UPDATE: Mar 20, 2016 - Added my new follow-up course on Deep Learning, which covers ways to speed up and improve vanilla backpropagation: momentum and Nesterov momentum, adaptive learning rate algorithms like AdaGrad and RMSProp, utilizing the GPU on AWS EC2, and stochastic batch gradient descent. We look at TensorFlow and Theano starting from the basics - variables, functions, expressions, and simple optimizations - from there, building a neural network seems simple! Deep learning is all the rage these days. What exactly is deep learning? Well, it all boils down to neural networks.

If you think deep, you'd realize the whole process of predictive modeling is a war. Consider the data set as your opponent. Your win will depend on the intuitive usage of your knowledge & strategy to get highest accuracy. So, how many times have you won this battle? I realized it last week.

Twitter started the trend of'People to Follow'. This later got replicated by other platforms such as Facebook, Linkedin, Quora and GitHub. This cool feature lets you connect with the rockstars of various domains and get an access to what is going on their end without bothering them much. For the influencers, this has become an effective way to communicate with their followers. The lives of people on GitHub doesn't appear to as tempting as you would observe on other platforms, but if you love coding, programming and data science, you'll surely enjoy the company of 9 million users on this platform!

Over a year ago, following an original presentation at MLConf, I wrote a blog post entitled "10 Lessons Learned from building ML systems". At that point, I was leading the Algorithms Engineering team at Netflix and those lessons reflected lessons we had learned there over the last few years. When you do a post/presentation like that, you don't really know how it is going to be received. Some things might be obvious to many while others might be controversial and some will not agree. It turns out though that it was very well received and referenced elsewhere (e.g.

The veteran provider of document capture and OCR software has a new product suite aimed at enabling business applications to read and understand natural language, and thus help their users make better-informed decisions. Is the time right for a resurgence of interest in text analytics?

H2O.ai is nurturing a grassroots movement of developers and data scientists to herald a new wave of discovery powered by machine learning. Come join us to change the world in a meaningful way! If you have a startup mentality and a desire to help some of the world's largest organizations leverage the world's most powerful algorithms we want to hear from you. A competitive salary, terrific health benefits, a great location, complimentary meals and snacks and unlimited PTO are only a few of the perks we offer.

DARPA has announced plans to hold a competition pitting electromagnetic spectrum receivers (think: TVs, wifi-enabled computers, radios). As the amount of mobile data traffic is growing at an exponentially rapid rate, DARPA believes we need to start refining the way we handle the crowded spectrum. The Spectrum Collaboration Challenge (SC2) will see research teams collaborating to create smart systems to share wavelengths using algorithms and artificial intelligence. Global mobile data traffic grew by 74 percent in 2015, and more than half a billion devices and connections were added to the overwhelmingly clogged spectrum, according to recent Cisco reports. Analysts expect that the monthly global mobile traffic rates will reach 30.6 exabytes by 2020 (rates are at 3.7 exabytes per month as of 2015).

IBM's Watson AI product is mostly rolled out live with machine learning halted to avoid "losing control" of its behaviour, Europe CTO Duncan Anderson has confirmed. Anderson said the idea of AI always learning and adjusting its behaviour is still something people are "a bit nervous about". "At the moment, you stop the learning before it goes live, so you don't get any surprises," Anderson told Computing at our Big Data & Analytics Summit 2016 in London. The aim, explained Anderson, is to "get a sensible kind of answer" from an AI in line with the business's expectations. He added that Watson's modular-based learning updates are now so advanced in specific areas of industry that it's now possible to sell "off the shelf" versions of the AI that can immediately get to grips with traditional tasks in a given area.

Artificial intelligence (AI) is here to stay. The AI revolution is changing lives but scaring millions. AIs are already everywhere, supporting the human mind in the pursuit to make the world better and our way of life even easier. They're in your phones and favorite apps, helping you stay connected and do what you do best in the shortest possible time. The opportunities are endless and uncertain.

Microsoft's attempt at engaging millennials with artificial intelligence has backfired hours into its launch, with waggish Twitter users teaching its chatbot how to be racist. The company launched a verified Twitter account for "Tay" – billed as its "AI fam from the internet that's got zero chill" – early on Wednesday. The chatbot, targeted at 18- to 24-year-olds in the US, was developed by Microsoft's technology and research and Bing teams to "experiment with and conduct research on conversational understanding". Related: How much should we fear the rise of artificial intelligence? "Tay is designed to engage and entertain people where they connect with each other online through casual and playful conversation," Microsoft said.