Mallard: Turn the Web into a Contextualized Prototyping Environment for Machine Learning
summary
Mallard is a Chrome extension that adds a prototyping IDE to the developer tools panel, letting users drag/drop or right‑click to acquire images, text, tables or video and convert them into JavaScript arrays for immediate use. It loads pretrained models via ML5 and TensorFlow.js (async), runs style transfer, sentiment, OCR and face‑recognition pipelines, visualizes outputs, and pushes replacements or widgets back to the page. The authors prototype nine ML browser apps, including Amazon/Twitter sentiment, OCR menu search, and real‑time face control for a Pac‑Man demo.
AI-written summary · about
chapters (8)
- 0:00Introduction to Mallard
- 1:56Challenges of ML prototyping
- 5:31Introducing Mallard extension
- 7:00Demo: Image style transfer
- 10:37Mallard workflow explained
- 11:45Case studies and examples
- 13:58Summary and future work
- 14:22Q&A session
related talks
24mThe next level apps for Web developers Machine learning on browsers 🤯
44mArtificial Intelligence and Machine Learning in the browser
41mPractical Machine Learning for Front-End Developers
36mPractical Machine Learning for Front End Developers
29mSuperpowers for next gen web apps: Machine learning
33mPrototype ML with Visual Blocks
19mML at Client Side
48mDeep Learning in the Browser