Andrew Ng sees an eternal springtime for AI

Former Google Brain leader and Baidu chief scientist Andrew Ng lays out the steps companies should take to succeed with artificial intelligence, and explains why there's unlikely to be another "AI winter" like in times past.

"We may be in the eternal spring of AI," says Andrew Ng, a luminary in the field of machine learning.

Ng, a co-founder and former director of Google's AI team, sat down for an interview with ZDNet to discuss his just-published "playbook" for how to use the technology, which is available as a free download.

He dismissed worries that artificial intelligence technology may be entering another one of its periodic "winters," when interest, and funding, drops off sharply.

andrew-ng-explains-playbook-for-ai.png

Andrew Ng explains the five principles of his "Playbook for AI."

Machine learning, in the form of "connectionist" theories that model computing loosely along the lines of neurons in the brain, has gone through boom and bust cycles, flowering initially with Frank Rosenblatt's "perceptron" in the late 1950s, cooling in the late 60s, emerging again in the late 1980s only to again fall out of favor, and now suddenly back in vogue in the last several years. Those periodic coolings have been termed an "AI winter."

Also: AI startup Petuum aims to industrialize machine learning

The current renaissance raises the specter of a multi-year cooling again, but Ng says, "I don't think we're in for another one."

"The earlier periods of hype emerged without much actual value created," observes Ng, "but today, it's [AI] creating a flood of value, I've seen it with my own eyes, I've built some of those systems, I've seen the revenue being generated."

"Lots of industries go through this pattern of winter, winter, and then an eternal spring," Ng says of other fields of endeavor in human history. "We may be in the eternal spring of AI."

To try and help make that be the case, Ng's 11-page manifesto, "The AI Transformation Playbook," offers ideas for institutions, especially corporations, that want to implement AI but don't have the resources of Google or Facebook or Baidu. (Ng was chief scientists for a time at Baidu and built up its substantial machine learning efforts.)

Also: Can IBM possibly tame AI for enterprises?

"Lot of CEOs feel an urgency because they've seen the rise of the internet disrupt many areas, and they believe AI could be the next wave to disrupt their own industry," Ng says.

"Contrary to what one might expect, helping your company is not random, it can be a repeatable playbook," Ng says.

He offers five basic steps, in the order in which they should be pursued:

  1. Pilot a few AI projects within an organization to build momentum. When I was at Google, there was a lot of skepticism about deep learning. My first customer was the speech recognition group. Making progress with them helped other teams at Google gain faith in what Google Brain had to offer. The second customer we had was the Google Maps team, and so on, building momentum within the organization to think about AI.
  2. Use your own your own internal team, don't outsource the work.
  3. Develop training to bring vision on the matter to the leaders of the company, so they know how to manage the products of AI within the enterprise.
  4. Form an AI strategy, which will inevitably be industry-specific. Some people think this should be step number one, but I resist that. I've seen how in some large companies, you form a strategy and then get a budget, and because companies are not yet familiar with AI, you end up with academic strategies, where you accumulate a lot of data, often stuff grabbed from the headlines, and strategies like that don't work. You first have to have some examples from step one, you have to do the due diligence.
  5. Communicate the strategy to employees so fears about job displacement and the like will be dispelled, etc.

Also: Google says 'exponential' growth of AI is changing nature of computer

The key, he says, is that it's not just one product, even for giants such as Google and Baidu. Instead, "they have a capability to drive a sequence of AI products, from advertising to Web search to computer security, that ultimately impact the bottom line."

Pursuing this course, Ng expects companies can use AI to contribute "major drivers" of global GDP growth.

Asked whether older companies can ever manage to compete with startups that have been "AI first," such as logistics company Uber, Ng replies, "It sounds like you've drunk the Silicon Valley KoolAid."


Must read


"The common view is that what happens in tech is, it's always the startups that win. But although Google and Facebook, and Amazon and Baidu won, nevertheless Microsoft and Apple did well as well. Who would have thought that Microsoft, a company from the era of desktop software, would become one fo the AI great companies?"

"I think with the rise of AI, there will be great startups created, but also incumbents that will do a good job migrating."

Ng's Playbook is being released online by a company he co-founded and runs as CEO, Landing AI, which contracts with companies to help them follow the steps he's outlined. Ng is also the founder and general partner of the AI Fund, which works with inventors to build AI companies from scratch. He's also the founder of deeplearning.ai, which runs courses on the topic.

Previous and related coverage:

What is AI? Everything you need to know

An executive guide to artificial intelligence, from machine learning and general AI to neural networks.

What is deep learning? Everything you need to know

The lowdown on deep learning: from how it relates to the wider field of machine learning through to how to get started with it.

What is machine learning? Everything you need to know

This guide explains what machine learning is, how it is related to artificial intelligence, how it works and why it matters.

What is cloud computing? Everything you need to know about

An introduction to cloud computing right from the basics up to IaaS and PaaS, hybrid, public, and private cloud.