I can't remember whose idea it was to go out for lunch. It might have been my idea. Of course it was my idea. I start thinking about lunch the second that last morsel of granola hits my taste buds and disappears down that endless cavern known as my belly. Should we order a salad?

Protests in the ethnic Russian enclave in Riga, Latvia have NATO on edge. Russian units in the Western Military District are on alert conducting snap exercises involving autonomous ground and air attack systems. The Russian president makes a speech promising to protect ethnic Russians wherever they are with military forces if necessary. In response, a U.S. Army brigade combat team bolstered by intelligence, air defense, and aviation support elements from U.S. Army Europe deploys. Their mission is to reassure Latvian forces, deter Russian aggression, and if necessary conduct a mobile defense.

In a world where personal information is ubiquitous and accessible, shouldn't you have the right to be forgotten? How should we deal with traces of our online selves? These are just two of many questions and issues explored in Sheila Jasanoff's new book, The Ethics of Invention, which published this week. Jasanoff, a professor of science and technology studies at the Harvard Kennedy School of Government, explores ethical issues that have been created by technological advances--from how we should deal with large-scale disasters such as Bhopal or Chernobyl to the more hidden conundrums of data collection, privacy, and our relationship with tech giants like Facebook and Google. Jasanoff believes we don't sufficiently acknowledge how much power we've handed over to technology, which, she writes, "rules us as much as laws do."

August saw the Defense Advanced Research Projects Agency (DARPA) host its first Cyber Grand Challenge – the first hacking competition not involving people. During this event, teams left their systems alone to single-handedly find, diagnose and fix software flaws in real time. Elsewhere, researchers at MIT are not only developing machine learning systems that automatically mine dark web marketplaces for vulnerabilities and zero-day attacks and reports them back as well as software that automatically fixes buggy code, but also a platform that can predict 85% of cyber-attacks. Machine learning, deep learning, and Artificial Intelligence (AI) are hot topics at the moment, and while there's plenty of research going on, there's also some practical applications that can be deployed right now to make life easier for cybersecurity professionals. A glut of new start-ups, from the likes of Darktrace, Cylance, Deep Instinct, and HackerONE, plus established player such as FireEye, IBM, and Forcepoint, are all working on bringing self-learning systems into the world of security.

On August 19, 2014 something exceptional happened. Queen Elizabeth II of England finally granted a posthumous pardon to Alan Turing (1912-1954), convicted in 1952 for homosexual acts. Thus ended a long process of the British state to apologize to one of its most outstanding scientific figures of the twentieth century, whose contributions had a historical impact. During World War II he played a key role in helping the Allies to decipher the secret communications of the Nazis. And before that he had launched an idea that transformed computers into the powerful and versatile machines they are today. Today computers do much more than computing.

Recently we have seen a slew of popular films that deal with artificial intelligence – most notably The Imitation Game, Chappie, Ex Machina, and Her. However, despite over five decades of research into artificial intelligence, there remain many tasks which are simple for humans that computers cannot do. Given the slow progress of AI, for many the prospect of computers with human-level intelligence seems further away today than it did when Isaac Asimov's classic I, Robot was published in 1950. The fact is, however, that today the development of neuromorphic chips offers a plausible path to realizing human-level artificial intelligence within the next few decades. Starting in the early 2000's there was a realization that neural network models – based on how the human brain works – could solve many tasks that could not be solved by other methods. The buzzphrase'deep learning' has become a catch-all term for neural network models and related techniques, as is shown by a plotting of the frequency of the phrase using Google Trends: Most deep learning practitioners acknowledge that the recent popularity of'deep learning' is driven by hardware, in particular GPUs .

Automation has disrupted work for centuries. Two hundred years ago in Britain, the Luddites rose in rebellion, smashing the machines that made their weaving skills obsolete. Today it's high status cognitive jobs that are under threat. Earlier this year ROSS, a legal version of IBM's Watson, was launched and hailed as the first artificially intelligent lawyer. Future iterations may put lawyers out of work.

We provide two fundamental results on the population (infinite-sample) likelihood function of Gaussian mixture models with $M \geq 3$ components. Our first main result shows that the population likelihood function has bad local maxima even in the special case of equally-weighted mixtures of well-separated and spherical Gaussians. We prove that the log-likelihood value of these bad local maxima can be arbitrarily worse than that of any global optimum, thereby resolving an open question of Srebro (2007). Our second main result shows that the EM algorithm (or a first-order variant of it) with random initialization will converge to bad critical points with probability at least $1-e^{-\Omega(M)}$. We further establish that a first-order variant of EM will not converge to strict saddle points almost surely, indicating that the poor performance of the first-order method can be attributed to the existence of bad local maxima rather than bad saddle points. Overall, our results highlight the necessity of careful initialization when using the EM algorithm in practice, even when applied in highly favorable settings.

A computer's victory over a human go master this past March reminds us of the pending "singularity" -- the rapidly approaching moment in time when artificial intelligence overtakes human intelligence. Machines will learn, and we won't be their teachers. Are we prepared for it? Can we prepare for it? Many futurists declare it inevitable, probably within a generation, maybe less.

US Transportation Secretary Anthony Foxx delivers an announcement in Washington, DC, in 2014. Automakers and ride-hail companies are racing to put self-driving cars on the road. In a few weeks, Uber passengers in Pittsburgh will be able to hail self-driving Volvos. Last month, Tesla announced its hopes to build an autonomous ride-hailing fleet. And this month, Ford said it plans to mass-produce autonomous vehicles by 2021.