Thursday, May 30, 2019

Manifold Episode #11: Joe Cesario on Police Decision Making and Racial Bias in Deadly Force Decisions



Manifold Show Page    YouTube Channel

Corey and Steve talk with Joe Cesario about his recent work which argues that, contrary to activist claims and media reports, there is no widespread racial bias in police shootings. Joe discusses his analysis of national criminal justice data and his experimental studies with police officers in a specially designed realistic simulator. He maintains that racial bias does exist in other uses of force such as tasering but that the decision to shoot is fundamentally different: it is driven by specific events and context, rather than race.

Cesario is associate professor of Psychology at Michigan State University. He studies social cognition and decision-making. His recent topics of study include police use of deadly force and computational modeling of fast decisions. Cesario is dedicated to reform in the practice, reporting, and publication of psychological science.

Is There Evidence of Racial Disparity in Police Use of Deadly Force? Analyses of Officer-Involved Fatal Shootings in 2015–2016
https://journals.sagepub.com/doi/abs/...

Example of officer completing shooting simulator
https://youtu.be/Le8zoqk-UVo

Overview of Current Research on Officer-Involved Shootings
https://www.cesariolab.com/police

Joseph Cesario Lab
https://www.cesariolab.com/


man·i·fold /ˈmanəˌfōld/ many and various.

In mathematics, a manifold is a topological space that locally resembles Euclidean space near each point.

Steve Hsu and Corey Washington have been friends for almost 30 years, and between them hold PhDs in Neuroscience, Philosophy, and Theoretical Physics. Join them for wide ranging and unfiltered conversations with leading writers, scientists, technologists, academics, entrepreneurs, investors, and more.

Steve Hsu is VP for Research and Professor of Theoretical Physics at Michigan State University. He is also a researcher in computational genomics and founder of several Silicon Valley startups, ranging from information security to biotech. Educated at Caltech and Berkeley, he was a Harvard Junior Fellow and held faculty positions at Yale and the University of Oregon before joining MSU.

Corey Washington is Director of Analytics in the Office of Research and Innovation at Michigan State University. He was educated at Amherst College and MIT before receiving a PhD in Philosophy from Stanford and a PhD in a Neuroscience from Columbia. He held faculty positions at the University Washington and the University of Maryland. Prior to MSU, Corey worked as a biotech consultant and is founder of a medical diagnostics startup.

Tuesday, May 28, 2019

NYTimes Op-Ed from the future (Ted Chiang): Genetics and Cognitive Enhancement

In this scenario Ted Chiang forecasts that recipients of government-funded genetic enhancement will not catch up to children of elites who receive similar enhancements. The latter are born to rich, highly educated parents and have access to elite social networks, better schools, etc. The system is still not entirely fair (i.e., invariant to accidents of birth), because many non-genetic advantages still exist. But can we ever achieve equality of outcome? At what cost?

Nevertheless, perhaps the beneficiaries of the Gene Equality Project are at least better off than their siblings who were not in the program?

It is interesting that the Times is already flirting with the idea of redistribution of genetic endowments. See also The Neanderthal Problem.
NYTIMES OP-ED FROM THE FUTURE

It’s 2059, and the Rich Kids Are Still Winning
DNA tweaks won’t fix our problems.

Ted Chiang is an award-winning science fiction writer.

Editors’ note: This is the first installment in a new series, “Op-Eds From the Future,” in which science fiction authors, futurists, philosophers and scientists write op-eds that they imagine we might read 10, 20 or even 100 years in the future.

Last week, The Times published an article about the long-term results of the Gene Equality Project, the philanthropic effort to bring genetic cognitive enhancements to low-income communities. The results were largely disappointing: While most of the children born of the project have now graduated from a four-year college, few attended elite universities and even fewer have found jobs with good salaries or opportunities for advancement. With the results in hand, it is time for us to re-examine the efficacy and desirability of genetic engineering.

