Projects
AI for Writers
We develop tools to help writers serve their readers better without putting words in their mouths.
Publications:
- Supporting Revision with AI-Generated Views (HAI-GEN workshop at IUI 2024)
- Generative Models can Help Writers without Writing for Them (HAI-GEN workshop at IUI 2021)
- Cognitive Engagement with AI Writing Suggestions (internal poster fair 2024)
- Interacting with Conditional Distributions to Enact Desired Edits (in preparation, demo available)
Also, students are currently working on deploying the Textfocals sidebar (described in the publications above) as part of a broader study; let me know if you’re interested in participating.
Predictive Text
We study how predictive text shapes what people write and develop interfaces that help people use it differently.
Publications:
- Predictive Text Encourages Predictable Writing (IUI 2020)
- Sentiment Bias in Predictive Text Recommendations Results in Biased Writing (GI 2018)
- On Suggesting Phrases vs. Predicting Words for Mobile Text Composition (UIST 2016)
See the Writing Prototypes space for one demo that re-imagines the chatbot interface as a predictive text interface.
Educational Technology
- Moodle-Palette is a keyboard-based interface for Moodle. It also serves as a playground for various ways of incorporating privacy-sensitive AI into instructors’ workflows for content creation and feedback.
- zylabs-libs is a collection of Python shim code to help instructors develop ZyLabs questions that give students better feedback.
- Calgen: a tool for helping instructors (and eventually students) import their course schedules into their calendars.
Other Projects
- X32 channel-mover script: a Streamlit app for moving channels around on a Behringer X32 soundboard.
- Nanomixer: A digital audio mixer, with ADAT I/O and processing on FPGA and web-based control