With the deluge of deep learning libraries and software innovation in the field over the last few years, it is an exciting time to be working on machine learning problems. Most of the libraries available evolved from fairly specialized computational code for large dense problems such as image classification into general frameworks for neural-network-based models offering marginal support for sparse models. At Netflix, our machine learning scientists deal with a wide variety of problems across a broad spectrum of areas: from tailoring TV and movie recommendations to your taste to optimizing encoding algorithms. A subset of our problems involve dealing with extremely sparse data; the total dimensionality of the problem at hand can easily reach tens of millions of features, even though every observation may only have a handful of non-zero entries. For these cases, we felt the need for a minimalist library that is specifically optimized for training shallow feedforward neural nets on sparse data in a single-machine, multi-core environment.

Movie studios have a long tradition of testing out new films to see how audiences react before launching them in wide release. But with their latest research innovation, Disney is taking it to a whole new level. Now as you're settling in to watch the latest Disney blockbuster, the movie could also be watching you. And while this could signal an exciting new era of responsive storytelling in which movies are shaped around our likes and dislikes in real time, it also raises some red flags about yet another frontier in personal data collection. At a conference in July, Disney Research presented a new process called factorized variational autoencoders (FVAEs).

Both industry and community must develop effective mechanisms to filter bias as well as negative sentiment in the data that AI learns from - ensuring AI does not perpetuate stereotypes. Users build a relationship with AI and start to trust it after just a few meaningful interactions. With trust, comes responsibility and AI needs to be held accountable for its actions and decisions, just like humans. Technology should not be allowed to become too clever to be accountable. We don't accept this kind of behaviour from other'expert' professions, so why should technology be the exception.

Amazon announced today that users of its streaming service Prime Music, which is free with a Prime membership, and its subscription-based Amazon Music Unlimited can now ask Alexa to find tunes appropriate for various activities. As of now, over 500 different activity-based requests are supported including music for meditation, partying and even "getting pumped." The new feature is available immediately to users with Alexa-enabled devices. The new voice controls were geared towards activities that have been requested most often by Alexa users and listeners of Amazon's music streaming services. In the announcement, the company said that 27 percent of all activity requests come from users who want to relax.

I know this might sound like a "google this for me question" but bare with me (I googled it). I'm working with a highly imbalanced data set where the minority class accounts for 1.5% of the total set. This leads to poor predictive performance by most models when nothing is done to address the problem because most algorithms will minimize cost on the majority class, to the detriment of the minority class, when training so as to decrease overall cost. So far I've tried out ANNs,RFs,XGBs, and SVMs and have found that XGB and RF outperform the others in this particular problem, so the remaining post will be about RF and XGB. I've tried penalizing classification on the minority class much more than the majority class to try to fix the imbalance on an algorithmic level but I've found undersampling and then training on the resulting data set to be more effective in my case.

Summary: Just how accurate are algorithms at spotting fake news and are we ready to turn them loose to suppress material they don't find credible. Here are some considerations and stories about some of the companies trying to build these fact-checkers. Which of the following headlines is fake? The first two are in fact fake. You may actually have seen them during last fall's election coverage.

Dear songwriters, I am sorry to tell you, but you are done. Remember the times, when singers actually had to hire you to write songs for them? Well, that time will be over soon. Because music can be generated by computers now -- and you, my friends, will be replaced by machine learning algorithms within the next years. But imagine the power such a program has.

The following piece is a guest post from Jessica Rovello, the CEO and co-founder of Arkadium, which provides interactive content to brands and publishers. Guest posts do not necessarily reflect the opinions of this publication. Is the growing adoption of Artificial intelligence (AI) products by digital publishers a much-needed ...

There are many newspapers available in the New York City area that cater to different segments of the population. This project focuses on the Wall Street Journal (WSJ), an international newspaper with a high circulation in the New York area. For this project, I scraped details of articles over a three-week period in order to analyze some basic metrics of the newspaper as well as some of the topics that the WSJ focuses on. In total, the project encompassed a total of 4,126 articles from three consecutive weeks in August 2016. As a result of collecting these metrics, several interesting findings emerged.

Filmmakers have been trying to adapt Stephen King's The Dark Tower series for more than a decade. But with time-jumping metanarratives and compulsive genre-switching, the eight novels proved tough to wrangle into one film-able narrative. Director Nikolaj Arcel's version of King's events finally hits theaters today. Written by no fewer than four writers (not including King), the movie arrives with a lean 95-minute runtime and the kind of Rotten Tomatoes score (21% and barely climbing) that studios fear. But is it possible the critics aren't being fair?