A Live, Multiple-Representation Probabilistic Programming Environment for Novices
summary
A live IDE for probabilistic programming built around Infer.NET and F# shows code, a graphical model view and probability-distribution visualizations that update immediately when evidence changes. A controlled study with 16 participants finds a significant reduction in keystrokes and faster task times and fewer deletions versus a text-only editor. The system compiles models slowly (Infer.NET bottleneck), doesn’t yet allow editing-by-manipulating the graph, and the prototype and code are available on GitHub.
AI-written summary · about
chapters (6)
- 0:00Introduction to talk
- 0:50What is probabilistic programming
- 2:24Differences from conventional programming
- 4:52Live IDE with visualizations
- 6:40User study results
- 8:49Q&A session
related talks
35mA Live Programming Experience
32mTotally Live Programming with Hazel (Progress Report)
22mEnhancing Liveness with Exemplars in the Newspeak IDE
38mThe Gamma: Programming Tools for Data Journalism
09mMicro-Versioning Tool to Support Experimentation in Exploratory Programming
39mPower of AI and Code Quality and Performance
14mExample-Based Live Programming for Everyone: Building Language-Agnostic Tools for Live Programming w
01h 03mMaking a desktop IDE distributed and collaborative