Looking for a way to train a model to recognize certain objects in a video motion detection system but lack enough training images to start from scratch? Try transfer learning.
With transfer learning you start with a pre-trained convolutional model, replace the last layer, then train it on the training images you do have. Results can be spectacular.

Training convolutional neural nets can be expensive and time consuming. With transfer learning businesses can add object detection capabilities to their software quickly and at low computing and data cost.

A good introduction to this technique is available at https://www.kaggle.com/dansbecker/transfer-learning

In order to compete in today's healthcare market, dentists running private practices need to also think of their practice as a business, know the channels by which their business obtains paying customers, then cultivate those channels.

For specialists like endodintists, the vast majority of their patients are referred to them by general dentists. Orienting their business focus towards improving their customer service to general dentists helps specialists build their primary sales channels and stand out in an increasingly competitive market.

AI may (or may not) turn out to be an evolution, but won't be a revolution. Blockchain is the real game changer, and this article from MIT does a good job explaining why.

Although the agile manifesto promotes the idea that "Continuous attention to technical excellence and good design enhances agility", there's been a movement in agile project management away from up-front software design. Inadequate up-front design, however, can reduce agility, resulting in a mess that is increasingly expensive to correct as a project progresses.

Up-front design can go a long way to improve agility. Since it isn't popular at the moment, make sure it is happening in all your projects so you can stay nimble down the road.