Build a UI that Learns - Intelligent Prefetching
summary
An actual intelligent prefetcher is built: React.lazy and dynamic imports load components on demand while a TensorFlow.js LSTM predictor consumes recent interaction sequences to estimate the next trigger. Inputs use one-hot encoding of triggers, model output is a probability over routes/components, and the app dynamically imports predicted modules to prepare them before navigation. Training uses sampled user sequences so the system adapts over time and demonstrates practical feasibility for faster perceived loads.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:50Content Context
- 2:00Prefetching
- 2:40Lazy Loading
- 3:16Question
- 5:42Training the Network
- 6:37Implementation
related talks
19mUsing the activity logs to predict user behaviour
34mPredictive preloads for ultimate web performance?
31mA Feasibility of On-Device ML for Adaptive User Interfaces in Frontend Development
25mMind Reading with Adaptive and Intelligent UIs in React
27mWhen To Fetch: Remixing React Router
43mRoute Recalculating
19mTake a load off with React
25mUsing Machine Learning To Improve UI/UX