The head of the US government's Intelligence Advanced Research Projects Activity (IARPA) is leaving the unit this month to take a job with a quantum computing company, WIRED has learned. Rick Muller's pending departure from IARPA comes amid broader efforts to downsize the United States intelligence community, including the Office of the Director of National Intelligence (ODNI), which oversees IARPA. A person familiar with Muller's plans confirmed to WIRED his departure from IARPA. Born during the aftermath of the September 11, 2001 terrorist attacks, IARPA is tasked with testing AI, quantum computing, and other emerging technologies that could aid the missions of spy agencies including the Central Intelligence Agency and National Security Agency. The Trump administration reportedly has been moving to cut the workforces of intelligence agencies as part of the president's broad efforts to dismantle diversity programs and streamline government operations.

The program optimizes three main areas of shipping logistics with the benefits of A.I.: use insights to make the FedEx network operate more efficiently; give customers more visibility and control over their supply chains; and solve e-commerce-related challenges. It's a big undertaking but one that Sriram Krishnasamy, chief executive officer of FedEx Dataworks, says is the direction that logistics companies lean today: using data-driven, digital insights to inform decision-making and increase transparency. In the first six months, the team launched a predictive insights-based sensor data for added visibility for packages, which he says had an immediate impact on vaccine distribution. At the time of publishing, the team has rolled out more than 40 additional solutions currently being used by FedEx and its partners. "For most companies, the data generated across every aspect of the supply chain is something that they are simply trying to'manage' rather than cultivate," says Krishnasamy.

Human-Centered AI (HCAI) refers to the research effort that aims to design and implement AI techniques to support various human tasks, while taking human needs into consideration and preserving human control. In this short position paper, we illustrate how we approach HCAI using a series of research projects around Data Science (DS) works as a case study. The AI techniques built for supporting DS works are collectively referred to as AutoML systems, and their goals are to automate some parts of the DS workflow. We illustrate a three-step systematical research approach (i.e., explore, build, and integrate) and four practical ways of implementation for HCAI systems. We argue that our work is a cornerstone towards the ultimate future of Human-AI Collaboration for DS and beyond, where AI and humans can take complementary and indispensable roles to achieve a better outcome and experience.

In recent years, the use of Machine Learning (ML) in computational chemistry has enabled numerous advances previously out of reach due to the computational complexity of traditional electronic-structure methods. One of the most promising applications is the construction of ML-based force fields (FFs), with the aim to narrow the gap between the accuracy of ab initio methods and the efficiency of classical FFs. The key idea is to learn the statistical relation between chemical structure and potential energy without relying on a preconceived notion of fixed chemical bonds or knowledge about the relevant interactions. Such universal ML approximations are in principle only limited by the quality and quantity of the reference data used to train them. This review gives an overview of applications of ML-FFs and the chemical insights that can be obtained from them. The core concepts underlying ML-FFs are described in detail and a step-by-step guide for constructing and testing them from scratch is given. The text concludes with a discussion of the challenges that remain to be overcome by the next generation of ML-FFs.

Domain translation is the task of finding correspondence between two domains. Several Deep Neural Network (DNN) models, e.g., CycleGAN and cross-lingual language models, have shown remarkable successes on this task under the unsupervised setting---the mappings between the domains are learned from two independent sets of training data in both domains (without paired samples). However, those methods typically do not perform well on a significant proportion of test samples. In this paper, we hypothesize that many of such unsuccessful samples lie at the fringe---relatively low-density areas---of data distribution, where the DNN was not trained very well, and propose to perform Langevin dynamics to bring such fringe samples towards high density areas. We demonstrate qualitatively and quantitatively that our strategy, called Langevin Cooling (L-Cool), enhances state-of-the-art methods in image translation and language translation tasks.

By accurately predicting industrial aging processes (IAPs), it is possible to schedule maintenance events further in advance, thereby ensuring a cost-efficient and reliable operation of the plant. So far, these degradation processes were usually described by mechanistic models or simple empirical prediction models. In this paper, we evaluate a wider range of data-driven models for this task, comparing some traditional stateless models (linear and kernel ridge regression, feed-forward neural networks) to more complex recurrent neural networks (echo state networks and LSTMs). To examine how much historical data is needed to train each of the models, we first examine their performance on a synthetic dataset with known dynamics. Next, the models are tested on real-world data from a large scale chemical plant. Our results show that LSTMs produce near perfect predictions when trained on a large enough dataset, while linear models may generalize better given small datasets with changing conditions.

SciTech Europa delves into the world of AI, defining what it means, giving examples of the real-life applications, and discussing the ethical questions it prompts. The computer scientist John McCarthy coined the term Artificial Intelligence in 1956, and defines the field of artificial intelligence as "the science and engineering of making intelligent machines." As well as the term for the scientific discipline, artificial intelligence refers to the intelligence of a machine, program, or system, in contrast to that of human intelligence. Alessandro Annoni, the head of the European Commission's Joint Research Centre, spoke at the Science Meets Parliaments conference at the European Parliament, Brussels in February 2019. He said: "Artificial intelligence should not be considered a simple technology…it is a collection of technologies. It is a new paradigm that is aiming to give more power to the machine. It's a technology that will replace humans in some cases."

We consider the problem of embedding a relation, represented as a directed graph, into Euclidean space. For three types of embeddings motivated by the recent literature on knowledge graphs, we obtain characterizations of which relations they are able to capture, as well as bounds on the minimal dimensionality and precision needed.

This year's sold out #shoptalk2019 conference was both engaging and enlightening. The event brought together 8,500 participants from the retail ecosystem to share trends, technologies, and innovations on how to be a better retailer – particularly in the age of digital. Anil Aggarwal, CEO of Shoptalk, introduced himself to the crowd with the comment, "[Retail] needs to transform for a modern digital age." For Anil, this was an imperative, not just a nice to have. For me, however, the experience was even more lucid than Anil's interpretation – especially around AI technology.

Image SEO used to be a huge aspect of content and site optimization, with the potential to drive tons of image search traffic. File names, alt tags, and image sitemaps were all super important. The "view image" button was added to Google Images in 2013. With the new implementation, sites saw an average decrease of 63 percent in image search traffic. While image optimization is still an extensive practice, it hasn't been effective for driving much traffic.