Going deeper, with less data — Quadrant’s Generative Machine Learning (GML)

D-Wave announced their new Quadrant business unit today, and their early results with Siemens, winning first place in the CATARACTS medical imaging grand challenge.

Generative Machine Learning allows for Deep Learning with a lot less data. To address the problem of overfitting, the Quadrant algorithms construct generative models which jointly model inputs and outputs. It is like developing a mental model of the structure of the problem on the fly. It combines the flexibility of deep neural nets with probabilistic graphical models.

While they have an eye to making quantum leaps, so to speak, with these algorithms on their quantum computers, the results so far are using classical GPUs. In the future, work done here should see further speedups running on their quantum computing cloud (they are testing that now).

P.S. They open with a quote that reminds me that everybody be jonesing on rocket ships:

“I think AI is akin to building a rocket ship. You need a huge engine and a lot of fuel. If you have a large engine and a tiny amount of fuel, you won’t make it to orbit. If you have a tiny engine and a ton of fuel, you can’t even lift off. To build a rocket you need a huge engine and a lot of fuel. The analogy to deep learning is that the rocket engine is the deep learning models and the fuel is the huge amounts of data we can feed to these algorithms.”
—Andrew Ng, Founder of Coursera, Google Brain and former Head of AI at Baidu

Here’s more info and a link to their White Paper, and today’s announcement about the new business unit.