Europe Needs to Shape Up
Here are five important AI stories from the week.
The U.S. and China are leading the AI race. Other players, such as Russia and India, are tier-two contenders. To compete with these AI powers, Europe has a lot of work to do. First, it needs to retain AI talent, which it so often loses to the U.S. Europe does not pay competitive enough salaries. Second, very stringent data privacy regulations are hampering access to good training data. Third, Europe is reliant on US chipmakers for its AI needs. Without progress on these fronts, Europe will be a weak contender in the global AI race.
This read explores what’s front and center for the various thought leaders in AI: Google, Amazon, Cloudera, OpenAI, and Microsoft.
Great read on the day-to-day life of the VP of Applied Deep Learning Research Bryan Catanzaro at Nvidia. His advice for people that are trying to break into AI: “Just get started and iterate quickly. You’ll find your way as long as you keep moving and keep adjusting.” This advice is similar to how a deep learning algorithm trains on data to find an optimal solution.
Generative adversarial networks (GANs) are capable of generating near-realistic synthetic data called deepfakes. While the news coverage on such deepfakes has been negative (e.g., raising alarms of fake news), the technology could also be applied positively. In medicine, GANs could generate synthetic medical images to help supplement the training set of real medical images. As the set of training data increases, the models for cancer diagnosis will become better, providing a very valuable boost to AI-enabled cancer diagnosis.
The CEO of Yoox Net-a-Porter discusses just how much data is available to data-first fashion companies and how that data is being used to deliver more personalized fashion and improve logistics.
More Stories Worth Reading and Watching…