Research with Agents: Reflections and Lessons: hci.prof/blog/#reflections-r…
congrats to @YoungseungJ and team for the recently-accepted #uist2026 paper.
in contrast to automating research, we take a symbiotic approach where AI is embedded in scientists' abductive reasoning process to support their discovery of novel molecular edits.
Excited to share that HALO has been accepted to UIST 2026!🎉
💻 Paper: arxiv.org/abs/2607.18564
HALO is a human-AI collaborative system that assists medicinal chemists in abductive reasoning to generate promising molecular hypotheses in drug discovery.
6
15
2,259
Most #prompting is just… text.
We introduce OOPrompt—a design space for prompting UIs that treats prompts as structured, object-oriented artifacts instead of strings.
→ modular
→ reusable
→ explicit intent
📑 arxiv.org/pdf/2604.19114
Here’s the pipeline to build it 👇
4/ Deployment
Now generate the actual prompt.
Two options:
- standard NL prompt
- object-oriented format
Benefits:
- reuse across tasks
- adapt outputs for different formats
- manage multiple variants easily
👉 generation becomes a compilation step
1
210
congrats to @UCLAWBB! as we celebrate the 1st 🏆, we shouldn't forget the n-1 years' efforts, setbacks, & frustrations. moments like this remind us to never give up on what we believe in.
3
377
Agents are changing how we design: mockups→vibe coding
Moreover, they’re changing *what* we design: GUI→chat-native interfaces
UX design is shifting too: create GUI→help users express intent to agents
On the co-evolution of agents, UX, and UX design: hci.prof/blog/#agents-ux-des…
this scene will live rent-free in my head for some time ... #hamnet
someday ai might be able to create contents indistinguishable from those by human creators
but i am not sure i will feel the same--there's something unmatchable about being created by fellow human beings.
6
722
🤔 Can you tell if a user is overrelying on LLM by just observing how they type ⌨️, click ↖, and scroll🖱️on the UI?
📊 We took a data-driven approach & surfaced key behavioral indicators that might help mitigate overreliance on LLM.
👏 Led by @_ChangLiu @qinyizhou2024 @Frilk3
This Post is from an account that no longer exists. Learn more
🔬 Science discovery shapes the future of humanity. We have been working on interactive & computational tools to support this process. Check out @YoungseungJ 's latest work 👇
Excited to share that RAGPPI has been accepted to EACL 2026 Main!🎉 We introduce the benchmark for evaluating RAG in identifying the biological impacts of Protein-Protein Interactions in Drug discovery.
💻 Paper: arxiv.org/abs/2505.23823
📂Dataset: huggingface.co/datasets/Youn…
1
2
900
Xiang 'Anthony' Chen retweeted
My team is hiring! @Dodgers Quantitative Analysis is adding a Junior Quantitative Analyst (BS/BA early-career) + Quantitative Analyst (MS/PhD+). LA-based. Looking for depth in deep learning (spatiotemporal) or Bayesian modeling. Application links below — DMs open for questions.
Some thoughts on important problems for HCI: hci.prof/blog/#important-pro…
Xiang 'Anthony' Chen retweeted
Some personal news! 🎉 So excited to share that after finishing my PhD at @UCLA, I’m joining @AdobeResearch as a Research Scientist! I’ll be working with @dingzeyuli to build the next generation of generative filmmaking tools. 🎬✨
Huge thanks to my advisor @_xiang_chen_ and all my mentors/friends for their support during my PhD journey.
🚀 P.S. I’m looking for interns for Summer 2026! If you're interested in GenAI + film, DM/Email me!
Xiang 'Anthony' Chen retweeted
Q: What can you can learn from Yamamoto?
HYESEONG KIM: He’s an incredibly consistent player. He does the same training every single day without taking a break. There aren’t many rest days in the Major Leagues, but even on those rare off days, when I go to work out, Yamamoto is there training too. Seeing that really motivates me.
Xiang 'Anthony' Chen retweeted
Why limit ourselves to a single-threaded AI?
🧵(1/5)
I love the idea of systems that keep thinking continuously (not just prompt → answer) and collaborate with the user in a shared workspace. Proactively.
What it means? 👇🏻
