From Touch to Integration: UChicago CS at UIST 2026

At this year’s ACM Symposium on User Interface Software and Technology (UIST), taking place November 2-5, 2026 in Detroit, the University of Chicago Department of Computer Science returns with research that spans tangible materials, wearable haptics, and the future of text. Faculty and students are contributing four full papers (two paired with live demos), two additional demos, a poster, and two workshops.
The work ranges from magnetic particles you can shape with your hands to a ring that talks back through your fingertip. Together, these projects show how interfaces are moving beyond the screen and into our hands, skin, muscles, and physical surroundings, while asking how people can stay in control as systems grow more capable.
Here’s a look at what UChicago CS will bring to UIST 2026:
PAPER PROGRAMS
UChicago authors in bold
Granular I/O: Enabling Bi-directional Malleable Interaction for Actuated TUIs with Magnetic Particles (Paper + Demo)
Chi Wang, Clara Cui, Alan Pham, Justice T Andersen, Tucker Rae-Grant, Ken Nakagaki
AxLab– program link
Granular materials, clusters of small particles, can be shaped, split, and merged with natural hand gestures, yet they’ve rarely been used for actuated tangible interfaces. Granular I/O uses millimeter-scale magnetic particle clusters, moved by external magnetic field controllers, to create malleable material that works as both input and output. The team presents a design space for this kind of interaction, along with a full implementation covering particle fabrication, two actuation platforms, and authoring software. They validate the system through technical evaluation and demonstrate a range of applications.
Electrotactile Improves Thermal Referral (Paper + Demo)
Wen Li, Rong Ni, Bozhi Tian, Pedro Lopes
Human-Computer Integration Lab – program link
Thermal referral is a perceptual illusion in which a tactile cue makes a nearby warm or cold sensation feel as if it comes from the touched spot, letting designers use fewer thermal actuators. The standard approach relies on vibration, which struggles to produce cold sensations and creates a strong “buzz” that makes it unsuitable for non-contact events like the chill of an open freezer in VR. This paper swaps vibration for electrotactile stimulation. Across two user studies the approach increased referral for cold, boosted thermal perception while minimizing tactile sensation, and improved realism in VR, including distinguishing contact from non-contact thermal events. These results led to a set of new design guidelines for building more immersive thermal experiences.
RingIO: An Eyes-Free Smart Ring with Inputs via Touch and Outputs via Touch (Paper)
Lonnie Chien, Pedro Lopes
Human-Computer Integration Lab – program link
Smart rings are popular, with 5.5 million units on the Oura ring alone, but their small size and limited power make traditional outputs like screens or vibration motors impractical. RingIO explores electrotactile stimulation, a very low-power form of haptic feedback, as an alternative. The prototype ring uses double-sided electrodes that sense touch on one side and deliver electrotactile animations on the other, enabling eyes-free input and output in the same place. In a user study, participants recognized the ring’s animations with up to roughly ~90% accuracy, making it useful for a whole new range of applications not currently possible in today’s smart rings.
Gridlock: Plain Text with Guides (Paper)
Sam Cohen, Ravi Chugh
Ravi Chugh’s Lab – program link
Plain text underpins much of software development, from code to documentation to data files, because it works with version control, search, and countless other tools. But keeping text neatly formatted, such as lining up columns in a table, often means tedious manual edits. Gridlock is a text editor that adds movable vertical and horizontal guides, which users drag to align, group, or hide parts of a document. The guides create a spreadsheet-like grid of cells, borrowing familiar ideas like rows, columns, resizing, and clipping, while the underlying file stays plain text that can still be edited directly. The team demonstrates the approach with examples spanning code, tabular data, and ASCII diagrams.
DEMOS AND POSTERS
NodeBots: Physicalizing Network-Based Data with SwarmUIs (Demo)
Ramarko Bhattacharya, Alan Pham, Ken Nakagaki
AxLab – program link
Data physicalization, which represents data through physical forms rather than screens, is a growing area of research. NodeBots brings it together with swarm user interfaces (SwarmUIs): groups of small robots that act as the nodes of a network graph. Users can physically interact with the nodes, creating hands-on metaphors for working with graphs. The team explores the system’s interaction space and demonstrates applications such as finding sparse cuts and shortest paths.
