PAPIs '15, the 2nd International Conference on Predictive APIs and Applications, took place in Sydney, Australia and featured 4 research presentations. The corresponding papers were compiled into proceedings that were published in the Journal of Machine Learning Research (Volume 50 of the Workshop & Conference Proceedings series; you can also download the whole proceedings in a single pdf here). The first paper of these proceedings gives us a behind-the-scenes look at Microsoft Azure ML, an MLaaS environment for authoring predictive models, experimenting with them, running them on a cloud infrastructure and publishing them as web APIs. The Azure ML team presents design principles, challenges encountered and lessons learnt while building the platform. While it is common for ML practitioners to measure models' performance via predictions' accuracy, the second paper of these proceedings by Brian Gawalt of Upwork focuses on concerns of software engineers who are in charge of deploying in production and scalability: models' throughput and response time.

A team of CMU researchers led by Associate Professor Anupam Datta have developed new measurement methods that provide important insight into how machine-learning algorithms make decisions about things like credit applications, job opportunities and medical diagnoses. Machine-learning algorithms increasingly make decisions about credit, medical diagnoses, personalized recommendations, advertising and job opportunities, among other things, but exactly how usually remains a mystery. Now, new measurement methods developed by Carnegie Mellon University researchers could provide important insights to this process. Was it a person's age, gender or education level that had the most influence on a decision? Was it a particular combination of factors?

If you're one of the 38 million American travelers this Memorial Day weekend, or enjoying one of the 800 hot dogs consumed per second during this holiday, then take a step back, relax and enjoy the week's top technology news! The question of human sustainability and sufficient resources on our planet has served as the basis for many science fiction movies. But unlike Hollywood's'Interstellar' film, our answer might be right here on Earth. AI is showing potential to completely revolutionize farming, and with it keep humanity's food supply coming. What if we could go past what we know as binary computing?

AI & deep learning are powering interactive messaging services known as chatbots & virtual assistants, which use conversational interfaces to create deeper, more personalised one-to-one customer experiences. The Chatbot Track will explore the technical advancements in deep learning, NLP & predictive intelligence to create conversational self-learning bots for messaging platforms, healthcare, personalised services & more.

On May 10, the world-class admin behind John Chambers's success was honored with one of the top awards in her field: Debbie Gross received the Colleen Barrett Award for Administrative Excellence. Clearly, Debbie is a force. This CNBC story gives us a peek into her life keeping John at the top of his game. People are rightly saying she's a role model for next-generation administrators. But what people aren't saying is that some next-gen admins are made not of flesh and blood like Debbie but of compute cycles.

May 27--As technologists race to invent the next big thing after the smartphone, many have overlooked what is starting to seem obvious: That the human voice is a powerful, perhaps the most powerful, mechanism for controlling the world around us. And, more importantly, speaking is a behavior that doesn't require a user manual. Now, Amazon, Google and Apple are speeding toward a not-too distant future when your voice will dictate commands to a personal assistant, powered by artificial intelligence, and ultimately supplant your PC and phone for most quotidian computing purposes. In other words, your voice -- not a screen -- will become the primary interface for accessing the Internet, and even controlling your home or car. "The entire Silicon Valley startup community was more enamored over virtual reality and augmented reality, and in the process completely overlooked the transformative nature of voice-first technology," said Brian Roemmele, a Temecula-based researcher and consultant in voice-based technologies and online commerce.

Magic Leap is apparently working on a virtual reality technology that's so revolutionary it justifies valuing the company at 4.5 billion, even though almost no one has seen this mysterious tech, or knows exactly what it will do. But according to a lawsuit filed by Magic Leap on May 26 and first uncovered by MIT Technology Review, the company isn't satisfied with working on one of the most talked-about technologies of the year; it also appears to be working on some sort of robot. Magic Leap filed the suit against two former employees, including its head of computer vision and AI work, Gary Bradski, for starting a competing company while still employed by Magic Leap. The suit alleges that the two employees used company time to work on a separate venture for at least a year. The lawsuit contains three references to robotics, and specifically Magic Leap's feelings that whatever Bradski and vice president of special projects Adrian Kaehler were up to competed with Magic Leap's own robotics work: Magic Leap's Proprietary Technologies are not limited to its head-mounted virtual retinal display and extend to many different applications and devices, including, but not limited to, robotics.

Would you share your home with a robot or work side by side with one? People are starting to do both, which has put the relationship we have with them under the spotlight and exposed both our love and fear of the machines that are increasingly becoming a crucial part of our lives. In Japan they grow so attached to their robot dogs that they hold funerals for them when they "die". Sony, the firm that began making the popular Aibo toys in 1999, decided to stop offering repairs in 2014, meaning once they broke down they were fit only for the scrapheap. But people weren't willing to throw them in the rubbish bin, wanting instead to say goodbye to them in the same way you would to a human or pet.

Five days from now Google will publish open-source tools that will focus its machine-learning engine on music and art. But one London startup, named Jukedeck, has been working on getting machines to automatically generate original music for years. You can even generate your own ditty, composed by artificial intelligence, right now on Jukedeck's website. Jukedeck lets anyone use its machine-learning engine to generate tunes hosted on its website. The engine then produces an original piece of music.

Though research is a slow moving and rigid process, one study shows that the rate of scientific study has exploded in the last 50 years. According to the paper, humanity's scientific output now doubles every nine years. In specific areas like healthcare, the doubling rate is even faster -- as much as every 3 years currently with an expected increase to every 73 days by the early 2020s. For overwhelmed researchers navigating the growing stack of science literature -- the value isn't in having so much new information, but finding relevant insights when they need them. According to Jacobo Elosua, a co-founder of Iris AI -- a Singularity University portfolio company -- the research process is very often tedious and unfruitful.