A person receives a test for diabetes during Care Harbor LA free medical clinic in Los Angeles, California September 11, 2014. DreaMed Diabetes, the Petah Tikva-based developer of personalized diabetes management solutions, has received US Food and Drug Administration (FDA) clearance for its artificial intelligence-powered insulin recommendations technology. The company's AI-based insulin dosing decision-support software, DreaMed Advisor Pro, aims to assist people with Type 1 diabetes (T1D) using insulin-pump therapy with continuous glucose sensors or blood glucose meters.

In the past few weeks, two important developments in artificial intelligence research have gone largely unheralded. Both hint at just how earth shaking – or at least industry-shattering – A.I.'s potential really is. The first item was news that a Hong Kong-based biotechnology startup, InSilico Medicine, working with researchers from the University of Toronto, had used machine learning to create a potential new drug to prevent tissue scarring. What's eye-popping here is the timescale: just 46 days from molecular design to animal testing in mice. Considering that, on average, it takes more than a decade and costs $350 million to $2.7 billion to bring a new drug to market, depending on which study one believes, the potential impact on the pharmaceutical industry is huge.

When terrorist bombings rocked Sri Lanka in April 2019, the Dataminr alerted its clients 26 minutes before conventional news outlets. It also kept the information rolling with more than 200 updates as the attacks unfolded. With its monitoring of social media and alternative data sources, the Dataminr platform sifted through the data landscape to identify and distribute the most relevant information to clients in real-time. For corporates with business operations and key people in the danger zone, the real-time alerts were critical intelligence. It helped them to frame their crisis management response to keep their assets safe.

Artificial intelligence is infiltrating every industry, allowing vehicles to navigate without drivers, assisting doctors with medical diagnoses, and mimicking the way humans speak. But for all the authentic and exciting ways it's transforming the tasks computers can perform, there's a lot of hype, too. As Jeremy Achin, CEO of newly minted unicorn DataRobot, puts it: "Everyone knows you have to have machine learning in your story or you're not sexy." The inherently broad term gets bandied about so often that it can start to feel meaningless and can be trotted out by companies to gussy up even simple data analysis. To help cut through the noise, Forbes and data partner Meritech Capital put together a list of private, U.S.-based companies that are wielding some subset of artificial intelligence in a meaningful way and demonstrating real business potential from doing so. One makes robots that can whir around shoppers to help workers restock shelves. Another scans recruiting pitches for unconscious bias. A third analyzes massive data sets to make street-by-street weather predictions. To be included on the list, companies needed to show that techniques like machine learning (where systems learn from data to improve on tasks), natural language processing (which enables programs to "understand" written or spoken language), or computer vision (which relates to how machines "see") are a core part of their business model and future success. Find all the details on our methodology here. The honorees span categories like human resources, security, insurance, and finance, with healthcare, transportation, and infrastructure startups best represented on the list.

At the recent ApacheCon North America, Denis Magda spoke on continuous machine learning with Apache Ignite, an in-memory data grid. Ignite simplifies the machine-learning pipeline by performing training and hosting models in the same cluster that stores the data, and can perform "online" training to incrementally improve models when new data is available. Magda, vice-president of product management at GridGain, began by describing some of the pain points of machine learning on large datasets, in particular the latency involved in moving data across the network from its storage location to the processors that perform training. Models also have to be deployed into a production system after they are trained, and retrained periodically after new data is collected. Because Ignite runs code on the same computers that host data, it can train, deploy, and update a machine-learning model without a time-consuming extract-transform-load (ETL) step.

How many of you believe robots and artificial intelligence will take jobs? How many of you believe machines will take your job? Robots, AI and other disruptive new technologies are expected to displace a significant number of office and manual labor jobs that pay $20 to $40 an hour, according to a 2014 Pew Research Center report. It won't happen all at once, and that's a blessing and a challenge. Slowly, then quickly, machines will replace certain human jobs. We might not hear about most of them because they'll happen in pockets of geographies and industries. But in the not too distant future, we'll see the great extent that robots are among us, and that people will no longer be able to apply for the jobs that some humans work at today. How robotics and AI will change jobs has been top of my mind since the national election's emphasis on bringing back jobs to the USA.

