By now the world has heard about the rise and fall of Microsoft's Tay, an artificially intelligent bot that lived on Twitter, Kik, and GroupMe. To better understand where exactly Microsoft went wrong with Tay, I spoke with Brandon Wirtz, the creator of Recognant, a cognitive computing and artificial intelligence (AI) platform designed to aid in understanding big data from unstructured sources. Tay's Twitter conversations started out innocently enough, proclaiming her love for humans and wishing that National Puppy Day was every day. Upon analyzing Tay's tweets, Broad Listening found that Tay made four times more negative tweets than that of popular teen celebrities from Disney such as Peyton List, Laura Marano, China McClain, and Kelli Berglund.

For years, automakers have rhapsodized about how our cars would become mobile offices and living spaces. And then they botched even the simple stuff, like letting you dial up the Backstreet Boys from your iPod on the car's sound system. The Mercedes-Benz F 015 will be the rolling roost of your dreams. You'll just have to wait at least until 2030, when Mercedes thinks this kind of hydrogen-powered, fully autonomous vehicle will become viable. Bigger than an S-Class, the Benz concept looks like a Clockwork Orange hipster lounge, with its walnut-veneered floor and wall-wrapping touch and gesture displays.

Atlas, the humanoid robot created by Alphabet (GOOGL, Tech30) company Boston Dynamics, can open doors, balance while walking through the snow, place objects on a shelf and pick itself up after being knocked down. The new version of Atlas is smaller and more nimble than its predecessor. It's fully mobile too -- the previous version had to be tethered to a computer. Atlas was created to perform disaster recovery in places unsafe for humans, such as damaged nuclear power plants. The robot made its debut in 2013 during a competition held by the Defense Advanced Research Projects Agency. The new version of Atlas is a result of seven computer research teams from around the world who were contracted to develop software to give Atlas a better brain.

From the street, through Beijing's heavy smog, it can sometimes be hard to make out IBM's Chinese headquarters: a towering office building with a distinctive undulating architectural flourish and a large company logo at the top. But just a short distance away, on the northeast outskirts of the capital, IBM computer scientists are using artificial intelligence to develop what they think will be a way to manage China's notorious and chronic pollution problem more successfully. The team is using complex computer models and machine learning to calculate how pollution will spread across the city. The researchers can now produce pollution forecasts, with a resolution of a kilometer square, up to 10 days in advance. These predictions can also tell the government how it might act to avoid the worst scenarios--for instance, by shutting certain factories, or by reducing the number of cars on the road.

Rolls-Royce presents a vision of a future land-based control centre in which a small crew of 7 to 14 people monitor and control a fleet of remote controlled and autonomous vessels across the world. The crew uses interactive smart screens, voice recognition systems, holograms and surveillance drones to monitor what is happening both on board and around the ship. Remote and autonomous ships are one of three elements of the company's innovative Ship Intelligence strategy, which will enable customers to transform their marine businesses by harnessing the power of big data. The film marks the final stage of research that will inform the design and construction of an effective remote operations centre which is essential to the company's plans to develop autonomous and remote controlled vessels. The film is the latest in a series to present Rolls-Royce's vision of future shipping known as the'oX' operator experience concept and introduced in 2014.

Many experts say combining Big Data and artificial intelligence will bring about the 4th Industrial Revolution. Park Se-young has the details. The Industrial Revolution brought huge changes to the economy through the mass production of manufactured goods. The second industrial revolution came about on the back of new innovations in steel production, petroleum and electricity …which led to the introduction of public transportation and airplanes. A third round of changes was driven by internet technology and renewable energy, creating a new industry for ICT and the energy sector.

The problem of recovering a structured signal $\mathbf{x} \in \mathbb{C}^p$ from a set of dimensionality-reduced linear measurements $\mathbf{b} = \mathbf {A}\mathbf {x}$ arises in a variety of applications, such as medical imaging, spectroscopy, Fourier optics, and computerized tomography. Due to computational and storage complexity or physical constraints imposed by the problem, the measurement matrix $\mathbf{A} \in \mathbb{C}^{n \times p}$ is often of the form $\mathbf{A} = \mathbf{P}_{\Omega}\boldsymbol{\Psi}$ for some orthonormal basis matrix $\boldsymbol{\Psi}\in \mathbb{C}^{p \times p}$ and subsampling operator $\mathbf{P}_{\Omega}: \mathbb{C}^{p} \rightarrow \mathbb{C}^{n}$ that selects the rows indexed by $\Omega$. This raises the fundamental question of how best to choose the index set $\Omega$ in order to optimize the recovery performance. Previous approaches to addressing this question rely on non-uniform \emph{random} subsampling using application-specific knowledge of the structure of $\mathbf{x}$. In this paper, we instead take a principled learning-based approach in which a \emph{fixed} index set is chosen based on a set of training signals $\mathbf{x}_1,\dotsc,\mathbf{x}_m$. We formulate combinatorial optimization problems seeking to maximize the energy captured in these signals in an average-case or worst-case sense, and we show that these can be efficiently solved either exactly or approximately via the identification of modularity and submodularity structures. We provide both deterministic and statistical theoretical guarantees showing how the resulting measurement matrices perform on signals differing from the training signals, and we provide numerical examples showing our approach to be effective on a variety of data sets.

In what could be the last time that a human taunts a robot with a hockey stick and lives to brag about it, the latest demonstration of the Atlas Robot has prompted renewed fears about the future of intelligent machines. Born in 2013, Atlas is a DARPA-funded robot developed by Boston Dynamics. Its latest iteration stands at a very human-proportioned 5'9, weighing 180lbs. Like Lee Majors circa 1974, each successive version of Atlas has gotten better, stronger, faster than it was before. Aside from scaring the bejesus out of genius technophobic Oxbridge physicists, it's intended to perform tasks in emergency situations too dangerous for humans.

The nonnegative matrix factorization (NMF) is widely used in signal and image processing, including bio-informatics, blind source separation and hyperspectral image analysis in remote sensing. A great challenge arises when dealing with a nonlinear formulation of the NMF. Within the framework of kernel machines, the models suggested in the literature do not allow the representation of the factorization matrices, which is a fallout of the curse of the pre-image. In this paper, we propose a novel kernel-based model for the NMF that does not suffer from the pre-image problem, by investigating the estimation of the factorization matrices directly in the input space. For different kernel functions, we describe two schemes for iterative algorithms: an additive update rule based on a gradient descent scheme and a multiplicative update rule in the same spirit as in the Lee and Seung algorithm. Within the proposed framework, we develop several extensions to incorporate constraints, including sparseness, smoothness, and spatial regularization with a total-variation-like penalty. The effectiveness of the proposed method is demonstrated with the problem of unmixing hyperspectral images, using well-known real images and results with state-of-the-art techniques.

Guest blog post, authored by Brandon Rohrer, Senior Data Scientist at Microsoft. Machine learning (ML) is the motor that drives data science. Each ML method (also called an algorithm) takes in data, turns it over, and spits out an answer. ML algorithms do the part of data science that is the trickiest to explain and the most fun to work with. That's where the mathematical magic happens.