Input gradient regularization is not thought to be an effective means for promoting adversarial robustness. In this work we revisit this regularization scheme with some new ingredients. First, we derive new per-image theoretical robustness bounds based on local gradient information, and curvature information when available. These bounds strongly motivate input gradient regularization. Second, we implement a scaleable version of input gradient regularization which avoids double backpropagation: adversarially robust ImageNet models are trained in 33 hours on four consumer grade GPUs. Finally, we show experimentally that input gradient regularization is competitive with adversarial training.

Quivering and hesitant, like a spoon-wielding toddler trying to eat soup without spilling it, the world's first raspberry-picking robot is attempting to harvest one of the fruits. After sizing it up for an age, the robot plucks the fruit with its gripping arm and gingerly deposits it into a waiting punnet. The whole process takes about a minute for a single berry. It seems like heavy going for a robot that cost £700,000 to develop but, if all goes to plan, this is the future of fruit-picking. Each robot will be able to pick more than 25,000 raspberries a day, outpacing human workers who manage about 15,000 in an eight-hour shift, according to Fieldwork Robotics, a spinout from the University of Plymouth.

The growing use of artificial intelligence comes with more than just technology challenges. Ethical questions are emerging as government agencies use AI for a range of purposes, from customer service chatbots to mission-critical tasks supporting military forces. This increasing reliance on AI introduces concerns about bias in its foundation, especially related to information on gender, race, socioeconomic status and age. Potential for bias can be built into AI algorithms by the humans who create them. This can happen intentionally or by accident resulting from biases developers don't realize they have.

Application development workloads keep growing, but developer teams are not. If anything, development skills are increasingly in precious short supply. That's the word from the latest survey of 3,300 IT leaders, conducted by OutSystems. The development skills shortage has been a crisis raging for a number of years now, and this latest survey shows no sign of abating. Fueling the demand is the rising tide of digital transformation, and with it, the reliance of business leaders on technology to amp up the customer experience and compete on data analytics.

NIIGATA/KYOTO - Self-driving tractors, tomato-picking robots, camera-mounted drones to survey fields and spot crop damage, and satellite data from the Japan Aerospace Exploration Agency (JAXA) to help farms keep track of climate and weather data. At over a dozen booths beside the G20 farm ministers' meeting venue earlier this month in the Sea of Japan city of Niigata, agricultural organizations and technology firms touted products and services they see as necessary tools to ensure a prosperous future for agriculture. "In today's Japan, the aging of farmers has become an issue, and the overall population of the country is decreasing. Collaboration between agriculture and nonagricultural sectors, such as satellite technology, IoT ("internet of things," internet connectivity into physical devices like tractors) and artificial intelligence has a key role to play in fostering agricultural innovation," said Susumu Hamamura, parliamentary vice minister at the Ministry of Agriculture, Forestry and Fisheries. The increased use of easily accessible data on tablet computers and smartphones to provide farmers with a wide range of agricultural data was a key message at the Niigata conference.

The Organisation for Economic Co-operation and Development unveiled the first intergovernmental standard for artificial intelligence policies Wednesday--and the organization's 36 member countries including America have initially signed on along with Argentina, Brazil, Colombia, Costa Rica, Peru and Romania. OECD, an international forum that unites stakeholders from many nations to work together to address challenges of globalization, released "Recommendations of the Council on Artificial Intelligence" to help foster a global policy ecosystem that leverages the evolving technology's benefits, while also protecting human rights and democratic values. OECD's Director of the Science, Technology and Innovation Directorate Andrew Wyckoff told reporters that the principles' creators hope they'll help shape a stable regulatory environment that promotes the tech's positive uses, while withstanding unethical abuses. "AI is what we would call a'general purpose technology.' It's going to change the way we do things in nearly every single sector of the economy--that's part of the reason we give so much importance to its development," he said.

In this work we address the problem of argument search. The purpose of argument search is the distillation of pro and contra arguments for requested topics from large text corpora. In previous works, the usual approach is to use a standard search engine to extract text parts which are relevant to the given topic and subsequently use an argument recognition algorithm to select arguments from them. The main challenge in the argument recognition task, which is also known as argument mining, is that often sentences containing arguments are structurally similar to purely informative sentences without any stance about the topic. In fact, they only differ semantically. Most approaches use topic or search term information only for the first search step and therefore assume that arguments can be classified independently of a topic. We argue that topic information is crucial for argument mining, since the topic defines the semantic context of an argument. Precisely, we propose different models for the classification of arguments, which take information about a topic of an argument into account. Moreover, to enrich the context of a topic and to let models understand the context of the potential argument better, we integrate information from different external sources such as Knowledge Graphs or pre-trained NLP models. Our evaluation shows that considering topic information, especially in connection with external information, provides a significant performance boost for the argument mining task.

Distributionally robust optimization (DRO) has attracted attention in machine learning due to its connections to regularization, generalization, and robustness. Existing work has considered uncertainty sets based on phi-divergences and Wasserstein distances, each of which have drawbacks. In this paper, we study DRO with uncertainty sets measured via maximum mean discrepancy (MMD). We show that MMD DRO is roughly equivalent to regularization by the Hilbert norm and, as a byproduct, reveal deep connections to classic results in statistical learning. In particular, we obtain an alternative proof of a generalization bound for Gaussian kernel ridge regression via a DRO lense. The proof also suggests a new regularizer. Our results apply beyond kernel methods: we derive a generically applicable approximation of MMD DRO, and show that it generalizes recent work on variance-based regularization.

In many scientific fields, such as economics and neuroscience, we are often faced with nonstationary time series, and concerned with both finding causal relations and forecasting the values of variables of interest, both of which are particularly challenging in such nonstationary environments. In this paper, we study causal discovery and forecasting for nonstationary time series. By exploiting a particular type of state-space model to represent the processes, we show that nonstationarity helps to identify causal structure and that forecasting naturally benefits from learned causal knowledge. Specifically, we allow changes in both causal strengths and noise variances in the nonlinear state-space models, which, interestingly, renders both the causal structure and model parameters identifiable. Given the causal model, we treat forecasting as a problem in Bayesian inference in the causal model, which exploits the time-varying property of the data and adapts to new observations in a principled manner. Experimental results on synthetic and real-world data sets demonstrate the efficacy of the proposed methods.

A recent study from the Federal Reserve Bank of San Francisco showed that the U.S. labor market is "at or beyond its full potential." While that might be good news for the economy, it's not for the tech companies still out there, searching desperately for qualified talent. Research firm Korn Ferry predicted that companies in technology, media and telecom will face a talent shortage of 1.1 million by 2020. By 2030, that number will grow to 4.3 million, the report said; and the result will be intense competition for a dwindling pool of tech professionals. The same survey showed that insufficient staff is already holding back digital transformation at 54 percent of companies surveyed.