Aochuan Chen

📢We're thrilled to present our #CVPR23 work during the poster session #253 Thu morning. Come and see the interpretability and advancement in visual prompting! @sijialiu17 @OptML_MSU @pinyuchenTW @MITIBMLab

📢We're thrilled to present our #CVPR23 work during the poster session #233 Wed afternoon. Curious about how to enhance 2D temporal video grounding performance? Discover the power of text-visual prompting! 🚀 Project page: optml-group.com/posts/2dtvp_cv…. @sijialiu17 @OptML_MSU @IntelAI

A special shoutout to our amazing collaborators Pranay Sharma, Parikshit Ram, Mingyi Hong, @krvarshney and @sijialiu17 for their invaluable contributions and to the support from @MITIBMLab @IBMResearch. Looking forward to exploring more frontiers together! #ICLR2023

We gain the "surprise" in both the accuracy and interpretability of VP in our #CVPR2023 work! Website: optml-group.com/posts/ilmvp_cv… Paper: arxiv.org/abs/2211.11635 Code: github.com/OPTML-Group/IL…

What is the challenge of (visual) prompting (a kind of parameter-efficient fine-tuning) in the vision domain? Our new #CVPR23 paper "Understanding and Improving Visual Prompting: A Label-Mapping Perspective" makes an effort to resolve the existing label mapping issue.

[#aaai2023] Interested in bi-level optimization (BLO) and its applications to ML? Feel free to join our BLO tutorial (Room 206, 2-6pm, Feb. 7th), with Mingyi Hong and @zyh2022 This Thursday, I will also give a new faculty highlights talk on Robust ML. Feel free to join!

We will have our next @trustworthy_ml reading at noon Sept. 15 EST of the ICLR spotlight paper on "How to Robustify Black-Box ML Models?" from @damon19950223 @yao_yuguang @jia_jinghan @CodeTerminator and @sijialiu17 please register if interested: forms.gle/TfJSfP1qbxKEWQ…