Interactive Learning and User‑Centered AI: A Retrospective
summary
Presents OCE 2024, an opportunistic composition engine that discovers local service components and assembles them using a multi-agent choreography trained with reinforcement learning and interactive user feedback. Implements three ambiguity-handling strategies—explicit disambiguation, proposing multiple assemblies, and uncertainty notification—and reports from a 2D campus prototype with 152 participants that uncertainty notification worked best while explicit disambiguation reduced usability.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:04Software composition
- 3:52Multiagent learning
- 6:25Ambiguity handling
- 8:12OCE 2024 demo
- 9:15User experiment results
- 11:48Experience feedback
related talks
39mAnticipated Future — Let's Orchestrate Intelligences
28mAgentic AI
22mAgentic AI for Engineering Leaders | Strategy & Execution Guide
16mUser Interaction Models for Disambiguation in Programming by Example
19mHypermed-AI: Where intelligence Emerges from Connection
30mKeynote: Fireside Chat
49mAI Agents: Collaboration in Real-Life Applications
19mBeyond agents: AI as a creative partner