A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of Elkan (2003). We argue that, when incorporating distance bounds into a mini-batch algorithm, already used data should preferentially be reused. To this end we propose using nested mini-batches, whereby data in a mini-batch at iteration t is automatically reused at iteration t+1. Using nested mini-batches presents two difficulties. The first is that unbalanced use of data can bias estimates, which we resolve by ensuring that each data sample contributes exactly once to centroids. The second is in choosing mini-batch sizes, which we address by balancing premature fine-tuning of centroids with redundancy induced slow-down. Experiments show that the resulting nmbatch algorithm is very effective, often arriving within 1% of the empirical minimum 100 earlier than the standard mini-batch algorithm.

A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of Elkan (2003). We argue that, when incorporating distance bounds into a mini-batch algorithm, already used data should preferentially be reused. To this end we propose using nested mini-batches, whereby data in a mini-batch at iteration t is automatically reused at iteration t + 1. Using nested mini-batches presents two difficulties. The first is that unbalanced use of data can bias estimates, which we resolve by ensuring that each data sample contributes exactly once to centroids. The second is in choosing mini-batch sizes, which we address by balancing premature fine-tuning of centroids with redundancy induced slow-down. Experiments show that the resulting nmbatch algorithm is very effective, often arriving within 1% of the empirical minimum 100 earlier than the standard mini-batch algorithm.

In 1987, then-CEO of Apple Computer, John Sculley, unveiled a vision that he hoped would cement his legacy as more than just a former purveyor of soft drinks. Keynoting at the EDUCOM conference, he presented a 5-minute, 45-second video of a product that built upon some ideas he had presented in his autobiography the previous year. The video is a two-hander playlet. The main character is a snooty UC Berkeley university professor. The other is a bot, living inside what we'd now call a foldable tablet.

Imagine a world where right kind of employee is chosen for the job at a fraction of cost what we currently incur to hire someone. Imagine a workplace where employees are able to work to their maximum potential and any problem which they face can be addressed in real time and the solutions be guided with volumes of data. Imagine a world where employees won't have to face the dreaded year-end ritual of getting appraised by their seniors or peers. All this and much more is possible now with the power of AI. The future of performance management is here as artificial intelligence meets performance management.

A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of Elkan (2003). We argue that, when incorporating distance bounds into a mini-batch algorithm, already used data should preferentially be reused. To this end we propose using nested mini-batches, whereby data in a mini-batch at iteration t is automatically reused at iteration t+1. Using nested mini-batches presents two difficulties. The first is that unbalanced use of data can bias estimates, which we resolve by ensuring that each data sample contributes exactly once to centroids. The second is in choosing mini-batch sizes, which we address by balancing premature fine-tuning of centroids with redundancy induced slow-down. Experiments show that the resulting nmbatch algorithm is very effective, often arriving within 1\% of the empirical minimum 100 times earlier than the standard mini-batch algorithm.

A flight over farmlands could be part of a future Uber ride. Recently there's been a lot of ink and pixels declaring 2016 the year of humanity's discontent. That's tough to deny in the realm of geopolitics, between awful conflicts in places like Syria, acts of terror worldwide and contentious elections in the UK and US. But the world went on, and so did important work in science and innovation. If you sweep the ugly parts of 2016 under the rug and then check the place out, it's not too shabby.

While the future of the Affordable Care Act, popularly known as Obamacare, is quite uncertain, parts of the famous healthcare reform act likely will have a lasting impact. "The transformative opportunity here is all about the reimbursement shift from fee-for-service to the accountable care model, with the Centers for Medicare and Medicaid Services the major driver behind it," said John Sculley, chairman of the board and chief marketing officer at RxAdvance who previously served as CEO of Apple and PepsiCo. "And as we know there are hundreds of accountable care organizations out there, some operating under one definition and some under others. This is an evolving thing." Sculley is looking at this transformative opportunity in healthcare through the lens of his company RxAdvance, a vendor of a cloud-based pharmacy benefit management platform.

A new algorithm is proposed which accelerates the mini-batch k-means algorithm of Sculley (2010) by using the distance bounding approach of Elkan (2003). We argue that, when incorporating distance bounds into a mini-batch algorithm, already used data should preferentially be reused. To this end we propose using nested mini-batches, whereby data in a mini-batch at iteration t is automatically reused at iteration t+1. Using nested mini-batches presents two difficulties. The first is that unbalanced use of data can bias estimates, which we resolve by ensuring that each data sample contributes exactly once to centroids. The second is in choosing mini-batch sizes, which we address by balancing premature fine-tuning of centroids with redundancy induced slow-down. Experiments show that the resulting nmbatch algorithm is very effective, often arriving within 1% of the empirical minimum 100 times earlier than the standard mini-batch algorithm.

Apple Watch has sold well, but the wearables category itself hasn't proven to be nearly as hot as Mp3, smartphone and tablets, all categories where Apple redefined the market. The 18-year-old south London student, who worked two jobs on the side to help afford the gadget, couldn't imagine being without the company's latest tech product. But before the year was out, he would sell his 38mm Apple Watch Sport. "The only advantage over my iPhone was the exercise tracking abilities, but they weren't worth the 350 I paid for it," he says. That sort of customer testimonial would keep most CEOs up at night. And it represents a rare miss for the world's most-valuable company, which over the past decade has risen to iconic status -- and a 575 billion market cap -- on the back of its once-groundbreaking, billion-selling iPhone.

In many real-world applications of machine learning classifiers, it is essential to predict the probability of an example belonging to a particular class. This paper proposes a simple technique for predicting probabilities based on optimizing a ranking loss, followed by isotonic regression. This semi-parametric technique offers both good ranking and regression performance, and models a richer set of probability distributions than statistical workhorses such as logistic regression. We provide experimental results that show the effectiveness of this technique on real-world applications of probability prediction.