The hysteria about the future of artificial intelligence (AI) is everywhere. There seems to be no shortage of sensationalist news about how AI could cure diseases, accelerate human innovation and improve human creativity. Just looking at the media headlines, you might think that we are already living in a future where AI has infiltrated every aspect of society. While it is undeniable that AI has opened up a wealth of promising opportunities, it has also led to the emergence of a mindset that can be best described as "AI solutionism". This is the philosophy that, given enough data, machine learning algorithms can solve all of humanity's problems.

Artificial intelligence are programming that copy the way people learn and take care of complex issue. These frameworks are not quite the same as different applications which mainly process exchanges and takes choices which are unequivocally modified. Such applications can't learn without anyone else.

I was recently asked to conduct a 2-hour workshop for the State of California Senior Legislators on the topic of "Big Data, Artificial Intelligence and Privacy." Honored by the privilege of offering my perspective on these critical topics, I shared with my home-state legislators how significant opportunities await the state. I reviewed the once-in-a-generation opportunities awaiting the great State of California ("the State"), where decision makers could vastly improve their constituents' quality of life, while creating new sources of value and economic growth. We have historical experiences and references to revisit in discerning what the government can do to nurture our "Analytics Revolution." Notably, the Industrial Revolution, holds many lessons regarding the consequences of late and/or confusing government involvement and guidance (see Figure 1).

Editor's Note: You can read our 2017 Ultimate Artificial Intelligence Resources Guide here. We saw the first human Go player defeated by a machine (AlphaGo), Saudi Arabia grant citizenship to an'empty-eyed humanoid' named Sophia and – more practically – the expansion of'smart speakers' like Google Home and Amazon Echo. From agricultural software to B2B marketing applications, from robotics to medical care, last year's achievements shifted the perception of AI from an emerging technology to a maturing market segment. There is no doubt that in 2018 the'race for AI' will continue to guide startup companies and tech giants alike to push to achieve that next big breakthrough. But, as the focus on AI continues to intensify, one top challenge remains: the already visible shortage of qualified AI talent.

Computers are getting ever better at recognising different faces - but on this week's Tech Tent we ask whether facial recognition technology is just too big a threat to privacy. That is certainly the view of the American Civil Liberties Union, the ACLU. This week the rights group urged Amazon to stop providing its Rekognition facial recognition technology to American police forces, saying a guide for the software "reads like a user manual for authoritarian surveillance". Amazon responded robustly, saying the quality of life would be much worse if new technologies were blocked because of how they might be used. But Matt Cagle, technology and civil liberties lawyer for the ACLU in California, says the tech firm has unlocked something really dangerous: "This technology can be turned against protesters - it can be targeted at immigrants, and it can be used to spy on entire neighbourhoods."

Policy gradient methods have been successfully applied to a variety of reinforcement learning tasks. However, while learning in a simulator, these methods do not utilise the opportunity to improve learning by adjusting certain environment variables: unobservable state features that are randomly determined by the environment in a physical setting, but that are controllable in a simulator. This can lead to slow learning, or convergence to highly suboptimal policies. In this paper, we present contextual policy optimisation (CPO). The central idea is to use Bayesian optimisation to actively select the distribution of the environment variable that maximises the improvement generated by each iteration of the policy gradient method. To make this Bayesian optimisation practical, we contribute two easy-to-compute low-dimensional fingerprints of the current policy. We apply CPO to a number of continuous control tasks of varying difficulty and show that CPO can efficiently learn policies that are robust to significant rare events, which are unlikely to be observable under random sampling but are key to learning good policies.

Internet companies are facing the need of handling large scale machine learning applications in a daily basis, and distributed system which can handle extra-large scale tasks is needed. Deep forest is a recently proposed deep learning framework which uses tree ensembles as its building blocks and it has achieved highly competitive results on various domains of tasks. However, it has not been tested on extremely large scale tasks. In this work, based on our parameter server system and platform of artificial intelligence, we developed the distributed version of deep forest with an easy-to-use GUI. To the best of our knowledge, this is the first implementation of distributed deep forest. To meet the need of real-world tasks, many improvements are introduced to the original deep forest model. We tested the deep forest model on an extra-large scale task, i.e., automatic detection of cash-out fraud, with more than 100 millions of training samples. Experimental results showed that the deep forest model has the best performance according to the evaluation metrics from different perspectives even with very little effort for parameter tuning. This model can block fraud transactions in a large amount of money \footnote{detail is business confidential} each day. Even compared with the best deployed model, deep forest model can additionally bring into a significant decrease of economic loss.

We study relations between evidence theory and S-approximation spaces. Both theories have their roots in the analysis of Dempster's multivalued mappings and lower and upper probabilities and have close relations to rough sets. We show that an S-approximation space, satisfying a monotonicity condition, can induce a natural belief structure which is a fundamental block in evidence theory. We also demonstrate that one can induce a natural belief structure on one set, given a belief structure on another set if those sets are related by a partial monotone S-approximation space.

We present a study on reinforcement learning (RL) from human bandit feedback for sequence-to-sequence learning, exemplified by the task of bandit neural machine translation (NMT). We investigate the reliability of human bandit feedback, and analyze the influence of reliability on the learnability of a reward estimator, and the effect of the quality of reward estimates on the overall RL task. Our analysis of cardinal (5-point ratings) and ordinal (pairwise preferences) feedback shows that their intra- and inter-annotator $\alpha$-agreement is comparable. Best reliability is obtained for standardized cardinal feedback, and cardinal feedback is also easiest to learn and generalize from. Finally, improvements of over 1 BLEU can be obtained by integrating a regression-based reward estimator trained on cardinal feedback for 800 translations into RL for NMT. This shows that RL is possible even from small amounts of fairly reliable human feedback, pointing to a great potential for applications at larger scale.

Society's operating system needs an upgrade. The model we have been using is simply not up to the challenges of the Fourth Industrial Revolution. A new era is unfolding at breakneck speed. It has huge potential to address some of the world's most critical challenges, from food security, to reducing congestion in big cities, to increasing energy efficiency, to accelerating cures to the most intractable diseases. But it also raises a host of social and governance issues that need addressing.