The intentions behind the Gene Equality Project were good. Therapeutic genetic interventions, such as correcting the genes that cause cystic fibrosis and Huntington’s disease, have been covered by Medicare ever since their approval by the Food and Drug Administration, making them available to the children of low-income parents. However, augmentations like cognitive enhancements have never been covered — not even by private insurance — and were available only to affluent parents. Amid fears that we were witnessing the creation of a caste system based on genetic differences, the Gene Equality Project was begun 25 years ago, enabling 500 pairs of low-income parents to increase the intelligence of their children.

The project offered a common cognitive-enhancement protocol involving modifications to 80 genes associated with intelligence. Each individual modification had only a small effect on intelligence, but in combination they typically gave a child an I.Q. of 130, putting the child in the top 5 percent of the population. This protocol has become one of the most popular enhancements purchased by affluent parents, and it is often referenced in media profiles of the “New Elite,” the genetically engineered young people who are increasingly prevalent in management positions of corporate America today. Yet the 500 subjects of the Gene Equality Project are not enjoying career success that is remotely comparable to the success of the New Elite, despite having received the same protocol.

A range of explanations has been offered for the project’s results. White supremacist groups have claimed that its failure shows that certain races are incapable of being improved, given that many — although by no means all — of the beneficiaries of the project were people of color. Conspiracy theorists have accused the participating geneticists of malfeasance, claiming that they pursued a secret agenda to withhold genetic enhancements from the lower classes. But these explanations are unnecessary when one realizes the fundamental mistake underlying the Gene Equality Project: Cognitive enhancements are useful only when you live in a society that rewards ability, and the United States isn’t one.

It has long been known that a person’s ZIP code is an excellent predictor of lifetime income, educational success and health. Yet we continue to ignore this because it runs counter to one of the founding myths of this nation: that anyone who is smart and hardworking can get ahead. Our lack of hereditary titles has made it easy for people to dismiss the importance of family wealth and claim that everyone who is successful has earned it. The fact that affluent parents believe that genetic enhancements will improve their children’s prospects is a sign of this: They believe that ability will lead to success because they assume that their own success was a result of their ability.

For those who assume that the New Elite are ascending the corporate ladder purely on the basis of merit, consider that many of them are in leadership positions, but I.Q. has historically had only a weak correlation with effectiveness as a leader. Also consider that genetic height enhancement is frequently purchased by affluent parents, and the tendency to view taller individuals as more capable leaders is well documented. In a society increasingly obsessed with credentials, being genetically engineered is like having an Ivy-League M.B.A.: It is a marker of status that makes a candidate a safe bet for hiring, rather than an indicator of actual competence.

This is not to say that the genes associated with intelligence play no role in creating successful individuals — they absolutely do. They are an essential part of a positive feedback loop: When children demonstrate an aptitude at any activity, we reward them with more resources — equipment, private tutors, encouragement — to develop that aptitude; their genes enable them to translate those resources into improved performance, which we reward with even better resources, and the cycle continues until as adults they achieve exceptional career success. But low-income families living in neighborhoods with underfunded public schools often cannot sustain this feedback loop; the Gene Equality Project didn’t offer any resources besides better genes, and without these additional resources, the full potential of those genes was never realized.

We are indeed witnessing the creation of a caste system, not one based on biological differences in ability, but one that uses biology as a justification to solidify existing class distinctions. It is imperative that we put an end to this, but doing so will take more than free genetic enhancements supplied by a philanthropic foundation. It will require us to address structural inequalities in every aspect of our society, from housing to education to jobs. We won’t solve this by trying to improve people; we’ll only solve it by trying to improve the way we treat people.

This doesn’t necessarily mean that the Gene Equality Project is something that never needs to be repeated. Instead of thinking of it as a cure to an illness, we could think of it as a diagnostic test — something we would conduct at regular intervals to gauge how close we are to reaching our goal. When the beneficiaries of free genetic cognitive enhancements become as successful as the ones whose parents bought the enhancements for them, only then will we have reason to believe that we live in an equitable society.

