Much of modern-day agriculture is dominated by monoculture, the practice of producing a single crop on a large swath of land. This approach makes it easier for farmers to manage their fields with tractors and other basic automated tools, but it also strips the soil of nutrients and reduces its productivity. As a result, many farmers rely heavily on nitrogen-based fertilizers, which can convert into nitrous oxide, a greenhouse gas 300 times more potent than carbon dioxide. Robots run on machine-learning software could help farmers manage a mix of crops more effectively at scale, while algorithms could help farmers predict what crops to plant when, regenerating the health of their land and reducing the need for fertilizers.

Once billionaire Jeff Bezos' Blue Origin lander makes it to the moon, the Amazon CEO says it won't have to go very far to re-fuel. In a space summit in Boston, Bezos told an audience that his somewhat mysterious moon lander will use ice harvested from the lunar surface to create fuel. 'We know things about the moon now we didn't know about during the Apollo days,' Bezos said at the conference as reported by CNBC. 'We can harvest that ice and use to make hydrogen and oxygen, which are rocket propellants.' Jeff Bezos says a recently discovered trove of water and ice in the moon's surface could fuel a lunar lander owned by his company Blue Origin.

AI is getting smarter, requiring less training data and moving from cloud to Edge. Finnish AI startups gathered last week in Business Finland's Customer Club to share relevant information for intelligent industry and to check the latest state of the art of AI solutions for industrial use. There are plenty of small, young Finnish companies that have created money saving and innovative AI solutions especially for pulp and paper industry, mining companies and oil refineries that are strong businesses in Finland. Possibilities for different profitable applications are numerous with solutions that combine the use of cloud and edge in storing and analyzing data. All data from industrial machines cannot be moved to the cloud because there is typically just too much data or the latency requirements don't allow it.

Traditionally, supercomputer performance is measured using the High-Performance Linpack (HPL) benchmark, which is the basis for the Top500 list that biannually ranks world's fastest supercomputers. The Linpack benchmark tests a supercomputer's ability to conduct high-performance tasks (like simulations) that use double-precision math. On June's Top500 list, announced Monday, Summit's 148 Linpack petaflops land it first place by a comfortable margin. Using that same machine configuration, Oak Ridge National Laboratory (ORNL) and Nvidia have tested Summit on HPL-AI and gotten a result of 445 petaflops. While the HPL benchmark tests supercomputers' performance in double-precision math, AI is a rapidly growing use case for supercomputers -- and most AI models use mixed-precision math.

"I see Amazon as a technology company that just happened to do retail," begins Werner Vogels, Amazon's chief technology officer. "When Jeff [Bezos] started Amazon, he wasn't thinking about starting a bookshop. He was really fascinated by the internet." Only "mortal humans", he tells me in an interview, ever saw Amazon as merely a retailer. So the question now is: what will Amazon become next?

NASA says its presence on the moon won't just be for show. With new technology, the agency hopes to mine natural resources on the lunar surface as well as reachable asteroids. Through NASA's Innovative Advanced Concepts (NIAC) program, the agency said it will begin to explore the feasibility of robotic rovers and mining technology that could make space mining a reality. To do so, it has green-lit two mission concepts this month. NASA wants to get a jump-start on mining in space with a tandem of proposals that would develop future technology.

The scientist from the 1950s has proven how amino acid building blocks can emerge from nothing, so to speak. All it takes is methane, ammonia, water, hydrogen, and electricity, in a certain mixture, at certain temperatures, time and iterations. What was it like 4 billion years ago, when the first amino acids decided to join together to form living cells? The first unicellular organisms formed 100% of the known life in the Precambrian age and for what it's worth, all the following life as well. Stromatolites can still be found today, in shelf areas, always formed by so-called cyanobacteria, which are able to produce oxygen.

Climate change is one of the greatest challenges facing humanity, and we, as machine learning experts, may wonder how we can help. Here we describe how machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate. From smart grids to disaster management, we identify high impact problems where existing gaps can be filled by machine learning, in collaboration with other fields. Our recommendations encompass exciting research questions as well as promising business opportunities. We call on the machine learning community to join the global effort against climate change.

Specifying utility functions is a key step towards applying the discrete choice framework for understanding the behaviour processes that govern user choices. However, identifying the utility function specifications that best model and explain the observed choices can be a very challenging and time-consuming task. This paper seeks to help modellers by leveraging the Bayesian framework and the concept of automatic relevance determination (ARD), in order to automatically determine an optimal utility function specification from an exponentially large set of possible specifications in a purely data-driven manner. Based on recent advances in approximate Bayesian inference, a doubly stochastic variational inference is developed, which allows the proposed DCM-ARD model to scale to very large and high-dimensional datasets. Using semi-artificial choice data, the proposed approach is shown to very accurately recover the true utility function specifications that govern the observed choices. Moreover, when applied to real choice data, DCM-ARD is shown to be able discover high quality specifications that can outperform previous ones from the literature according to multiple criteria, thereby demonstrating its practical applicability.

There has been a recent trend in training neural networks to replace data structures that have been crafted by hand, with an aim for faster execution, better accuracy, or greater compression. In this setting, a neural data structure is instantiated by training a network over many epochs of its inputs until convergence. In applications where inputs arrive at high throughput, or are ephemeral, training a network from scratch is not practical. This motivates the need for few-shot neural data structures. In this paper we explore the learning of approximate set membership over a set of data in one-shot via meta-learning. We propose a novel memory architecture, the Neural Bloom Filter, which is able to achieve significant compression gains over classical Bloom Filters and existing memory-augmented neural networks.