Advances in artificial intelligence (AI) will transform modern life by reshaping transportation, health, science, finance, and the military. To adapt public policy, we need to better anticipate these advances. Here we report the results from a large survey of machine learning researchers on their beliefs about progress in AI. Researchers predict AI will outperform humans in many activities in the next ten years, such as translating languages (by 2024), writing high-school essays (by 2026), driving a truck (by 2027), working in retail (by 2031), writing a bestselling book (by 2049), and working as a surgeon (by 2053). Researchers believe there is a 50% chance of AI outperforming humans in all tasks in 45 years and of automating all human jobs in 120 years, with Asian respondents expecting these dates much sooner than North Americans. These results will inform discussion amongst researchers and policymakers about anticipating and managing trends in AI. This article is part of the special track on AI and Society.

You will learn how to use Python to analyze data (big data analytics), create beautiful visualizations (data visualization) and use powerful machine learning algorithms. You will specifically get to learn how to use NumPy, Seaborn, Matplotlib, Pandas, Scikit-Learn, Machine Learning, Plotly, Tensorflow and more.

Hundreds of companies and thousands of individuals, many of them researchers and engineers prominent in the fields of robotics and artificial intelligence, vowed on Wednesday never to apply their skills toward the creation of autonomous killing machines. Led by the Future of Life Institute, a Boston-based nonprofit, as many as 160 AI-related companies in 36 countries, and 2,400 individuals in 90 countries, signed the pledge stating that autonomous weapons posed a "clear and present danger to the citizens of every country in the world," and that they would not participate in their development. "Artificial intelligence (AI) is poised to play an increasing role in military systems," the pledge states. "There is an urgent opportunity and necessity for citizens, policymakers, and leaders to distinguish between acceptable and unacceptable uses of AI." The signatories, who join 26 United Nations countries that have explicitly called for a ban on lethal autonomous weapons, include DeepMind, Google's top AI research team; the European Association for AI; ClearPath Robotics/OTTO Motors; the XPRIZE Foundation, the Swedish AI Society; and University College of London, among others. Leading AI researchers Demis Hassabis, Stuart Russell, Yoshua Bengio, Anca Dragan, Toby Walsh, and Tesla and SpaceX founder Elon Musk are among the individuals who also signed the pledge.

Humans should get'out of the loop' of artificial intelligence systems, UTS roboticist Professor Mary-Anne Williams argued last week at an Australian Human Rights Commission technology conference in Sydney. AI needn't consult a flesh and blood individual even when making life or death decisions, said Williams, director of The Magic Lab at the university's Centre of Artificial Intelligence. The rise of autonomous weapons systems which operate without human control is being campaigned against around the world. "States must draw the line now against unchecked autonomy in weapon systems by ensuring that the decision to take human life is never delegated to a machine," the Campaign to Stop Killer Robots states. In Australia, 122 AI experts last year signed a letter to Prime Minister Malcolm Turnbull urging him to "take a firm global stand" against weapons systems that remove "meaningful human control" when selecting targets and deploying lethal force.

This speech was given by Dr Larry Marshall at the AFR Innovation Summit on Monday 30 July 2018. I would like to begin by acknowledging the Gadigal people of the Eora nation as the Traditional Owners of the land that we are on today, and pay my respect to their Elders past and present. It's great to be back at the AFR Innovation Summit, and thank you to our morning speakers for setting a strong focus on the power of innovation to shape our future. Before I was the Chief Executive of Australia's national science agency, I was an inventor, an entrepreneur, a venture capitalist and of course a kid. I grew up in a time where people were obsessed with the power of the computer and its ability to replace humans, our parents told us to study computer programing.

We consider Markov models of stochastic processes where the next-step conditional distribution is defined by a kernel density estimator (KDE), similar to Markov forecast densities and certain time-series bootstrap schemes. The KDE Markov models (KDE-MMs) we discuss are nonlinear, nonparametric, fully probabilistic representations of stationary processes, based on techniques with strong asymptotic consistency properties. The models generate new data by concatenating points from the training data sequences in a context-sensitive manner, together with some additive driving noise. We present novel EM-type maximum-likelihood algorithms for data-driven bandwidth selection in KDE-MMs. Additionally, we augment the KDE-MMs with a hidden state, yielding a new model class, KDE-HMMs. The added state variable captures non-Markovian long memory and signal structure (e.g., slow oscillations), complementing the short-range dependences described by the Markov process. The resulting joint Markov and hidden-Markov structure is appealing for modelling complex real-world processes such as speech signals. We present guaranteed-ascent EM-update equations for model parameters in the case of Gaussian kernels, as well as relaxed update formulas that greatly accelerate training in practice. Experiments demonstrate increased held-out set probability for KDE-HMMs on several challenging natural and synthetic data series, compared to traditional techniques such as autoregressive models, HMMs, and their combinations.