Finally, let’s recall one of the arguments made during the original debate about legalizing genetic cognitive enhancements. Some proponents claimed that we had an ethical obligation to pursue cognitive enhancements because of the benefits to humanity that would accrue as a result. But there have surely been many geniuses whose world-changing contributions were lost because their potential was crushed by their impoverished surroundings.

Our goal should be to ensure that every individual has the opportunity to reach his or her full potential, no matter the circumstances of birth. That course of action would be just as beneficial to humanity as pursuing genetic cognitive enhancements, and it would do a much better job of fulfilling our ethical obligations.
This is one of the Reader Picks comments:
Mark
Philadelphia May 27

I have mixed feelings about the concept of this article. Surely, private schools confer numerous advantages to their students, who are from wealthy backgrounds and connections to higher education and corporate America.

But, look at Stuyvesant. The super intelligent and successful students are very often from middle class, lower-middle class, and even poor backgrounds. They are often first generation immigrants. They are just smart and hard working and their families care desperately about education.

Some kids are just smart, while others, are just average, or below average. You really think if you went into a school in the South Bronx and donated $1 billion the students would start cranking out perfect SATs?

Ask Zuckerberg how is $50 million donation to Newark public schools went. Darwinism is cruel, but some people aren't just cut out to be good students or white collar professionals.

Much of this has little to do with class and everything to do with drive and innate ability.

Saturday, May 25, 2019

Polygenic Risk Scores

I've collected some recent links related to polygenic risk scores below.

Many experts anticipate large scale clinical use of these scores within the next few years. Research progress has been very rapid -- it will be interesting to see how long it takes for these breakthroughs to be applied in health care. Graph below shows number of papers per year.

See also Harvard Business Review: AI and the Genetic Revolution (podcast).


Guardian article on UK Health Minister's proposal for widespread genetic testing:
"The latest predictive tests for a range of common diseases take a different approach: they aggregate the tiny contributions to risk made by hundreds or even thousands of genes to give a personalised score. Because the risk is spread out over many genes, people can end up at the very high-risk end of the spectrum by chance, without having a family history of a particular illness."

"Prof John Bell, a professor of medicine at Oxford university who led a recent government-commissioned review of the life sciences industry, said the approach could have a “quite profound” effect on the ability to manage disease. ... David Spiegelhalter, professor for the public understanding of risk at the University of Cambridge, agrees that genetic tests could allow the NHS to rapidly identify those who may need closer monitoring."
Editorial in New England Journal of Medicine anticipates broad clinical use of polygenic scores:
Has the Genome Granted Our Wish Yet?

"It is likely that tailoring decisions about prescribing preventive medicines or screening practices will be the main future use of genetic risk scores. If a PRS adds to existing clinical predictors of risk such as the Framingham Risk Score or the Q index for heart disease, it could be incorporated into preventive care as readily as any other biomarker."

"There seems little doubt that interpretation of these scores will become an accepted part of clinical practice in the future..."
Article in The Conversation by two Australian professors of Public Health (discussion refers to both monogenic and polygenic risks, AFAICT):
Population DNA testing for disease risk is coming. Here are five things to know

"As DNA testing becomes cheaper, it becomes more feasible to screen large numbers of healthy people for their risk of disease."

"We modelled the health and economic benefits of offering population DNA screening in Australia, focusing on young adults aged 18-25 years (about 2.6 million Australians). ... At A$200 per test (which could be realistic in the near future), savings in treatment costs could outweigh screening costs, saving the health-care system money and saving lives."
MyHeritage DNA (T2D, Heart Disease and Breast Cancer) joins 23andMe (T2D) in offering polygenic risk scores using common SNPs in their health reports. FDA regulatory stance allows DTC (Direct To Consumer) reports of this type as long as they are provided as information to be discussed with a physician, and not to diagnose a condition or prescribe care.

