang
▶️Kiran Garimella (Rutgers University) @gvrkiran
▶️Laura Burdick (University of Michigan)
See you soon!
Happy and honored to receive the 2022 Apple Scholar in AI/ML PhD Fellowship! Big thanks to my advisor @mohitban47 for all the support & mentorship. This award is not possible without the help from my labmates at @uncnlp, collaborators, & mentors, from whom I learned a great deal!
I am very honored to receive a @Bloomberg Data Science Ph.D. Fellowship!
Many thanks to my advisor @mohitban47 for all the support and guidance I received. I’d also like to thank all my labmates @uncnlp, collaborators, and mentors for their help all along!
We conclude that current LMs do not truly understand what content they consume is biased & how it is biased & we hope this will encourage future work on enhanced self-supervision during pre-training toward building LMs that are aware and adept at handling bias.
Excited to share our new work: "Analyzing the Limits of Self-Supervision in Handling Bias in Language"!
(diagnosis, identification, extraction, rephrasing)
Paper: https://t.co/6P8YQFcAor
w. karthik_g @gspandana yangliu @mohitban47@dilekhakkanitur
(@amazonscience @uncnlp)
We find that LMs are capable of performing these tasks to widely varying degrees for different types of bias. We observe internal model biases & tasks with high description sensitivity, tasks where descriptions are not useful, and tasks that models struggle on.
Incredibly honored to be recognized on the 2022
@Forbes 30 under 30 list in Science!
✨*HUGE* thanks to my collaborators and mentors,
@mulix, @anas_ant, @gneubig
✨More about my recent work: https://t.co/Qbum8a2qvL
@ForbesUnder30#ForbesUnder30
https://t.co/xfdWhMffXP