Best Practices from the Team That Built Bixby
summary
Demo shows a working Space Resorts capsule in the Bixby Developer Studio with a live simulator, a calendar input and fake reservation flow, plus UI layouts and training examples. The team claims preference learning ranks results per user by annotating model properties, and selection learning auto-chooses inputs using reusable strategies (examples: number of astronauts, Dublin CA vs Ireland) that users can correct. They demonstrate importing time and geo library capsules to parse relative dates and locations so the capsule works without custom NLP.
AI-written summary · about
chapters (7)
- 0:00Introduction to Bixby
- 5:00Space Resorts capsule tour
- 9:32Accelerating interactions
- 18:33Selection learning
- 25:27Using library capsules
- 33:46Implementation and testing
- 40:48Marketplace preparation
related talks
37mBixby Developer Studio Tour
44mNatural Language Training: Best Practices for Bixby
48mBixby 101: Getting Started
31mBuilding Bixby Conversational Experiences for Devices with and without Screens
49mBixby: A New Take on the Intelligent Assistant
42mBixby Developer Center and Marketplace
44mGame Changing Voice Assistant: Industry Perspectives
46mBecome a Leader in Voice AI: Join the Bixby Premier Developer Program