COURSE / HCI
Human-Computer Interface
Human-Computer Interface Link to heading
Course notes, interaction concepts, prototypes, and evaluations.
Practical labs: Machine learning in Unity includes trained reaching and chasing policies, training curves, and 400 recorded comparison trials. Gaussian splatting adds a real Brush training run on public posed images, recorded validation metrics, filtering, and a working Unity desktop viewer. The mobile adaptation adds touch controls and a signed Android APK with eight static checks passed. Android device acceptance is deferred.
Progress: Homework 1 covers Lectures 1-3, with illustrated interface cases and redesigns. Lab 1 documents the blog setup and the existing Unity environment. Homework 2 covers Lectures 4-6. Homework 3 includes three AR concepts, a browser prototype, and feedback from three people. Homework 4 covers Lectures 10-12. The talking-character lab has a working Unity client using the DeepSeek API, six automated API timing records, and three rated turns from participant use with the typing indicator enabled. Multiple participants shared one code, so the exact participant count is unknown and an A/B user comparison is still pending.
Work in this course
- HCI Lab: Streaming Chat in Unity with the DeepSeek API
A Unity chat client using DeepSeek, with measured streaming latency and three rated turns from an informal participant pilot.
- HCI Lab 2: Touch Controls and an Android Build
Adapting Signal Chase for a touch screen. Android device testing is deferred.
- HCI Lab 1: Machine Learning in Unity with ML-Agents
Two PPO policies in Unity, checked against random actions in 100 independent trials per task.
- HCI Lab: Gaussian Splatting from Images to Unity
A recorded Brush training run, simple splat filtering, and a tested Unity desktop viewer.
- HCI Homework 4: Phone Shapes and Everyday MR
Why phones change shape, what remains of Weiser's vision, and three everyday MR app concepts.
- HCI Homework 3: An AR Guide for Finding a Room
Three AR ideas, an indoor navigation prototype, and feedback from three people trying its routes.
- HCI Homework 2: Body Interfaces and the Ultimate Display
Pedro Lopes's research, Sutherland's display vision, and a wristband that reads muscle activity.
- HCI Homework 1: Affordances, Gestalt Laws, and Dark Patterns
Door affordances, two Gestalt problems in forms, and two dark patterns with simpler alternatives.
- HCI Lab 1b: Unity Setup with an Existing Project
Reusing the Corebound Unity project to document the environment, scenes, runtime evidence, and a UI readability fix.
- HCI Lab 1a: Setting Up the Course Blog
A Hugo Extended course portfolio, GitHub Pages publishing, and the image-path problem after deployment.