Someone tweeted me this photo from a recent conference presentation: increase of AUC (predictive power) with sample size. ~100k cases enough to capture most of common SNP heritability for diseases such as Testicular or Breast Cancer?



Figure below from our paper Genomic Prediction of Complex Disease Risk (bioRxiv).

Wednesday, May 22, 2019

Tomaso Poggio on AI, Neuroscience, and Physics



Highly recommended interview with MIT professor Tomaso Poggio, which I listened to recently on a plane. IIRC, I largely agreed with his positions except that I'm a bit more optimistic about AGI. I think his estimate for AGI was 100 or 200 years from now, whereas I think by the end of my lifetime is a distinct possibility.

Poggio (trained in theoretical physics) starts by describing the effect that Special Relativity had on him as a kid. It is a striking realization that pure thought experiments of the kind originally formulated by Einstein can have such far-reaching implications. See Physics as a Strange Attractor:
I suspect that Special Relativity, because it is easy to introduce (no mathematics beyond algebra is required), yet deep and beautiful and counterintuitive, stimulates many people of high ability to become interested in physics.
I notice (perhaps unsurprisingly) a lot of similarities in Poggio's views and those of his former student Demis Hassabis of DeepMind.
Tomaso Poggio is a professor at MIT and is the director of the Center for Brains, Minds, and Machines. Cited over 100,000 times, his work has had a profound impact on our understanding of the nature of intelligence, in both biological neural networks and artificial ones. He has been an advisor to many highly-impactful researchers and entrepreneurs in AI, including Demis Hassabis of DeepMind, Amnon Shashua of MobileEye, and Christof Koch of the Allen Institute for Brain Science. This conversation is part of the Artificial Intelligence podcast and the MIT course 6.S099: Artificial General Intelligence. The conversation and lectures are free and open to everyone. Audio podcast version is available on https://lexfridman.com/ai/

Thursday, May 16, 2019

Manifold Episode 10: Ron Unz on the Subprime Mortgage Crisis, The Unz Review, and the Harvard Admissions Scandal



Ron Unz is the publisher of the Unz Review, a controversial but widely read alternative media site hosting opinion outside of the mainstream, including from both the far right and the far left. Unz studied theoretical physics at Harvard, Cambridge and Stanford. He founded the software company Wall Street Analytics, acquired by Moody’s in 2006, and was behind the 1998 ballot initiative that ended bilingual education in California.

Podcast transcript

The Unz Review

The Myth of American Meritocracy - How corrupt are Ivy League admissions?

The Myth of American Meritocracy and Other Essays


man·i·fold /ˈmanəˌfōld/ many and various.

In mathematics, a manifold is a topological space that locally resembles Euclidean space near each point.

Steve Hsu and Corey Washington have been friends for almost 30 years, and between them hold PhDs in Neuroscience, Philosophy, and Theoretical Physics. Join them for wide ranging and unfiltered conversations with leading writers, scientists, technologists, academics, entrepreneurs, investors, and more.

Steve Hsu is VP for Research and Professor of Theoretical Physics at Michigan State University. He is also a researcher in computational genomics and founder of several Silicon Valley startups, ranging from information security to biotech. Educated at Caltech and Berkeley, he was a Harvard Junior Fellow and held faculty positions at Yale and the University of Oregon before joining MSU.

Corey Washington is Director of Analytics in the Office of Research and Innovation at Michigan State University. He was educated at Amherst College and MIT before receiving a PhD in Philosophy from Stanford and a PhD in a Neuroscience from Columbia. He held faculty positions at the University Washington and the University of Maryland. Prior to MSU, Corey worked as a biotech consultant and is founder of a medical diagnostics startup.

