Using hand gestures when we talk is an ordinary part of communication between humans, adding emphasis and flavor to our speech. It's second nature for us, but can robots be taught to take on such a human-like habit? Now, researchers have found that when humanoid robots talk with their hands, we understand them just as well as we do our fellow human beings. Humanoid robots, or avatars, have been around for several decades and are becoming big business. From artificial intelligence to social media and psychotherapy to high-end video games, they are used to sell things, to solve problems, to teach us and to entertain us. As they become more sophisticated and more sought after, getting your message across with your avatar is becoming more important than ever.

Diplomats from around the world met in Geneva last week for the United Nations' third Informal Expert Meeting on lethal autonomous weapons systems (LAWS), commonly dubbed "killer robots". Their aim was to make progress on deciding how, or if, LAWS should be regulated under international humanitarian law. A range of views were expressed at the meeting, from Pakistan being in favour of a full ban, to the UK favouring no new regulation for LAWS, and several positions in between. Despite the range of views on offer, there was some common ground. It is generally agreed that LAWS are governed by international humanitarian law.

When Hollywood actress Angelina Jolie found out she carried the BRCA1 gene, her doctors told her she had an 87% chance of developing breast cancer. Armed with this knowledge, she chose to undergo a double mastectomy in 2013 to reduce the risk to around 5%. This kind of genetic testing can now be done much faster and at lower cost, giving clinicians the ability to target treatments more effectively. And combining this technological breakthrough with cloud computing and artificial intelligence is giving pharmaceutical companies the tools to develop drugs faster and with greater chance of success. One beneficiary of this new approach is Eric Dishman, founder of tech giant Intel's first health research and innovation laboratory in 1999 and a founding member of its digital health group in 2005.

HEFFNER: I'm Alexander Heffner, your host on The Open Mind. TED Talk curator Chris Anderson joined us recently to consider the danger of artificial intelligence, namely its potential to drive away or make obsolete the moral compass of human beings and civilization as we know it. Of course sometimes, we're our own worst enemy, and we would rather not embrace the present reality. So I've invited today the leading ethicist in the arena of innovation. He's going to help us understand the term techno sapiens as he calls it, with our drones, our supercomputers, our designer babies, and now our 3D printers too. Wendell Wallach is the author of A Dangerous Master: How to Keep Technology From Slipping Beyond Our Control.

Call me old and grumpy, but there are certain things in life that an artificial intelligence or a robot cannot do for me. If I buy gifts for my family and friends, I love to do the research. I like to read product descriptions and reviews, to compares stores to find the best prices and to weigh the different payment options. All this effort on my part, in the end, provides a sense of satisfaction that simply cannot be achieved through the blunt response of a bot. Also, I love investing my time this way because my family and friends are worth the effort to me.

You could be sent to prison and charged with a felony for shooting a drone from the sky. According to the federal law, 18 USC S 32, anyone who willfully'sets fire to, damages, destroys, or wrecks an aircraft' will be fined or imprisoned no more than 20 years or both. And the FAA says drones fall into the category of'aircraft' and threatening anyone operating a drone is also punishable with jail time. According to the federal law, 18 USC S 32, anyone who willfully'sets fire to, damages, destroys, or wrecks an aircraft' will be fined or imprisoned no more than 20 years or both. And experts say drones fall into the category of'aircraft' and threatening anyone operating a drone falls is also punishable with jail time The law says that if you attempt to shoot down a flying robot from the sky, you could face up to two decades behind bars, and/or be handed a fine up to a quarter of a million dollars.

Forecasting models play a key role in money-making ventures in many different markets. Such models are often trained on data from various sources, some of which may be untrustworthy.An actor in a given market may be incentivised to drive predictions in a certain direction to their own benefit.Prior analyses of intelligent adversaries in a machine-learning context have focused on regression and classification.In this paper we address the non-iid setting of time series forecasting.We consider a forecaster, Bob, using a fixed, known model and a recursive forecasting method.An adversary, Alice, aims to pull Bob's forecasts toward her desired target series, and may exercise limited influence on the initial values fed into Bob's model.We consider the class of linear autoregressive models, and a flexible framework of encoding Alice's desires and constraints.We describe a method of calculating Alice's optimal attack that is computationally tractable, and empirically demonstrate its effectiveness compared to random and greedy baselines on synthetic and real-world time series data.We conclude by discussing defensive strategies in the face of Alice-like adversaries.

Gradient backpropagation (BP) requires symmetric feedforward and feedback connections — the same weights must be used for forward and backward passes. This "weight transport problem'' (Grossberg 1987) is thought to be one of the main reasons to doubt BP's biologically plausibility. Using 15 different classification datasets, we systematically investigate to what extent BP really depends on weight symmetry. In a study that turned out to be surprisingly similar in spirit to Lillicrap et al.'s demonstration (Lillicrap et al. 2014) but orthogonal in its results, our experiments indicate that: (1) the magnitudes of feedback weights do not matter to performance (2) the signs of feedback weights do matter — the more concordant signs between feedforward and their corresponding feedback connections, the better (3) with feedback weights having random magnitudes and 100% concordant signs, we were able to achieve the same or even better performance than SGD. (4) some normalizations/stabilizations are indispensable for such asymmetric BP to work, namely Batch Normalization (BN) (Ioffe and Szegedy 2015) and/or a "Batch Manhattan'' (BM) update rule.

This paper presents a novel nonmyopic adaptive Gaussian process planning (GPP) framework endowed with a general class of Lipschitz continuous reward functions that can unify some active learning/sensing and Bayesian optimization criteria and offer practitioners some flexibility to specify their desired choices for defining new tasks/problems. In particular, it utilizes a principled Bayesian sequential decision problem framework for jointly and naturally optimizing the exploration-exploitation trade-off. In general, the resulting induced GPP policy cannot be derived exactly due to an uncountable set of candidate observations. A key contribution of our work here thus lies in exploiting the Lipschitz continuity of the reward functions to solve for a nonmyopic adaptive epsilon-optimal GPP (epsilon-GPP) policy. To plan in real time, we further propose an asymptotically optimal, branch-and-bound anytime variant of epsilon-GPP with performance guarantee. We empirically demonstrate the effectiveness of our epsilon-GPP policy and its anytime variant in Bayesian optimization and an energy harvesting task.

We consider the problem of modeling temporal textual data taking endogenous and exogenous processes into account. Such text documents arise in real world applications, including job advertisements and economic news articles, which are influenced by the fluctuations of the general economy. We propose a hierarchical Bayesian topic model which imposes a "group-correlated" hierarchical structure on the evolution of topics over time incorporating both processes, and show that this model can be estimated from Markov chain Monte Carlo sampling methods. We further demonstrate that this model captures the intrinsic relationships between the topic distribution and the time-dependent factors, and compare its performance with latent Dirichlet allocation (LDA) and two other related models. The model is applied to two collections of documents to illustrate its empirical performance: online job advertisements from DirectEmployers Association and journalists' postings on BusinessInsider.com.