DUBAI, UNITED ARAB EMIRATES – The assault on the beating heart of Saudi Arabia's vast oil empire follows a new and dangerous pattern that's emerged across the Persian Gulf this summer of precise attacks that leave few obvious clues as to who launched them. Beginning in May with the still-unclaimed explosions that damaged oil tankers near the Strait of Hormuz, the region has seen its energy infrastructure repeatedly targeted. Those attacks culminated with Saturday's assault on the world's biggest oil processor in eastern Saudi Arabia, which halved the oil-rich kingdom's production and caused energy prices to spike. Some strikes have been claimed by Yemen's Houthi rebels, who have been battling a Saudi-led coalition in the Arab world's poorest country since 2015. Their rapidly increasing sophistication fuels suspicion among experts and analysts however that Iran may be orchestrating them -- or perhaps even carrying them out itself as the U.S. alleges in the case of Saturday's attack.

Richard Stallman, pictured in 2015, resigned from his posts as President of the Free Software Foundation and visiting scientist at MIT's Computer Science & Artificial Intelligence lab. Richard Stallman, pictured in 2015, resigned from his posts as President of the Free Software Foundation and visiting scientist at MIT's Computer Science & Artificial Intelligence lab. Free software pioneer and renowned computer scientist Richard Stallman resigned from his post at MIT following recent comments about one of Jeffrey Epstein's sex-trafficking victims. He also resigned as president of the Free Software Foundation. On Monday, Stallman, a visiting scientist at the university's Computer Science and Artificial Intelligence Laboratory, posted a brief message on his blog announcing the decision.

First principles modeling of physical systems has led to significant technological advances across all branches of science. For nonlinear systems, however, small modeling errors can lead to significant deviations from the true, measured behavior. Even in mechanical systems, where the equations are assumed to be well-known, there are often model discrepancies corresponding to nonlinear friction, wind resistance, etc. Discovering models for these discrepancies remains an open challenge for many complex systems. In this work, we use the sparse identification of nonlinear dynamics (SINDy) algorithm to discover a model for the discrepancy between a simplified model and measurement data. In particular, we assume that the model mismatch can be sparsely represented in a library of candidate model terms. We demonstrate the efficacy of our approach on several examples including experimental data from a double pendulum on a cart. We further design and implement a feed-forward controller in simulations, showing improvement with a discrepancy model.

Bio-oil molecule assessment is essential for the sustainable development of chemicals and transportation fuels. These oxygenated molecules have adequate carbon, hydrogen, and oxygen atoms that can be used for developing new value-added molecules (chemicals or transportation fuels). One motivation for our study stems from the fact that a liquid phase upgrading using mineral acid is a cost-effective chemical transformation. In this chemical upgrading process, adding a proton (positively charged atomic hydrogen) to an oxygen atom is a central step. The protonation energies of oxygen atoms in a molecule determine the thermodynamic feasibility of the reaction and likely chemical reaction pathway. A quantum chemical model based on coupled cluster theory is used to compute accurate thermochemical properties such as the protonation energies of oxygen atoms and the feasibility of protonation-based chemical transformations. However, this method is too computationally expensive to explore a large space of chemical transformations. We develop a graph neural network approach for predicting protonation energies of oxygen atoms of hundreds of bioxygenate molecules to predict the feasibility of aqueous acidic reactions. Our approach relies on an iterative local nonlinear embedding that gradually leads to global influence of distant atoms and a output layer that predicts the protonation energy. Our approach is geared to site-specific predictions for individual oxygen atoms of a molecule in comparison with commonly used graph convolutional networks that focus on a singular molecular property prediction. We demonstrate that our approach is effective in learning the location and magnitudes of protonation energies of oxygenated molecules.