Wednesday, May 08, 2019

Harvard Business Review: AI and the Genetic Revolution (podcast)


Harvard Business Review podcast with Azeem Azhar (Exponential View).
AI and the Genetic Revolution

Michigan State University senior vice president Stephen Hsu, a theoretical physicist and the founder of Genomic Prediction, demonstrates how the machine learning revolution, combined with the dramatic fall in the cost of human genome sequencing, is driving a transformation in our relationship with our genes. Stephen and Azeem Azhar explore how the technology works, what predictions can and cannot yet be made (and why), and the ethical challenges created by this technology.

In this podcast, Azeem and Stephen also discuss:

FDA approval of the first genetic treatment for monogenic conditions and the work towards developing treatments for polygenic conditions like diabetes and cancer.

How this technology might exacerbate existing social inequalities or create new ones; is it just an issue of access, or does it go further?

Developing best practice protocols for governance and regulation of genomic technologies.
In the interview I mention that the number of genomics papers on polygenic risk scores has exploded just in the last year or so:

Tuesday, May 07, 2019

Embryo Screening: Polygenic Traits and Disease Risk

Several people asked me to comment on this paper, which appeared recently on biorxiv. It seems to be an update of earlier (simulation) analyses by Gwern [16] and Shulman and Bostrom [15] (cited in the paper) on potential gains from embryo selection using quantitative trait predictors (e.g., height, cognitive ability). In the paper the authors analyze real families using actual genetic and phenotype data.

The main limitations given current technology are the number of embryos available from which to select, and the accuracy of the polygenic predictors. The latter will almost certainly improve significantly for some traits in the near future, and for all traits eventually. The number of embryos available for selection may also increase if new methods allow oocytes (eggs) to be produced using stem cell technology (already demonstrated in mice; video).
Screening human embryos for polygenic traits has limited utility
E. Karavani et al.

Genome-wide association studies have led to the development of polygenic score (PS) predictors that explain increasing proportions of the variance in human complex traits. In parallel, progress in preimplantation genetic testing now allows genome-wide genotyping of embryos generated via in vitro fertilization (IVF). Jointly, these developments suggest the possibility of screening embryos for polygenic traits such as height or cognitive function. There are clear ethical, legal, and societal concerns regarding such a procedure, but these cannot be properly discussed in the absence of data on the expected outcomes of screening. Here, we use theory, simulations, and real data to evaluate the potential gain of PS-based embryo selection, defined as the expected difference in trait value between the top-scoring embryo and an average, unselected embryo. We observe that the gain increases very slowly with the number of embryos, but more rapidly with increased variance explained by the PS. Given currently available polygenic predictors and typical IVF yields, the average gain due to selection would be ≈2.5cm if selecting for height, and ≈2.5 IQ (intelligence quotient) points if selecting for cognitive function. These mean values are accompanied by wide confidence intervals; in real data drawn from nuclear families with up to 20 offspring each, we observe that the offspring with the highest PS for height was the tallest only in 25% of the families. We discuss prospects and limitations of PS-based embryo selection for the foreseeable future.
The authors of the paper seem to define "utility" in terms of expected gain in trait value. However, there is also utility in eliminating very negative outcomes, even if they have small probability. This does not shift the average very much but may still be highly desirable. For example, the odds of my house being destroyed by fire or earthquake in the next decade are small, but the outcome is negative enough that I will act to insure against it. If there is a 1% chance of a $100k house being destroyed, the expected loss is only $1k over the period. But without insurance the outcome might be devastating to a family.

One can compare this to screening for Down Syndrome, which has an incidence of roughly 1% (depending on parental age, etc.) but very serious consequences (see podcast discussion below).

At Genomic Prediction we have focused on screening against disease risk rather than on selection for quantitative traits, for both ethical and practical reasons. Even noisy (imperfect) predictors allow the identification of individuals who are high risk outliers -- e.g., are 5x times more likely to get the disease than a typical person.



