The ND HCI crew at #chi2024!! So excited to have such a big group of talented researchers from diverse areas of HCI representing Notre Dame CSE

Super interesting work by Tuhin Chakrabarty. It is kinda shocking yet reasonable to see how genAI performs poorly in creative writing. Thought: rather than relying ai for generating creative content, it makes more sense to use it atomically as a tool to refine/inspire/…

Location changed to Exhibit Hall II from 316A due to an issue in the room!

from Hawaii shirt to cap and gown: 1) #CHI2024. chatting with new people & old friends. 2) Commencement. Graduated from University of Notre Dame. What a crazy week…

[Please RT] We are looking for {grad students, researchers} who will be giving a {conference, defense, or any academic} talk soon, and planning a rehearsal! Participate in our study and receive a $20 Gift Card. Sign up here: ucsd.co1.qualtrics.com/jfe/form/SV_di…

This is the 5th SV lottery that I joined since 2022 and finally got in…😶‍🌫️

Excited to see Ningzhi Tang presenting our latest work at #VLHCC2024! Interested in how developers validate and repair LLM-generated code & how the provenance of code (AI-generated or not) affects it? Check out our paper📄! arxiv.org/pdf/2405.16081

New vision-to-language models can provide detailed image descriptions on-demand but are context-free. How can we make them context-aware? Check out our #ASSETS2024 paper: “Context-Aware Image Descriptions for Web Accessibility” 🔗ananyagm.com/context-aware-… 📄arxiv.org/abs/2409.03054…

Sangho is THE BEST mentor and collaborator! As a mentor, he’s always down to answer any questions from students and to offer advice. As a researcher, he asks valuable questions and produces top-notch works. Check his post out!!

Vision language models that power apps like BeMyAI now give really long responses to Q's about images ("What is this?", "Can I wear these together?"). What is in these long answers? Are the long answers useful? Check out Mina Huh's 🎉Oral Spotlight🎉 talk today at #COLM2024!

One of my favorite papers of this year’s UIST! Karl Toby Rosenberg is very cool! He’s on market!

📢After years at Notre Dame CSE with amazing advisor Toby J. Li😺 (he/him), I'm on the job market now! My research lies at developing interactive AI-powered systems that provide adaptive, context-sensitive support in cognitive tasks. Please learn more about my work here: zhengzhang.me

Join us!

Recent AI models can suggest endless video edits, offering many alternatives to video creators. But how can we easily sift through them all? In our #CHI2025 paper, we propose VideoDiff, an AI video editing tool designed for editing with alternatives.

I’m hiring students who are interested in multimodal generative AI / UI agent-related topics. My current vision is that we should have human-in-the-loop controllable UI generation models that adapt UIs to the diverse needs of designers and users + enhance human creativity.

Last week, we visited the Texas School for the Blind and Visually Impaired (TSBVI) and introduced AVscript to their Film and Media class. 🎥🎬 Students tried editing videos with our tool!

Solving complex problems with CoT requires combining different skills. We can do this by: 🧩Modify the CoT data format to be “composable” with other skills 🔥Train models on each skill 📌Combine those models Lead to better 0-shot reasoning on tasks involving skill composition!

LLMs trained to memorize new facts can’t use those facts well.🤔 We apply a hypernetwork to ✏️edit✏️ the gradients for fact propagation, improving accuracy by 2x on a challenging subset of RippleEdit!💡 Our approach, PropMEND, extends MEND with a new objective for propagation.

The #DynamicAbstractions Reading Group returns this Friday! 🎉 📍Topic: Simulations and Understanding 🗓️ Date & Time: June 27, 12pm EST / 9am PST 🎙️ Speaker: Andy Matuschak (andymatuschak.org) ✉️ Full letter: buttondown.com/dynamic_abstra… 🔁 Please #RT to share! 🧵 More:

Happy to share that EvalAgent has been accepted to #COLM2025 Conference on Language Modeling 🎉🇨🇦 We introduce a framework to identify implicit and diverse evaluation criteria for various open-ended tasks! 📜 arxiv.org/pdf/2504.15219