AI: unleashing the next wave
summary
Intel outlines concrete claims and tools: edge cameras and self-driving sensors produce 4000 GB per day and hospitals 3000 GB per week of data, requiring cloud and edge inference. Benchmarks show software+hardware wins — older stacks gain 3.5x with MKL 2017 and newer setups reach 18x with MKL plus next‑gen Xeon; frameworks include Intel MKL‑DNN, Intel‑optimized Caffe and Nervana, and hardware ranges from Xeon/Xeon Phi to FPGAs and Movidius sticks, with specialized ASICs promising ~100x training reduction.
AI-written summary · about
chapters (15)
- 0:00Intro
- 3:32Why is it happening now
- 5:35Data generation
- 7:11India
- 9:52Opportunities
- 12:43Pain Points
- 14:05Hardware
- 18:19Case studies
- 18:50Hardware and software
- 21:46Hardware for AI
- 24:41Forum
- 25:35AI Academy
- 26:27Intel Software Innovator
- 28:14Special events
- 28:33Conclusion
related talks
50mAI acceleration at the edge
24mTools & Techniques for a cost optimized Analytics & AI pipeline
50mLakshmi Narasimhan of Intel at Machine Learning Developers Summit 2019
09mThe AI stack from chip to analytics
09mTackling edge computing with Intel's AI Builders program
05mAI and ML are changing traditional business
04m4 real-world AI deployments at enterprise scale
36mSession on 'Learning to learn'