Matthew Ramirez was teaching writing classes to students at the University of California at Berkeley when he started to get frustrated. Mixing his experience as a teacher with some advanced learning technology, he and his business partner started WriteLab – a Berkeley, California-based software company that helps students strengthen their writing skills by providing quick, customized feedback. WriteLab can even adapt its feedback over time to students' individual writing styles. "Focus on problems that eliminate waste – wasted time, wasted energy, or wasted space," Ramirez advises.

Matthew Ramirez was teaching writing classes to students at the University of California at Berkeley when he started to get frustrated. He was spending the majority of his time giving repetitive feedback to students, and there wasn't enough time to provide truly constructive, in-depth feedback to each individual before the next essay was due. "Given the time constraints of a semester and the number of students in a class, it wasn't humanly possible to respond to everything I wanted to," Ramirez said in an interview. "Nor was it possible to work through multiple drafts with individual students." Ramirez took it upon himself to build a solution.

With new neural network architectures popping up every now and then, it's hard to keep track of them all. Knowing all the abbreviations being thrown around (DCIGN, BiLSTM, DCGAN, anyone?) can be a bit overwhelming at first. So I decided to compose a cheat sheet containing many of those architectures. Most of these are neural networks, some are completely different beasts. Though all of these architectures are presented as novel and unique, when I drew the node structures… their underlying relations started to make more sense. One problem with drawing them as node maps: it doesn't really show how they're used. For example, variational autoencoders (VAE) may look just like autoencoders (AE), but the training process is actually quite different. The use-cases for trained networks differ even more, because VAEs are generators, where you insert noise to get a new sample. AEs, simply map whatever they get as input to the closest training sample they "remember". I should add that this overview is in no way clarifying how each of the different node types work internally (but that's a topic for another day).

One of the biggest concerns surrounding #artificialintelligence is that the technology will eliminate #jobs as it becomes increasingly more sophisticated. The first field to be impacted by artificial intelligence - chatbots, digital assistants, robotic process automation, and cognitive computing - will be customer service, followed by truck and taxi drivers and then the automotive industry. The general theme is that jobs that can be easily automated will be threatened by #robots, as companies will likely prefer cost-efficient robots over humans.

IBM has launched three Power8 Linux servers designed to accelerate artificial intelligence, deep learning, and advanced analytics applications. The new systems tap the Nvidia NVLink technology to move data five times faster than any competing platform, said Stefanie Chiras, an IBM vice president, in an interview with VentureBeat. These systems and their operating systems are part of a larger business group that generates about 2 billion a quarter for IBM. And the A.I. markets they're going after have exploded in the past couple of years. IBM claims that the combination of Power8 processors and Linux software results in systems that deliver 80 percent more performance per dollar than the latest x86-based (Intel or AMD) servers.

Like what you just read? To receive business news directly in your inbox each Tuesday and Thursday, sign up for our free e-newsletter, Briefcase, and be sure to join the conversation by commenting below.

In her opening keynote at the Intel Analytics Summit, which was mostly about machine learning, Intel's executive vice president Diane Bryant said that we are now "reaching a tipping point where data is the game changer." With the rapid growth of machine-to-machine data exchange we should expect more, according to Bryant: Autonomous vehicles will produce 4 terabytes of data each day, a connected plane will transmit 40 terabytes of data, and the automated, connected factory will generate one petabyte (one million gigabytes) daily. Another presenter, CB Bohn, Senior Database Engineer at Etsy, the online marketplace, speculated that the tipping point has already happened--when the value of the data exceeded the cost of its storage. Historical data has lots of value left in it, so "why throw it away?" Cheap storage, added Debora Donato, Director of R&D at Mix Tech, a content discovery platform, has changed the attitudes of businesses towards data and what they can do with it.

This post is authored by Jaya Mathew & Hong Lu, Data Scientists at Microsoft, in collaboration with the team at ServusNet. Governments worldwide are actively promoting the use of renewable energy because of the growing concern over climate change and carbon emissions. Among the available renewable energy sources, harnessing energy through wind turbines offers a compelling alternative. There are a growing number of wind farms throughout the world now, and, as the industry matures at a rapid pace, it is increasing its focus on operational efficiency and cost-competitiveness. Given the inherent variability in wind energy production, the ability to analyze the data generated by wind turbines to characterize historical behavior and create a framework to predict future performance can be a critical asset and business differentiator.

Elos Medtech Pinol in Gørløse, Denmark, is one of Europe's leading development and production partners for medical technology products and components. In recent years, increasing demands for quality, process documentation, and product consistency have forced customers to use fewer suppliers. Customers have selected Elos Medtech Pinol largely due to targeted optimization, standardization, and automation in the last decade. The automated factory – a mantra used within the company that impacts all employees – includes ambitious investments in user-friendly and flexible production equipment and the latest automation technology, including an MiR100 robot from Mobile Industrial Robots (MiR). In 2011, Danish robot scientist Niels Jul Jacobsen came up with a new product idea that a few years later was transformed into a product that handles internal transport logistics.

Training neural networks involves quite a few tricky bits. We try to make everything clear and easy to understand, to get you training your neural networks as quickly as possible. Theano allows us to write relatively concise code that follows the structure of the underlying maths. We will train a network to classify digits. More precisely, we want a network that when presented with an image of a hand-written digit will tell us what digit it is ('0', '1', ..., '9').