WorldQuant Hedge Fund Deploys Scientists for Cancer Research

• Igor Tulchinsky has teamed up with Weill Cornell Medicine
• Hedge fund data scientists serve as fellows in genomics lab

Some 1,000 conference attendees took part this week in this research project, led by Chris Mason, a metagenomicist at Weill Cornell Medicine. The microbial data gathered will help create the first global map to track hot spots of resistance to antibiotics, a growing threat that will likely kill more people than all cancers by 2050, according to Centers for Disease Control data presented at the conference.

Igor Tulchinsky, founder of the $5 billion hedge fund WorldQuant, is teaming up with Mason to accelerate research in the burgeoning fields of computational biomedicine and genomics. Researchers, armed with troves of new genetic data and the wizardry of artificial intelligence, are aiming for breakthroughs in everything from predicting the risk of cancer and treating it to forecasting the spread of pathogens around the globe.

The collaboration goes beyond the $5 million that Tulchinsky gave last year to support the research. WorldQuant has sent three data scientists to work in Mason’s lab, applying their algorithms, typically homed in on chasing profits, to metagenomic research.

“What intrigues me about data prediction is how it is changing the world,” Tulchinsky said in an email interview. “For predictive medicine, it is saving more lives. If I can make a difference in someone’s life, I want to do that. Applying our expertise to a biomedical research partnership is an extremely compelling opportunity.”

The results from Milken are already in: Mason’s team found 2,150 microbial species — bacteria, viruses and fungi — and 560 antibiotic resistance markers at the Milken Institute Global Conference in Beverly Hills. The data is part of a 3-year-old project that started with swabbing things people touch in New York City subways.

Researchers have gathered and analyzed the genetic makeup of thousands of germs across 55 cities on every continent. The mapping project, set to be completed in 2020, could inform health officials on everything from possible threats of pandemics to predicting new drugs to address the overuse of antibiotics.

“Our goal is to be able to predict where antibiotic resistance is moving and where it presents the most existential threat,” Mason said in an interview.

Machine Learning

Tulchinsky, a computer scientist who founded New York-based WorldQuant in 2007, has been at the forefront of quantitative investing. The firm, which manages money for Millennium Management, has PhDs and other researchers in 15 countries using machine learning to comb through data for trading signals.

After a decade of running WorldQuant, Tulchinsky, 51, has sought to expand his reach beyond finance. He found an eager partner in Mason, a Yale-trained PhD with big ideas of his own. Mason, 39, has spoken publicly about the necessity and moral obligation for humans to colonize Mars in the coming centuries because the sun will eventually engulf the Earth, destroying all species. The two scientists became fast friends after meeting over lunch
a few years ago.

Tulchinsky is helping Mason’s lab on several fronts. Today, practitioners often struggle to accurately diagnose the cause of an infection by trying to culture it. So Mason’s lab seeks to genetically sequence and mine all DNA and RNA known to humanity, which would help diagnose infections in the future.

“Researchers will be able to determine the type of infection in a matter of seconds, rather than having to wait days trying to culture and identify it,” Mason said.

‘Rocket’ Fuel

The WorldQuant scientists, who serve fellowships in Mason’s lab, have mined sequencing data to find new drugs. One fellow contributed to a paper that Mason expects will be published in a peer-review journal. He says the fellows possess a fearlessness, and as outsiders, they bring fresh perspectives and see new research approaches.

“When WorldQuant steps into one of our projects, it’s like adding rocket booster fuel,” Mason said.

What does WorldQuant get out of the deal? The fellows may learn new predictive methods at the lab to bring back to the hedge fund in a cross-fertilization between finance and medicine.

“Top talent is seeking employment that engages them beyond core expertise,” Tulchinsky said. “Participation in our program has tremendous interest.”

https://bloom.bg/2KzPETU

CMS innovation unit ‘hobbled,’ data collection possibly aimless

The CMS Innovation Center is being crippled by the Trump administration and might not be collecting the correct data from new payment models, according to a group of researchers.

The focus on voluntary participation in programs by the Center for Medicare & Medicaid Innovation Center (CMMI) is hurting payment reform by possibly providing an inaccurate depiction of how such models will perform nationwide, researchers including Rahul Rajkumar, M.D., chief medical officer at Blue Cross Blue Shield of North Carolina, University of Michigan law professor Nicholas Bagley and Yale professor Scott Levy wrote in the New England Journal of Medicine.