Deep neural networks bring in impressive accuracy in various applications, but the success often relies on the heavy network architecture. Taking well-trained heavy networks as teachers, classical teacher-student learning paradigm aims to learn a student network that is lightweight yet accurate. In this way, a portable student network with significantly fewer parameters can achieve a considerable accuracy which is comparable to that of teacher network. However, beyond accuracy, robustness of the learned student network against perturbation is also essential for practical uses. Existing teacher-student learning frameworks mainly focus on accuracy and compression ratios, but ignore the robustness. In this paper, we make the student network produce more confident predictions with the help of the teacher network, and analyze the lower bound of the perturbation that will destroy the confidence of the student network. Two important objectives regarding prediction scores and gradients of examples are developed to maximize this lower bound, so as to enhance the robustness of the student network without sacrificing the performance. Experiments on benchmark datasets demonstrate the efficiency of the proposed approach to learn robust student networks which have satisfying accuracy and compact sizes.

We propose to incorporate adversarial dropout in generative multi-adversarial networks, by omitting or dropping out, the feedback of each discriminator in the framework with some probability at the end of each batch. Our approach forces the single generator not to constrain its output to satisfy a single discriminator, but, instead, to satisfy a dynamic ensemble of discriminators. We show that this leads to a more generalized generator, promoting variety in the generated samples and avoiding the common mode collapse problem commonly experienced with generative adversarial networks (GANs). We further provide evidence that the proposed framework, named Dropout-GAN, promotes sample diversity both within and across epochs, eliminating mode collapse and stabilizing training.

Social media has become prevalent and important for social networking and opinion sharing in recent years [1]. By changing the way we perceive and interact with the world, social media has changed our lives profoundly [2], [3]. With millions of posts and replies uploaded every day on social media such as Facebook, Twitters and YouTube, it is an abundant and informative data source of public opinions; thus, it has attracted lots of attention from both academia and industry to understand people and society [4]-[6]. Most previous text mining-based social media analysis focused on Twitter and Facebook [7]. YouTube, generally considered as a video platform, the values of its text comments below videos have long been underestimated. Being the second most popular website in the world [8] and having 1.9 billion active users [9], YouTube is an attractive source of research in social media analysis with immense potentials. Recent developments in autonomous vehicle technology have helped bring self-driving vehicles to the forefront of public interest [10].

Answer Set Programming (ASP) is a well-known declarative problem solving approach based on nonmonotonic logic programs, which has been successfully applied to a wide range of applications in artificial intelligence and beyond. To address the needs of modern applications, HEX-programs were introduced as an extension of ASP with external atoms for accessing information outside programs via an API style bi-directional interface mechanism. To evaluate such programs, conflict-driving learning algorithms for SAT and ASP solving have been extended in order to capture the semantics of external atoms. However, a drawback of the state-of-the-art approach is that external atoms are only evaluated under complete assignments (i.e., input to the external source) while in practice, their values often can be determined already based on partial assignments alone (i.e., from incomplete input to the external source). This prevents early backtracking in case of conflicts, and hinders more efficient evaluation of HEX-programs. We thus extend the notion of external atoms to allow for three-valued evaluation under partial assignments, while the two-valued semantics of the overall HEX-formalism remains unchanged. This paves the way for three enhancements: first, to evaluate external sources at any point during model search, which can trigger learning knowledge about the source behavior and/or early backtracking in the spirit of theory propagation in SAT modulo theories (SMT). Second, to optimize the knowledge learned in terms of so-called nogoods, which roughly speaking are impossible input-output configurations. Shrinking nogoods to their relevant input part leads to more effective search space pruning. And third, to make a necessary minimality check of candidate answer sets more efficient by exploiting early external evaluation calls. As this check usually accounts for a large share of the total runtime, optimization is here particularly important. We further present an experimental evaluation of an implementation of a novel HEX-algorithm that incorporates these enhancements using a benchmark suite. Our results demonstrate a clear efficiency gain over the state-of-the-art HEX-solver for the benchmarks, and provide insights regarding the most effective combinations of solver configurations.