Opex AI Roundup — December 2018
by Gabriella Runnels and Macon McLean
The Opex AI Roundup provides you with our take on the coolest and most interesting Artificial Intelligence (AI) news and developments each month. Stay tuned and feel free to comment on any you think we missed!
___________________________________________________________________
A Many-Fold Improvement
Do you remember DeepMind? Not the supercomputer tasked with answering the Ultimate Question of Life, but Google’s artificial intelligence research group responsible for the A.I. program that beat the world champion Go player back in 2015. Since then, its members have been working on plenty of exciting new research, including a recent breakthrough related to “one of the core challenges in biology,” the protein folding problem. DeepMind’s new A.I. system, AlphaFold, has made impressive progress toward predicting the structure of a protein using genomic data. A protein’s physical shape dictates its function, and because proteins are responsible for pretty much everything our bodies can do, being able to determine a protein’s 3D structure would be a boon to potential advances in medicine.
AlphaFold’s achievement is particularly noteworthy because, in the status quo, modeling all the possible interactions of a protein’s amino acids becomes computationally infeasible very quickly. In fact, “it would take longer than the age of the universe to enumerate all the possible configurations of a typical protein before reaching the right 3D structure” — that’s even longer than the seven and a half million years it took Deep Thought to come up with the answer to life, the universe, and everything!
Mechanical, But Not Turks
When you think about global leaders in A.I., a few major American companies might come to mind: Google, Facebook, Amazon, etc. But America may not be the dominant power in A.I. for much longer. A combination of looser privacy laws, which allow greater access to personal data, and large pools of cheap labor, which enable the manual creation of massive amounts of training data, may give China the edge it needs to overtake America as the leading world superpower in A.I. over the next decade. Huge “data factories” are popping up around China, in which humans manually tag data that will be fed into machine learning models. “If China is the Saudi Arabia of data” — and if data is the new oil — then “these businesses are the refineries, turning raw data into the fuel that can power China’s A.I. ambitions.”
Help Wanted… Please, Seriously
We all know that the Harvard Business Review named data scientist as “The Sexiest Job of the 21st Century” in 2012, and we’ve been given no reason to doubt that proclamation in the interim. Artificial intelligence is an “exploding discipline,” says Science Magazine (and everybody else), and interest in the power and applications of A.I. does not appear to have slowed down one bit. This article gives an overview of the current state of A.I. careers, including the skill sets and kinds of expertise companies are looking for, opportunities to combine the rigor of academia with the fast pace of industry, and the creative recruiting methods companies are using to find top talent. The demand for talented data scientists and machine learning engineers is large and still growing — let’s hope the supply can keep up.
Transformation: More Than Meets the A.I.
You may have heard of Andrew Ng from his status as one of the world’s preeminent computer scientists, or as a co-founder of online learning platform Coursera, but he’s added to his extensive resume a new position as the chairman and CEO of Landing AI, a company he built to transform businesses with AI. Something tells me his experience with this company helped him publish his new AI Transformation Playbook.
Also drawing on his experience at Google Brain and Baidu, Ng lays out what he considers to be the five main steps to successfully opening your business up to the full capabilities of AI. His final note is possibly the most important — “Even if a shopping mall built a website and sold things on a website, that by itself did not turn the shopping mall into a true internet company.” An organization needs to not only embrace A.I., but emphasize and expand on the things that A.I. lets it do well.
Giving a Brand New Meaning to “Continuous Learning”
The foundation of deep learning is the neural network, a sequence of mathematical functions that work together to process inputs and create outputs. These networks are organized into multiple layers, each representing one of those aforementioned functions that perform a transformation of input data. But this design, focused on discrete, independent layers, tends to be less effective for continuous processes. This mode of thinking is getting turned on its head, thanks to a new approach that’s been built from the ground up to succeed in those very cases where traditional neural networks tend to miss the mark. Based on ordinary differential equations, the “ODE net” will hopefully see more practical application in the near future. They’re working on further generalizing this design, and hopefully they can… smooth out the edges on this approach soon.
Roundup-ception
Not that you’d ever need another roundup to sate your hunger for AI reading material, but we thought this one by former Quora and Netflix executive Xavier Amatriain was particularly good. His broad experience and deep expertise gives him a well-rounded view of AI, including notable events of the past year and trends that he thinks will characterize 2019 and beyond. Notably, Xavier is of the opinion that there weren’t any true breakthroughs this year, and it’s attributable to data scientists’ current inclination to apply existing approaches to practical problems in new ways instead of creating completely new techniques. Read more of Xavier’s interesting observations in the link below.
That’s it for this month’s roundup! Check back in January for more of the most interesting news and developments in the AI community (from our point of view, of course). Happy New Year!
___________________________________________________________________
If you liked this blog post, check out more of our work, follow us on social media (Twitter, LinkedIn, and Facebook), or join us for our free monthly Academy webinars.
