Session on 'Learning to learn'
summary
Intel and Gnani.AI present concrete results: a lab car records 1.5 TB per 15 minutes of camera data, and Intel releases the IDD dataset of 50,000 frames with 34 India-specific classes that breaks models trained on KITTI/Cityscapes. Drive-assist analytics surface moving “gray spots” for city planners and work with Karnataka. Speech and NLP stacks, optimized with Intel libraries, run 40% faster and decode four streams per Xeon core (≈40 on a 10-core) in real-time.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:02AI Complete Problems
- 3:04Cognitive Systems
- 6:21Three Factors
- 8:20Data deluge
- 9:49AI will touch us
- 11:22Why AI has come of age
- 14:04Analytics with AI
- 15:10Opportunities for AI
- 23:42AI as a platform
- 24:18Indian Road Conditions
- 25:12Data Diversity
- 27:31Derived Data
- 28:48Work with Government
- 30:00Intel AI in India
- 30:48Success Story