When considering disease risk the key metric is not the polygenic score itself, because odds ratios are nonlinear functions of the score (or score percentile). For example (note, this is entirely hypothetical), consider 3 embryos with disease risk percentile scores (e.g., Breast Cancer, Type 1 Diabetes, Atrial Fibrillation, Coronary Artery Disease) given by column:

    #1    33   57   64   51

    #2    62   39   36   49

    #3    26   22   52   99.5

Even though the linear averages of the four risk percentiles for all three embryos are similar (contrived to be near 50), embryo #3 has unusually high risk for one condition (e.g., Coronary Artery Disease) and embryos #1 and #2 might be preferred.

Quantifying the utility to the family from this kind of screening is much more complex than for quantitative traits such as height or cognitive ability.

For more on ethical questions related to genetic engineering and embryo selection, see this podcast discussion with Sam Kerstein, chair of the philosophy department at the University of Maryland.

Friday, May 03, 2019

Janelia (HHMI) talk: Genomic Prediction of Complex Traits and Disease Risks via AI and Large Genomic Datasets



Janelia is the research campus of the Howard Hughes Medical Institute (HHMI), located near Washington DC. It is reputed to be heaven on earth for scientists :-)

I'll be visiting there next week (see title and abstract below). If you're at Janelia and want to meet with me there is still a little space on my schedule. Or just come to the talk and try to grab me afterward.

My talk is Tuesday May 7 12:30 – 1:30.
Genomic Prediction of Complex Traits and Disease Risks via AI and Large Genomic Datasets

Abstract: The talk is divided into two parts. The first gives an overview of the rapidly advancing area of genomic prediction of disease risks using polygenic scores. We can now identify risk outliers (e.g., with 5 or 10 times normal risk) for about 20 common disease conditions, ranging from diabetes to heart diseases to breast cancer, using inexpensive SNP genotypes (i.e., as offered by 23andMe). We can also predict some complex quantitative traits (e.g., adult height with accuracy of few cm, using ~20k SNPs). I discuss application of these results in precision medicine as well as embryo selection in IVF, and give some details concerning genetic architecture. The second part covers the AI/ML used to build these predictors, with an emphasis on "sparse learning" and phase transitions in high dimensional statistics.




Thursday, May 02, 2019

Manifold Episode #9: Philosopher S. Kerstein on the Morality of Genome Engineering



Corey and Steve speak with Samuel Kerstein, Professor of Philosophy and expert in Medical Ethics at the University of Maryland. They discuss the ethics of genome engineering and preimplantation embryo selection, and the inequality and narrowing of human diversity that might result from widespread adoption of these technologies. Among the topics covered: Why genome engineering at this time is immoral. Should we always pick the healthiest embryo? In the future will parents have a moral obligation to engineer their children? Will there be an arms race between countries to engineer their populations? Is Star Trek’s Khan a more advanced person (Steve) or just another smart psychopath (Sam) or both?

Samuel J. Kerstein

How to Treat Persons by Samuel J. Kerstein

CRISPR Babies – Episode #1

Transcript


man·i·fold /ˈmanəˌfōld/ many and various.

In mathematics, a manifold is a topological space that locally resembles Euclidean space near each point.

Steve Hsu and Corey Washington have been friends for almost 30 years, and between them hold PhDs in Neuroscience, Philosophy, and Theoretical Physics. Join them for wide ranging and unfiltered conversations with leading writers, scientists, technologists, academics, entrepreneurs, investors, and more.

Steve Hsu is VP for Research and Professor of Theoretical Physics at Michigan State University. He is also a researcher in computational genomics and founder of several Silicon Valley startups, ranging from information security to biotech. Educated at Caltech and Berkeley, he was a Harvard Junior Fellow and held faculty positions at Yale and the University of Oregon before joining MSU.

Corey Washington is Director of Analytics in the Office of Research and Innovation at Michigan State University. He was educated at Amherst College and MIT before receiving a PhD in Philosophy from Stanford and a PhD in a Neuroscience from Columbia. He held faculty positions at the University Washington and the University of Maryland. Prior to MSU, Corey worked as a biotech consultant and is founder of a medical diagnostics startup.

Blog Archive

Labels