Five takeaways from Japan’s AI Tech Leaders Summit

August 26, 2017

Some of the Asia’s leading experts in artificial intelligence mingled yesterday with a small group of corporate acquirers, startups, and investors at the first ever Japan AI Summit. Here are my five takeaways from the interactions:

5. The sovereign gap is widening

In advanced AI research, the U.S. still has the lead, but China has the momentum. Other developed nations need to accelerate or will never close the gap.

4. Data is “misunderestimated”

Perhaps the most controversial statement of the day, but the contention was that with techniques like simulation learning, massive data sets are no longer as critical. So size doesn’t matter so much after all…

3. Hardware is overrated

Robotics folks subscribe religiously to the necessity of hardware for data collection to feed AI systems; however, the consensus yesterday disputed this mindset. The prevalence of data already in existence focuses the question more on “how to use it” rather than “how to collect it.”

2. Analog is back, baby !

Ok, I’m getting ahead of myself, but I’ve been claiming for years that rumours of analog’s death had been greatly exaggerated. Well, it turns out that analog chips or mixed signal ICs will soon play an essential role in AI systems. CPU::Intel. GPU::Nvidia. Neuromorphic::X. If anyone can accurately tell me who X is, I’ll give you a medal.

1. Privacy arbitrage opportunities

The disparity in regulations on data privacy may create opportunities for “privacy arbitrage” across countries. For example, obtain access to medical records in China in order to train an AI system on a health care solution with applications globally.

Special congratulations to Dreamquark and Snips for making the trek over from France !

tags: , , ,
posted in technology, venture capital by mark bivens

Follow comments via the RSS Feed | Leave a comment | Trackback URL

Visit Us On TwitterCheck Our Feed