Unlocking the potential of low-power embedded devices with Edge AI software
summary
Panelists claim concrete wins: quantization and pruning push models toward binary networks, fine tuning on-device cuts memory from typical 10× training overhead toward inference‑level footprints, and small language models and agentic edge AI run on MCUs. Imagimob and the EDGE AI FOUNDATION surface toolchains and learning resources, research yields startups and demos, and the field points to 75% of data created at the edge, decentralization, lower latency and privacy as measurable drivers.
AI-written summary · about
chapters (12)
- 0:00Edge AI introduction
- 1:46Defining edge AI
- 3:13Optimizing AI models
- 5:42Decentralization benefits
- 8:50Academic innovations
- 11:41AI research challenges
- 12:25Backpropagation limits
- 14:54Low-power AI accessibility
- 17:25Simplifying research
- 18:09Future of Edge AI
- 19:56Agentic AI at edge
- 20:16Closing remarks
related talks
55mEdgeAI - AI in the Palm of Your Hand
10mThe Rise and Impact of Edge AI
48mThe Power of Edge AI: Revolutionizing Future of Real-Time, Autonomous Processing
31mInfiniEdge AI: Bringing Open-Source Generative AI To the Edge
42mThe Stakes of Green AI for Embedded Systems
36mApplied AI on the edge
29mEdge Intelligence Revolution
09mWelcome + Opening Remarks, Program Committee Members