|
Briefing
|
- Brain-Like Chip – Intel developed self-learning neuromorphic chip, called Loihi, that works like human brain with 130,000 neurons and 130 million synapses, enabling it to learn and operate based on feedback from environment
- Improved Learning – Demonstrated learning rates improvement of up to one million times compared to typical spiking neural networks
- More Energy Efficient – Up to 1,000 times more energy-efficient than general purpose computers required in typical training systems
- Benefits – Can make machine learning faster and more efficient while consuming less power as well as enable machines to become autonomous and adapt instead of waiting for software update from the cloud
- Institutional Debut – Will be shared with university and research institutions in first half of 2018 to advance AI
- Sample Use Cases – Include health monitoring systems that track heart activity and can identify when irregular heartbeats occur, as well as cybersecurity systems that can alert of abnormal activities, such as breaches, after learning what is normal from data streams
|
