Variational auto-encoders (VAE) are scalable and powerful generative models. However, the choice of the variational posterior determines tractability and flexibility of the VAE. Commonly, latent variables are modeled using the normal distribution with a diagonal covariance matrix. This results in computational efficiency but typically it is not flexible enough to match the true posterior distribution. One fashion of enriching the variational posterior distribution is application of normalizing flows, i.e., a series of invertible transformations to latent variables with a simple posterior. In this paper, we follow this line of thinking and propose a volume-preserving flow that uses a series of Householder transformations. We show empirically on MNIST dataset and histopathology data that the proposed flow allows to obtain more flexible variational posterior and competitive results comparing to other normalizing flows.

We study the effectiveness of neural sequence models for premise selection in automated theorem proving, one of the main bottlenecks in the formalization of mathematics. We propose a two stage approach for this task that yields good results for the premise selection task on the Mizar corpus while avoiding the hand-engineered features of existing state-of-the-art models. To our knowledge, this is the first time deep learning has been applied to theorem proving on a large scale.

Ever since algorithms began recognizing patterns faster and better than humans, computers have been making doctors' lives easier and diagnoses more accurate. But widely used tools like automated cell counters, which can quickly point to diseases like malaria and leukemia by getting a head count on different kind of blood cells, are beginning to look quaint next to the deep learning and neural networks coming online. Today, hospitals can outfit their existing computer systems with a $1,000 graphics processor and speed-boost their capacity up to 260 million images per day. That's basically equivalent to all the MRIs, CT scans, and other images that all the radiologists in America look at each day. Unleashing that kind of AI on the medical world's mountains of patient data could speed up diagnoses and get patients on the path to recovery much sooner.

It's scary enough making a doctor's appointment to see if a strange mole could be cancerous. Imagine, then, that you were in that situation while also living far away from the nearest doctor, unable to take time off work and unsure you had the money to cover the cost of the visit. In a scenario like this, an option to receive a diagnosis through your smartphone could be lifesaving. Universal access to health care was on the minds of computer scientists at Stanford when they set out to create an artificially intelligent diagnosis algorithm for skin cancer. They made a database of nearly 130,000 skin disease images and trained their algorithm to visually diagnose potential cancer.

A number of research organizations are working to create artificial intelligence systems capable of developing machine-learning software. Several research organizations, including Google Brain and DeepMind, are working to create artificial intelligences (AI) that can in turn develop machine-learning software. In many cases, the results coming from machines programming other machines match or exceed work done by humans. If self-programming AI techniques become practical, they could increase the pace at which machine learning is adopted throughout the economy without requiring more machine-learning experts, who already are in short supply. One set of experiments from DeepMind suggests self-teaching methods could alleviate the problem of AI software needing to consume massive amounts of data on a specific task.

The Friday afternoon news dump, a grand tradition observed by politicians and capitalists alike, is usually supposed to hide bad news. So it was a little weird that Elon Musk, founder of electric car maker Tesla, and Sam Altman, president of famed tech incubator Y Combinator, unveiled their new artificial intelligence company at the tail end of a weeklong AI conference in Montreal this past December. But there was a reason they revealed OpenAI at that late hour. It wasn't that no one was looking. It was that everyone was looking. When some of Silicon Valley's most powerful companies caught wind of the project, they began offering tremendous amounts of money to OpenAI's freshly assembled cadre of artificial intelligence researchers, intent on keeping these big thinkers for themselves. The last-minute offers--some made at the conference itself--were large enough to force Musk and Altman to delay the announcement of the new startup.

Every day we create billions of bits of data. Ever faster and more powerful computers can use that big data to learn, predict events and carry out key tasks. Surveillance, voice recognition and driving vehicles are all areas where people are becoming superfluous. The BBC's Colm O'Regan explores the process known as "deep learning" - a branch of machine learning used to develop artificial intelligence.

Artificial intelligence is once again in the media spotlight. But what is it exactly? And how does it relate to developments in machine learning and deep learning? Below we spell out the various interpretations of AI and look back on how the technology has developed over the years. "The fundamental challenge is that, alongside its great benefits, every technological revolution mercilessly destroys jobs and livelihoods – and therefore identities – well before the new ones emerge."

Deep learning is taking on dermatology. An algorithm trained in image recognition has matched dermatologists in its ability to identify certain types of skin cancer based on photographs of skin lesions. "I'm certain this is how melanomas are going to be identified in the future," says Richard Weller, a consultant dermatologist at the Royal Infirmary of Edinburgh in the UK, who was not involved in the work. Researchers led by Andre Esteva and Brett Kuprel at Stanford University trained a neural network on more than 129,000 images of skin lesions associated with 2000 different diseases. They then pitted it against 21 certified dermatologists on new sets of images to find out whether deep-learning algorithms could reliably pick out cancerous moles and lesions.

This 3-hour course (video slides) offers developers a quick introduction to deep-learning fundamentals, with some TensorFlow thrown into the bargain. Deep learning (aka neural networks) is a popular approach to building machine-learning models that is capturing developer imagination. If you want to acquire deep-learning skills but lack the time, I feel your pain. In university, I had a math teacher who would yell at me, "Mr. Görner, integrals are taught in kindergarten!"