“Voluntary programs don’t always provide insight into whether a payment approach ought to be rolled out on a nationwide basis,” they said.

While payment reform, like value-based care and population health, is popular throughout the healthcare sector, providers have been slow to adopt to changes and have pushed back on aggressive new payment changes. Those providers have found refuge in the Trump administration’s approach, which has watered down or completely ended some mandatory programs.

The researchers specifically called out the administration’s “backpedaling” of mandatory programs, like bundled payments for hip and femur fractures last year. Following the rollback, the agency, as expected, released a voluntary replacement bundle.

Last fall, CMS Administrator Seema Verma called for industry input to guide the innovation center in a “new direction.”

The CMMI was initiated under the Affordable Care Act as a way to test payment models that have the potential to lower cost and improve patient outcomes. The researchers said the center is not living up to expectations.

Congress “would not have delegated to CMMI the extraordinary power to reshape Medicare and Medicaid while prohibiting the agency from amassing the highest-quality evidence about which models are effective,” the authors wrote.

The Department of Health and Human Services (HHS) has given mixed signals over the past year on how it plans to proceed with new CMMI models. Tom Price, President Donald Trump’s original, and short-lived, HHS secretary strongly opposed mandatory models, while current secretary, Alex Azar, voiced more support.

Last month, Azar appointed former Landmark Health CEO Adam Boehler to lead the Innovation Center.

CMS administrator Seema Verma has signaled that a larger quantity of voluntary programs is the best way to control costs by giving providers more choices, but one healthcare expert said that will only make things worse.

“The CMS has urged on the side of having too many programs and models, and its a little bit overwhelming for many organizations to try and keep up with and for the agency to manage,” Michael Abrams, co-founder and managing partner of Numerof & Associates, told FierceHealthcare.

Abrams also agreed with the authors that voluntary models are not always reflective of how the entire industry will react to a new program. He added that fewer payment models provide a better focus on data, which would lead to better research on what programs work.

https://bit.ly/2JS19F8

AI researchers allege that machine learning is alchemy

Ali Rahimi, a researcher in artificial intelligence (AI) at Google in San Francisco, California, took a swipe at his field last December—and received a 40-second ovation for it. Speaking at an AI conference, Rahimi charged that machine learning algorithms, in which computers learn through trial and error, have become a form of “alchemy.” Researchers, he said, do not know why some algorithms work and others don’t, nor do they have rigorous criteria for choosing one AI architecture over another. Now, in a paper presented on 30 April at the International Conference on Learning Representations in Vancouver, Canada, Rahimi and his collaborators document examplesof what they see as the alchemy problem and offer prescriptions for bolstering AI’s rigor.

“There’s an anguish in the field,” Rahimi says. “Many of us feel like we’re operating on an alien technology.”

The issue is distinct from AI’s reproducibility problem, in which researchers can’t replicate each other’s results because of inconsistent experimental and publication practices. It also differs from the “black box” or “interpretability” problem in machine learning: the difficulty of explaining how a particular AI has come to its conclusions. As Rahimi puts it, “I’m trying to draw a distinction between a machine learning system that’s a black box and an entire field that’s become a black box.”

Without deep understanding of the basic tools needed to build and train new algorithms, he says, researchers creating AIs resort to hearsay, like medieval alchemists. “People gravitate around cargo-cult practices,” relying on “folklore and magic spells,” adds François Chollet, a computer scientist at Google in Mountain View, California. For example, he says, they adopt pet methods to tune their AIs’ “learning rates”—how much an algorithm corrects itself after each mistake—without understanding why one is better than others. In other cases, AI researchers training their algorithms are simply stumbling in the dark. For example, they implement what’s called “stochastic gradient descent” in order to optimize an algorithm’s parameters for the lowest possible failure rate. Yet despite thousands of academic papers on the subject, and countless ways of applying the method, the process still relies on trial and error.

Rahimi’s paper highlights the wasted effort and suboptimal performance that can result. For example, it notes that when other researchers stripped most of the complexity from a state-of-the-art language translation algorithm, it actually translated from English to German or French better and more efficiently, showing that its creators didn’t fully grasp what those extra parts were good for. Conversely, sometimes the bells and whistles tacked onto an algorithm are the only good parts, says Ferenc Huszár, a machine learning researcher at Twitter in London. In some cases, he says, the core of an algorithm is technically flawed, implying that its good results are “attributable entirely to other tricks applied on top.”

