Artificial intelligence is becoming a central part of people's lives, even if they don't realize it. So many everyday functions have an artificial intelligence component – from auto-correct on text messages to map routes, from home loan approvals to Netflix suggestions. But while that may sound innovative, questions of "fairness" have arisen. For example, when it comes to home mortgages, white males tend to be approved more often and get lower interest rates, compared with others. But is it fair for that group to get an advantage over others?

My newest Ask-A-Data-Scientist post was inspired by a computer science student who wrote in asking for advice on how to pursue a career in policy making related to the societal impacts of AI. I realized that there are many great resources out there, and I wanted to compile a list of links all in one place. You can find my previous Ask-A-Data-Scientist advice columns here. Everyone in tech should be concerned about the ethical implications of our work and actively engaging with such questions. The humanities and social sciences are incredibly relevant and important in addressing ethics questions.

My name is Kevin, and I have a phone problem. And if you're anything like me – and the statistics suggest you probably are, at least where smartphones are concerned – you have one, too. I don't love referring to what we have as an "addiction". That seems too sterile and clinical to describe what's happening to our brains in the smartphone era. We might someday evolve the correct biological hardware to live in harmony with portable supercomputers that satisfy our every need and connect us to infinite amounts of stimulation. But for most of us, it hasn't happened yet.

So many people were asking Alexa what song they just heard and who sang it that Amazon responded with a new Echo feature. Rolling out first to Amazon Music Unlimited subscribers and Prime Music listeners in the US, Song ID will tell you the title and artist of each song before it plays on your Echo device. That should cut down on the "hundreds of thousands" of music questions Alexa hears daily. Don't worry, Song ID is optional, and you'll be able to turn it on and off with a simple voice command. When you use the feature, Amazon hopes it will "aid in music discovery" and make the speaker more of a personal DJ.

By using feature inversion to visualize millions of activations from an image classification network, we create an explorable activation atlas of features the network has learned which can reveal how the network typically represents some concepts.

It has come to Netflix's attention (perhaps a long time back) that its subscribers have been sharing their Netflix credentials with their loved ones. As harmless as this may seem on paper, Netflix reports the magnitude of revenue deficits it is experiencing because of this common practice. Netflix is the United States' most popular paid video streaming channel. Globally, 37% of the population that uses the internet is connected to Netflix in one way or the other. Netflix is a behemoth and is known to heavily invest in the production of its original series.

When the movie Her was released back in 2013, the concept of love between a human and a virtual entity was one that many people, myself included, had never really given much thought to. But as we watched the relationship blossom between Theodore, a recently-divorced writer struggling with loneliness, and Samantha, his charismatic and surprisingly empathetic operating system assistant, many of us began to wonder: Could this ever be possible? Fast-forward to today and this question isn't as far-fetched as it once may have seemed, thanks largely to the efforts of London-based software firm Spirit AI. While they have made a name for themselves using artificial intelligence to keep cyberbullying in check, they have also used this same technology to bring us 50 steps closer to the world of Her with Character Engine. Designed for videogame developers who understand the power of character to drive a story, Character Engine is an authoring tool and SDK that uses natural language processing and machine learning to create virtual personas who not only seem humanlike, but also process their environments in remarkably human ways.

State-of-the-art meta reinforcement learning algorithms typically assume the setting of a single agent interacting with its environment in a sequential manner. A negative side-effect of this sequential execution paradigm is that, as the environment becomes more and more challenging, and thus requiring more interaction episodes for the meta-learner, it needs the agent to reason over longer and longer time-scales. To combat the difficulty of long time-scale credit assignment, we propose an alternative parallel framework, which we name "Concurrent Meta-Reinforcement Learning" (CMRL), that transforms the temporal credit assignment problem into a multi-agent reinforcement learning one. In this multi-agent setting, a set of parallel agents are executed in the same environment and each of these "rollout" agents are given the means to communicate with each other. The goal of the communication is to coordinate, in a collaborative manner, the most efficient exploration of the shared task the agents are currently assigned. This coordination therefore represents the meta-learning aspect of the framework, as each agent can be assigned or assign itself a particular section of the current task's state space. This framework is in contrast to standard RL methods that assume that each parallel rollout occurs independently, which can potentially waste computation if many of the rollouts end up sampling the same part of the state space. Furthermore, the parallel setting enables us to define several reward sharing functions and auxiliary losses that are non-trivial to apply in the sequential setting. We demonstrate the effectiveness of our proposed CMRL at improving over sequential methods in a variety of challenging tasks.

Human-robot interaction is easy to do badly, and very difficult to do well. One approach that has worked well for robots from R2-D2 to Kuri is to avoid the problem of language--rather than use real words to communicate with humans, you can do pretty well (on an emotional level, at least) with a variety of bleeps and bloops. But as anyone who's watched Star Wars knows, R2-D2 really has a lot going on with the noises that it makes, and those noises were carefully designed to be both expressive and responsive. Most actual robots don't have the luxury of a professional sound team (and as much post-production editing as you need), so the question becomes how to teach a robot to make the right noises at the right times. At Georgia Tech's Center for Music Technology (GTCMT), Gil Weinberg and his students have a lot of experience with robots that make noise of various sorts, and they've used a new deep learning-based technique to teach their musical robot Shimi a basic understanding of human emotions, and how to communicate back to those humans in just the right way, using music.

A new camera app for iPhone lets you erase annoying tourists, cars or other objects that block your perfect shot. Spectre uses AI technology to stabilise long-exposure photos, focusing on the stationary part of a frame erasing moving parts, including people and objects. Near-perfect holiday snaps of beautiful views cluttered with people appear as if you have beaten the crowds and arrived to enjoy scenic vistas all to yourself. A new camera app for iPhone lets you erase annoying tourists, cars or other objects that block your perfect shot. Halide, the company behind the software, says its new app can generate such'magic' photos by using AI technology in several ways.