Embodied-AI: Providing Users with Physical Assistance by Constraining Multimodal-AI with Embodied Knowledge (Demo)
Romain Nith, Yun Ho, Pedro Lopes
Human-Computer Integration Lab – program link
Electrical muscle stimulation (EMS) can physically assist people with tasks, such as shaking a spray can before painting. But today’s EMS systems are narrow: each task needs its own fixed program, and none account for context. A can of cooking spray, for example, doesn’t need shaking the way paint does. Embodied-AI takes a different approach, using computer vision and large language models to generate muscle-stimulation instructions based on the user’s pose, location, and surroundings. The AI is constrained by a muscle-stimulation knowledge base and joint limits. The result is a more general system that can support new interactions, such as opening a pill bottle, while also reproducing existing systems without task-specific programming.
StepLens: Supporting Inspection and Evaluation in Agentic Tabular Data Work to Narrow Down the Gulf of Evaluation (Poster)
Danni Liu, Muhammad Imam Luthfi Balaka, Raul Castro Fernandez
Raul Castro Fernandez’s group – program link
AI agents can answer data questions in plain language, but users need to see how an answer was produced before they act on it. StepLens is an interface built on the Pneuma-Seeker agentic system that organizes the agent’s intermediate work into structured, inspectable stages. In a controlled study with 36 participants, readiness to act clearly differed between correct and incorrect answers. Structured visibility also helped users trace results, though it sometimes made inconsistencies harder to resolve.
WORKSHOPS
Drawing on Materials and Processes: A Hands-on Workshop on Building Creative Tools with Grain Spaces
Lea Albaugh (Carnegie Mellon University), Ilan E Moyer (MIT), Tucker Rae-Grant, Anup Sathya, Ken Nakagaki
AxLab, with Carnegie Mellon University and MIT – workshop link
In this half-day workshop, attendees use the “grain space” framework to build creative tools for pen plotters. After learning what a grain space is, participants will discuss grains suited to pen plotting, define a parametric space for a Truchet tile grain, and implement interactions with it. The session also explores how HCI researchers can use this approach to guide and refine systems that work with novel computational materials.
Workshop on Augmenting Human Dexterity: From Multisensory Feedback and Motor Capabilities to Skill Acquisition
Shan-Yuan Teng (National Taiwan University), Tan Gemicioglu (Cornell Tech), Kyungyeon Lee (University of Maryland), Dishita G Turakhia (Emory University), Patricia Alves-Oliveira (University of Michigan), Jennifer Jacobs (UC Santa Barbara), Hideki Koike (Institute of Science Tokyo), Pedro Lopes, Jun Nishida (University of Maryland)
Human-Computer Integration Lab, with institutions across the US and Asia – workshop link
Wearable and mobile technology is opening new ways to augment human dexterity, going beyond traditional assistive tools to support precise, complex tasks like playing an instrument or working at a bench. Research on the topic is scattered across technologies and application areas, and UIST has not yet had a dedicated forum for it. This workshop brings together researchers from engineering, neuroscience, learning science, and other fields for talks, demonstrations, and discussion, with the goal of mapping out a research agenda for augmenting dexterity more holistically. Notably the workshop is led by two alumni of our department, Shan-Yuan Teng (who was a PhD student and is now a Professor at National Taiwan University) and Jun Nishida (who was a Postdoc and is now a Professor at the University of Maryland).
FROM TOUCH TO IMPACT
UChicago researchers at UIST 2026 are rethinking where an interface lives. Magnetic particles become a material that can be both shaped and moved, a ring senses and speaks through the same fingertip, and a thermal sensation can be created without a thermal source. Network graphs become swarms of robots that people can physically manipulate, and AI-driven muscle stimulation offers help with everyday physical tasks.
Alongside this work, projects on text editing, creative tools, and agentic data analysis ask how people can understand and steer the systems they use. The two workshops widen the lens, from building tools with new computational materials to augmenting human dexterity.
As these projects move from the lab into classrooms, studios, and daily life, they share a common thread: interfaces designed around how people actually perceive, touch, and make sense of the world.