Rahimi offers several suggestions for learning which algorithms work best, and when. For starters, he says, researchers should conduct “ablation studies” like those done with the translation algorithm: deleting parts of an algorithm one at a time to see the function of each component. He calls for “sliced analysis,” in which an algorithm’s performance is analyzed in detail to see how improvement in some areas might have a cost elsewhere. And he says researchers should test their algorithms with many different conditions and settings, and should report performances for all of them.

Ben Recht, a computer scientist at the University of California, Berkeley, and coauthor of Rahimi’s alchemy keynote talk, says AI needs to borrow from physics, where researchers often shrink a problem down to a smaller “toy problem.” “Physicists are amazing at devising simple experiments to root out explanations for phenomena,” he says. Some AI researchers are already taking that approach, testing image recognition algorithms on small black-and-white handwritten characters before tackling large color photos, to better understand the algorithms’ inner mechanics.

Csaba Szepesvári, a computer scientist at DeepMind in London, says the field also needs to reduce its emphasis on competitive testing. At present, a paper is more likely to be published if the reported algorithm beats some benchmark than if the paper sheds light on the software’s inner workings, he says. That’s how the fancy translation algorithm made it through peer review. “The purpose of science is to generate knowledge,” he says. “You want to produce something that other people can take and build on.”

Not everyone agrees with Rahimi and Recht’s critique. Yann LeCun, Facebook’s chief AI scientist in New York City, worries that shifting too much effort away from bleeding-edge techniques toward core understanding could slow innovation and discourage AI’s real-world adoption. “It’s not alchemy, it’s engineering,” he says. “Engineering is messy.”

Recht sees a place for methodical and adventurous research alike. “We need both,” he says. “We need to understand where failure points come so that we can build reliable systems, and we have to push the frontiers so that we can have even more impressive systems down the line.”

https://bit.ly/2HOEU6o

Businesses drop marijuana from drug test requirements

FPI Management, a property company in California, wants to hire dozens of people. Factories from New Hampshire to Michigan need workers. Hotels in Las Vegas are desperate to fill jobs.

Those employers and many others are quietly taking what once would have been a radical step: They’re dropping marijuana from the drug tests they require of prospective employees.

Marijuana testing — a fixture at large American employers for at least 30 years — excludes too many potential workers, experts say, at a time when filling jobs is more challenging.

“It has come out of nowhere,” said Michael Clarkson, head of the drug testing practice at Ogletree Deakins, a law firm. “I have heard from lots of clients things like, ‘I can’t staff the third shift and test for marijuana.'”

https://bit.ly/2wf9d0V

China Biologic says ‘will be challenging to meet’ previous FY18 guidance

For the full year of 2018, the Company previously published a guidance of total sales growth of 18% to 20% in RMB terms and non-GAAP adjusted net income growth of 16% to 18% in RMB terms over 2017 financial results. Excluding TianXinFu, sales for 2018 are expected to grow 6% to 8% in RMB terms, and non-GAAP adjusted net income is expected to grow 3% to 4% in RMB terms over 2017 financial results. The 2018 non-GAAP adjusted net income projection excludes non-cash employee share-based compensation expenses and non-cash intangible assets amortization expense associated with the TianXinFu acquisition. However, given the worse-than-expected first quarter results due to ongoing impact of regulatory changes, the Company expects that it will be challenging to meet its previously published full year guidance. The Company is actively evaluating the evolving regulatory environment and competition dynamics and may lower its full year guidance should there be no significant improvement in the business operating conditions for the remainder of the year. FY18 EPS/revenue consensus is $5.35 and $439.6M, respectively

https://bit.ly/2HSp1b9

Nokia to exit digital health market

• Nokia is washing its hands of digital health. The company announced plans Wednesday to sell its digital health business to Eric Carreel, co-founder and chairman of French startup Withings, which Nokia purchased in 2016.
• The deal, which is expected to close late in the second quarter, includes trackers, connected scales, blood pressure monitors and other digital health devices for consumers and enterprise customers.
• Nokia took a 141 million euro writedown on the Withings deal in October, according to Bloomberg. Nokia said the sale is part of a strategy to become a business-to-business and licensing company.

Nokia’s short foray into digital health is an example of how a company can misread the tea leaves when it expands into new business areas.

The Finnish tech firm acquired Withings, a consumer electronics company, for $191 million in April 2016. The move signaled a new focus following the sale of Nokia’s mobile phone business to Microsoft in 2014, and a bet that there was plenty of room for growth and investment in wearable fitness trackers. But that gamble turned south as Apple and other competitors added fitness tracking features to their smartwatches.

