Scientists have created a robot chemist that could revolutionise the way new molecules are discovered, using machine learning techniques. Chemists from the University of Glasgow have trained an artificially-intelligent organic chemical synthesis robot to automatically explore a very large number of chemical reactions. The system, underpinned by machine-learning algorithms, can uncover new reactions and molecules. It is hoped the results could lead to lower costs for discovering new molecules for drugs, new chemical products including materials, polymers and molecules for high tech applications like imaging. The team demonstrated the system's potential by searching around 1,000 reactions using combinations of 18 different starting chemicals.

Most consumers would be dismayed with how little we know about the majority of chemicals. Only 3 percent of industrial chemicals – mostly drugs and pesticides – are comprehensively tested. Most of the 80,000 to 140,000 chemicals in consumer products have not been tested at all or just examined superficially to see what harm they may do locally, at the site of contact and at extremely high doses. I am a physician and former head of the European Center for the Validation of Alternative Methods of the European Commission (2002-2008), and I am dedicated to finding faster, cheaper and more accurate methods of testing the safety of chemicals. To that end, I now lead a new program at Johns Hopkins University to revamp the safety sciences.

We have devised and implemented a novel computational strategy for de novo design of molecules with desired properties termed ReLeaSE (Reinforcement Learning for Structural Evolution). On the basis of deep and reinforcement learning (RL) approaches, ReLeaSE integrates two deep neural networks--generative and predictive--that are trained separately but are used jointly to generate novel targeted chemical libraries. ReLeaSE uses simple representation of molecules by their simplified molecular-input line-entry system (SMILES) strings only. Generative models are trained with a stack-augmented memory network to produce chemically feasible SMILES strings, and predictive models are derived to forecast the desired properties of the de novo–generated compounds. In the first phase of the method, generative and predictive models are trained separately with a supervised learning algorithm. In the second phase, both models are trained jointly with the RL approach to bias the generation of new ...

It's been a busy week for the CRM software formerly known as ProsperWorks. The company who makes it announced that it has now officially changed its name to Copper along with and, at about the same time, launched a significantly revamped website. Is the new branding better? The company insists that "copper" connotes "timeless quality, clarity and simplicity, and its relationship to energy and currency" and that its name, according to Jon Lee, Copper CEO's and co-founder "truly represents our vision and plans for the future of the CRM industry." To Lee, people (like minerals such as copper) are "every company's most valuable resource and it's critical that CRM reflects that mindset and provides systems that put people's needs first." Copper's CRM application aims to be a valuable resource (like copper) that makes a company's most valuable resource (its people) more productive and efficient.

It's often the case that the more useful a robot is, the less exciting it is. The robots that do the hardest jobs tend to be straightforward solutions to straightforward problems, because that's what works. The (self-declared) world's largest robot is an efficient, grubby example of this--it's an autonomous train that recently hauled 28,000 metric tons of iron ore 280 kilometers across the Australian desert. Australia is a big place, and it takes a lot of effort to get material out of the middle of Australia (where it's not useful) to the coast (where it can be taken somewhere that it is). Trains are the most efficient way of doing this, and they travel back and forth through a whole lot of nothing, taking ore from mine to port and bringing the empty cars back again.

A demonstration video that veteran University College, London neuroscientist John O'Keefe often presents in lectures shows a rat moving around the inside of a box. Every time the rat heads for the top-left corner, loud pops play through a speaker; those sounds are the result of the firing of a specific neuron attached to an electrode. The neuron only fires when the rat moves to the same small area of the box. This connection of certain neurons to locations led O'Keefe and student Jonathon Dostrovsky to name those neurons "place cells" when they encountered the phenomenon in the early 1970s. Today, researchers such as Huajin Tang, director of the Neuromorphic Computing Research Center at Sichuan University, China, are using maps of computer memory to demonstrate how simulated neurons fire in much the same way inside one of their wheeled robots.

Most existing knowledge graphs (KGs) in academic domains suffer from problems of insufficient multi-relational information, name ambiguity and improper data format for large-scale machine pro- cessing. In this paper, we present AceKG, a new large-scale KG in academic domain. AceKG not only provides clean academic information, but also offers a large-scale benchmark dataset for researchers to conduct challenging data mining projects including link prediction, community detection and scholar classification. Specifically, AceKG describes 3.13 billion triples of academic facts based on a consistent ontology, including necessary properties of papers, authors, fields of study, venues and institutes, as well as the relations among them. To enrich the proposed knowledge graph, we also perform entity alignment with existing databases and rule-based inference. Based on AceKG, we conduct experiments of three typical academic data mining tasks and evaluate several state-of- the-art knowledge embedding and network representation learning approaches on the benchmark datasets built from AceKG. Finally, we discuss several promising research directions that benefit from AceKG.

A new type of plastic that can heal itself when damaged could mean satellites will be able to stay in orbit for longer, scientists have revealed. The polymer heals cracks when exposed to certain light by converting from a rigid structure to a much softer, malleable substance. Under certain conditions, the plastic used by researchers could become up to ten times softer and more dynamic. Such plastics could also be used to coat vehicles on Earth, including cars, giving them the ability to heal after being involved in crashes. A new type of plastic that can heal itself when damaged could mean satellites may stay in orbit for even longer, scientists have revealed.

Maura Kolb is more familiar with analyzing rock samples than computer code, but as exploration manager at Goldcorp's Red Lake gold mine, she now finds herself in charge of an innovative approach to exploration that involves both. The project, in partnership with IBM, uses IBM's Watson artificial intelligence platform to identify exploration targets at the Ontario mine. Since the project was launched in March 2017 at the Disrupt Mining event at PDAC, Goldcorp says the time it takes to process survey data has fallen by 97 per cent. "Being able to go through [our data] quickly, that's something no one else can do right now," said Kolb. "We can ask questions of our data that we haven't been able to in the past." The main goal of the project is to use Watson to help identify high-grade areas at Red Lake that were previously overlooked by human eyes.

Machine learning is beginning to make a large impact in catalysis research, according to Bryan Goldsmith, Jacques Esterhuizen, and Jin-Xun Liu of the Univ. of Michigan, Christopher Bartel of the Univ. of Colorado Boulder, and Christopher Sutton of the Fritz Haber Institute of the Max Planck Society in their July AIChE Journal Perspective article, "Machine Learning for Heterogeneous Catalyst Design and Discovery." Novel catalysts are crucial for several applications, such as energy generation and storage, sustainable chemical production, and pollution mitigation. The current trial-and-error approaches to new catalyst discovery and synthesis are expensive and time-consuming. As an alternative, machine learning can be used to identify the top catalyst candidates before experimental testing, thereby accelerating catalyst discovery and design. Goldsmith and colleagues highlight several examples where machine learning is making an impact on heterogeneous catalysis research, such as: accelerating the determination of catalyst active sites and catalyst screening; finding descriptors and patterns in catalysis data; determining interatomic potentials for catalyst simulation; and discovering and analyzing catalytic mechanisms.