🚨 I am recruiting for 2 new PhD students and 1 postdoctoral researcher to join my lab at UC Berkeley for summer 2025 start dates! If you have a background in mechanical engineering, hci, and/or cognitive science + an interest in changing the future of design, please apply!

Congratulations to @jdzamfi for receiving a 2024 Google PhD Fellowship in HCI!

have you ever been playing a game like "imagine if it let me do this thing? that would be cool" what if games could like... generate the code and assets to support new behaviors, right there, live at runtime...?! with LLMs, they can! …but should they? 👀 https://t.co/N0pFh87gPJ

Also today at #UIST2024, @sh_reya will present "Who Validates the Validators?", bringing human preferences into LLM evaluation, so it's not just turtles all the way down. This has already been adopted by half a dozen companies. 2:15pm, A3. https://t.co/k5osGfgzox

Today at #UIST2024, Nicholas will present "What's the Game, then?" (🏆). We built a system for realtime behavior generation in games, and interviewed developers about the opportunities and perils of it. 9:15am, Allegheny 3. https://t.co/dbiQxpj5l4

Also today at #UIST2024, see @cvachha and team's poster on Dreamcrafter, a VR-based 3D scene editing system for radiance fields. Poster Session A, 3:10 PM - 4:10 PM. https://t.co/4q3Agk401V

Today at #UIST2024, Peitong will present UICrit. We created a novel dataset with ~3k design critiques and ratings, and show how such a dataset can improve the performance of automatic UI design evaluation. Dataset here: https://t.co/U7aivFTSr6. Talk at 9:15am, Allegheny 3.

The demo session is always a highlight at #UIST2024. If you weren't at the conference, or want to revisit your favorites, descriptions of @bobtianqiwei's demo and many others are online in the UIST Adjunct Proceedings:

📣 The Student Volunteer (SV) lottery for #UIST2024 is now open! Don’t miss your chance to be part of the amazing organizing team that makes this conference unforgettable. 🤩 🔗 Register now: https://t.co/lwHXcLZk0b 🗓️ Apply by: August 23, 2024 (AoE)

Amazing Faculty Position here at UC Berkeley! ⚡️Very short application window (16 Sept!!!) Apply now and spread the word – Assistant/Associate/Full Professor - Cluster Hire in Artificial Intelligence, Inequality, and Society https://t.co/bqcnvgzWw1

A reminder to send nominations for the @ACMUIST Lasting Impact Award for #UIST2024 by August 19th!

Hi! May I interview you? If you’re interested (and also, an artist) please fill out the form at the top of my linktree! https://t.co/HbRkYXZ0Vl I’m working on a research project, hoping to study and report on artists’ perspectives regarding the use of artwork online.

Congratulations to Cyrus Vaccha and Nicholas Jennings on two great MS exit talks this week! Cyrus will start his PhD at Princeton and Nicholas is headed to KTH for an exchange program. Details on their awesome research in due time...

Talk on Wed 10:00-10:15 in 316A. https://t.co/lb83osCSRq

[4] Design Space Exploration for Board-level Circuits: Exploring Alternatives in Component-based Design, presented by Richard Lin, augments the Polymorphic Blocks platform for circuit board design with novel design space exploration interactions.

Rambler supports gist-level manipulation of dictated text to narrow the gap between spontaneously spoken words and well-structured writing. Talk on Wed 2:45-3:00pm in 310 Lili’u Theater.

[3] Rambler: Supporting Writing With Speech via LLM-Assisted Gist Manipulation, presented by @susan_birdy and @jdzamfi, asks how we might construct better dictation-based mobile writing interfaces.

It introduces LLM-assisted prompt construction tools hosted inside a spreadsheet interface. A deployment with artists reveals strategies for TTI design space exploration. Talk on Wed 12:00-12:15pm in 323C. https://t.co/3WcWDo7FJJ

[2] Prompting for Discovery: Flexible Sense-Making for AI Art-Making with Dreamsheets, presented by @shmu_h and @jdzamfi, asks how interfaces can better support artists in prompt-space explorations for text-to-image models.

We implemented a Figma plug-in and compared automatic vs human feedback, fit with practice, different LLMs and prompts. Performance varies by type of guideline, but could usefully augment existing design practices. Talk on Tue 2:15-2:30 in 316A.