Model Soups, extending Stochastic Weights Averaging (SWA), combine models fine-tuned with different hyperparameters. Yet, their adoption is hindered by computational challenges due to subset selection issues. In this paper, we propose to speed up model soups by approximating soups performance using averaged ensemble logits performances. Theoretical insights validate the congruence between ensemble logits and weight averaging soups across any mixing ratios. Our Resource ADjusted soups craftINg (RADIN) procedure stands out by allowing flexible evaluation budgets, enabling users to adjust his budget of exploration adapted to his resources while increasing performance at lower budget compared to previous greedy approach (up to 4% on ImageNet).

When you're a young male, you spend your money on cheap booze, hard drugs, and fast women – and the rest you waste. As you enter old age, you increase your spending on drugs, the type needed to treat the chronic problems you developed from abusing your body so much during your youth. The end result is more than $1.2 trillion spent last year on drugs produced by the pharmaceutical industry. Drugs are big business, but investing in biotech stocks is extremely risky due to the regulatory risks involved, the uncertainty around a drug's efficacy and safety, and the cost of taking a drug candidate to market which a pharmaceutical company incurs whether the drug gets approved or not. From an investor's perspective, it's best to find a "pick and shovel" business model somewhere in the pharma food chain that makes money regardless of whether or not drugs make it past human trials.

Transportation "can also become truly contactless if needed," says James Peng, who is CEO of ... [ ] self-driving startup Pony.ai. The first few months of 2020 have radically reshaped the way we work and how the world gets things done. While the wide use of robotaxis or self-driving freight trucks isn't yet in place, the Covid-19 pandemic has hurried the introduction of artificial intelligence across all industries. Whether through outbreak tracing or contactless customer pay interactions, the impact has been immediate, but it also provides a window into what's to come. The second annual Forbes' AI 50, which highlights the most promising U.S.-based artificial intelligence companies, features a group of founders who are already pondering what their space will look like in the future, though all agree that Covid-19 has permanently accelerated or altered the spread of AI. "We have seen two years of digital transformation in the course of the last two months," Abnormal Security CEO Evan Reiser told Forbes in May.

Transportation "can also become truly contactless if needed," says James Peng, who is CEO of ... [ ] self-driving startup Pony.ai. The first few months of 2020 have radically reshaped the way we work and how the world gets things done. While the wide use of robotaxis or self-driving freight trucks isn't yet in place, the Covid-19 pandemic has hurried the introduction of artificial intelligence across all industries. Whether through outbreak tracing or contactless customer pay interactions, the impact has been immediate, but it also provides a window into what's to come. The second annual Forbes' AI 50, which highlights the most promising U.S.-based artificial intelligence companies, features a group of founders who are already pondering what their space will look like in the future, though all agree that Covid-19 has permanently accelerated or altered the spread of AI. "We have seen two years of digital transformation in the course of the last two months," Abnormal Security CEO Evan Reiser told Forbes in May.

Washington – Can your Fitbit or Apple Watch detect a coronavirus infection before the onset of symptoms? Researchers are increasingly looking at these devices and other such wearables as a possible early-warning system for the deadly virus. Last month, scientists at the West Virginia University Rockefeller Neuroscience Institute said they had created a digital platform that can detect COVID-19 symptoms up to three days before they show up using the Oura ring, a wearable fitness and activity tracker. An app developed by the researchers uses artificial intelligence to forecast the onset of COVID-19 related symptoms such as fever, coughing, breathing difficulties and fatigue, with over 90 percent accuracy, according to the university. The researchers said the system could offer clues of infection in people not yet showing symptoms -- helping address one of the problems in detection and containment of the deadly outbreak.

Artificial intelligence company twoXAR--a specialist in separating signals from noise in data-rich drug discovery projects--has raised $10 million in a first-round financing led by SoftBank Ventures. Co-founder and CEO Andrew Radin says the proceeds will be used to bring forward preclinical drug candidates including candidates for diabetes, liver cancer and rheumatoid arthritis, build the pipeline with new partnering deals and spinouts, and add to its headcount. The company's business model is to use its AI technology to identify promising drug candidates, validate them through preclinical studies and then work with partners to bring them into the clinic. The Palo Alto, California-based company--which was the first investment for Silicon Valley entrepreneur Andreessen Horowitz's $200 million biotech fund in 2015--already has some collaborations under its belt. In 2016, it signed deals with the Universities of Chicago and Stanford for drug discovery programs in atherosclerosis and liver disease, respectively, as well as Mount Sinai Medical in the area of diabetic neuropathy.

Artificial intelligence (AI) is steadily reshaping healthcare from all sides, introducing technologies we wouldn't have thought possible five or 10 years ago. It's happening in the clinic (see HealthTap's Doctor A.I.), it's happening in diagnostics (see IBM Watson), and now it's moving into earlier-stage drug discovery with Palo Alto, California-based twoXAR. "In the couple years that we have been around, we've been told hundreds of times that computers cannot do this; that biology is too complex; that this will never work," said Andrew A. Radin, CEO of the AI-driven biopharmaceutical company. "Yet, in every single disease program where we have run proof-of-concept studies on our novel AI-identified candidates, we have generated efficacious results across standard end points." Using a custom-built computational platform, twoXAR works to identify what it calls "unanticipated associations between drug and disease."