Here's a great introduction to neural networks that explains the core concepts without getting too deep into the mathematical weeds.

Kolibree, a smart oral care company, has unveiled what it claims is the first toothbrush embedded with AI at CES 2017. The Ara is clearly no ordinary mouth scrubber. Just read how Kolibree founder and CEO Thomas Serval describes Ara's tech: "Patented deep learning algorithms are embedded directly inside the toothbrush on a low-power processor," he said in a release. "Raw data from the sensors runs through the processor, enabling the system to learn your habits and refine accuracy the more it's used." Ara picks up precisely where you're brushing, and will log information, such as frequency, duration and location, in either on or offline mode.

WHAT: TechCode, through a strategic partnership with the new AI Frontiers Conference, will host a demo day on Thursday, January 12, featuring startups with a niche focus on artificial intelligence. The TechCode AI Accelerator Demo Day will include demos from fifteen startups and a Q&A session from a panel of Venture Capitalists. In its first year, the AI Frontiers Conference brings together leaders in applied artificial intelligence from companies such as Google, Facebook, Microsoft and Amazon to share insights into the latest deep learning advancements and products. The conference will cover six major areas including speech-enabled assistants, Internet of Things, natural language processing, computer vision, deep-learning frameworks and autonomous driving. All the startups featured in the demo day will be available for in-person interviews at the event.

Predicting the future is more art than science, yet it's always an interesting exercise to engage in as a new year comes upon us. So with the close of what was a difficult, though interesting year in the technology business, here's a look at my predictions for the top 10 tech developments of 2017. One of the key metrics for the relative health of the tech industry has always been the measurement of unit shipments and/or revenues for various categories of hardware-based tech devices. From PCs, tablets and smartphones, through smartwatches, smart TVs and head-mounted displays, there has been a decades-long obsession with counting the numbers and drawing conclusions from how the results end up. The problem is, the lines between these categories have been getting murkier and more difficult to distinguish for years, making what once seemed like well-defined groupings become increasingly arbitrary.

Amazon is set to take on eBay by allowing people to sell their secondhand goods. The program, which is still in its pilot phase, has been launched as'Sell as Individual' and is only available in Bangalore so far. The service includes doorstep pick-up, packing and delivery and will compete with other companies that allow people to sell their used goods such as eBay, Quikr and Olx. The secondhand products are listed on the website and when someone orders it, an Amazon staff visits the seller and collects their product from their home at a time the seller chooses. Sellers are charged 15 cents (10 rupees) for items below $15 (1,000 rupees), 75 cents (50 rupees) for items between $15 and $75 (5,000 rupees), and $1.50 (100 rupees) for items sold at more than $75. Amazon has unveiled a convenience store in Seattle that uses technologies found in self-driving cars to replace cashiers.

In the 1.4 release of Seldon we have added an alpha release of gRPC endpoints to complement our REST and Javascript endpoints. Remote Procedure Calls (RPC) and Google's version of this (gRPC) provides several advantages over REST. However, gRPC may be unfamiliar to many and requires a certain expertise for the developer building a gRPC client or server. For the 1.4 release of Seldon we have added gRPC as an external prediction endpoint as well as allowing prediction microservices to be deployed as gRPC servers internal to Seldon. The stages to deploy a gRPC service are discussed in detail in our docs and our summarized below.

As we kick off what will surely be another very exciting year of progress in artificial intelligence, machine learning and data science, we start with a quick recap of our "Top 10" most popular posts (based on aggregate readership) from the year just concluded. We announced a major set of updates to the Data Science VM (DSVM) in September last year. DSVM gives you a comprehensive set of tools for data movement, storage, exploration/visualization, modeling with ML/AI algorithms, and operationalization – and using multiple languages in either Linux or Windows environments. Just last month, we announced our latest and most powerful version of Microsoft R Server. Supporting popular operating systems and a variety of data sources, MRS 9.0 helps you create and deploy sophisticated analytics models for real world problems, efficiently and at scale.

Snapchat debuted its camera-equipped Spectacles this fall. Snapchat is adding ways to optimize campaign performance with the help of machine learning. Earlier this month, Snap Inc. began rolling out what it calls Goal-Based Bidding (GBB). The option, available to marketers buying ads through Snapchat's API, uses machine learning to know which users are most likely to swipe a certain type of ad. Here's how it works: With goal-based bidding, advertisers can inform Snapchat of when their main goal is increasing swipe-ups--perhaps for app installs, web views or movie trailers--instead of focusing solely on impressions.

I wanted to try and detect glioblastoma in diseased brain MRI from healthy ones, but I changed it to tiles because I guess for medical imaging it was a nightmare trying to scour the web and collect images to train on.

Next year, IoT-related conversations will everything from augmented reality to machine learning to voice recognition, according to three industry experts we spoke with. While the field of IoT companies ballooned in 2016, a handful of winners will likely become evident in 2017. Here's their advice on how to stay competitive in a maturing marketplace. THINKstrategies believes organizations can jumpstart their corporate IoT initiatives in 2017 by turning around the "long-tail" idea. "For many organizations, the IoT idea is too big to gain momentum until corporate decision-makers are more confident with their IoT readiness," says THINKstrategies' managing director, Jeff Kaplan.