Nokia’s exit comes as more companies are eyeing a piece of the digital health pie. According to PwC’s Health Research Institute, 84% of Fortune 50 companies are invested in healthcare, up from 76% five years ago. In all, 967 deals were cut in the U.S. health services market last year and 2018 shows no signs of slowing down.

During the first quarter of this year, digital health funding hit a record high of $1.62 billion across 77 deals, up from $1.41 billion the previous year, according to a recent Rock Health report. The haul included three $100 million-plus deals, with an average deal size of $21 million.

The top-funded categories were digital diagnostic solutions, disease monitoring, diabetes, consumer health information, R&D catalysts, on-demand health services and health benefits administration.

Notably missing among current top value propositions were fitness trackers and wearables. That’s a message Nokia apparently learned too little and too late.

https://bit.ly/2jxbvzl

De Blasio backs plan to open supervised injection sites in NYC

Mayor de Blasio threw his support behind a plan to open spots in New York City where addicts can inject drugs under watchful eyes.

No city in the United States has opened the so-called supervised injection facilities, which are meant to prevent overdose deaths — but there are more than 100 around the world.

De Blasio backed the controversial idea as the Health Department released a long-awaited report on the concept — finding that the program could prevent 130 overdose deaths each year.

The administration said it decided to embrace the idea despite obstacles — especially federal law, which makes it a crime to run a location that facilitates drug use.

If the plan proceeds, sites would open as a one-year pilot program in up to four locations — Gowanus in Brooklyn, Midtown West and Washington Heights in Manhattan, and Longwood in the Bronx.

There are currently needle exchanges at each proposed site. The injection facilities would be run by the nonprofit Research for a Safer New York, and would not get city money.

“The opioid epidemic has killed more people in our city than car crashes and homicides combined,” de Blasio said.

“After a rigorous review of similar efforts across the world, and after careful consideration of public health and safety expert views, we believe overdose prevention centers will save lives and get more New Yorkers into the treatment they need to beat this deadly addiction.”

The injection facilities would be run by the nonprofit Research for a Safer New York if the plan proceeds. (Marcus Santos)

In 2017, 1,441 people died of drug overdoses in New York City — driven by an epidemic of opioids like heroin and fentanyl. Fatalities have surged 166% since 2010.

In order to proceed, the city says it will need approval from the state Department of Health, an agreement by district attorneys to protect clients and operators from prosecution, and the support of the Council members representing the sites.

Deputy mayor Herminia Palacio wrote to state health commissioner Howard Zucker Thursday asking him to license a pilot research study to create the facilities, which the city is calling Overdose Prevention Centers.

If those parties sign off, the city will take another six months to a year to form a community advisory board before opening the injection sites.

Advocates have been pushing de Blasio to embrace the injection sites — including at a protest Wednesday where 11 people were arrested.

“More people are dying of overdose in New York City than ever before and there’s no sign of stopping anytime in the near future,” said Alyssa Aguilera, Co-Executive Director of the group VOCAL-NY. “The reality is that people use drugs, and forcing individuals to inject in public bathrooms and parks is unsafe and inhumane.”

While no injection facilities currently exist in the United States, three cities — Seattle, San Francisco, and Philadelphia — are moving toward opening them.

In addition to federal laws against drug possession, a law known as the “crack house statute” makes it a crime to run a location that facilitates drug use. Violators could be fined or jailed, and property used in such an operation could be seized.

But it would be up to the feds to decide whether to vigorously enforce the law, or look the other way as they largely have with states that legalize marijuana.

Ed Mullins, president of the Sergeants Benevolent Association, blasted the injection facilities plan.

“It’s total insanity,” he said. “Drugs are illegal. So we are now telling people it’s OK to break the law.”

The city has been studying the issue since the City Council put up $100,000 for the effort in 2016. “These sites will save lives and connect addicts with treatment options and trained professionals that could lead them to recovery,” said Council Speaker Corey Johnson.

A spokesman for the state health department said they’d review the city’s request.

Brooklyn district attorney Eric Gonzalez said officials can’t “sit by and let people die when there are proven interventions that can save lives,” and the Manhattan district attorney’s office also said they support the idea. Bronx DA Darcel Clark said she was “open minded” and would study the issue.

https://nydn.us/2FHyjVI