Wednesday, March 23, 2016

Writers Guild President Howard Rodman mourns the passing of Ken Howard

Los Angeles – Writers Guild of America, West President Howard A. Rodman released the following statement today on the passing of SAG-AFTRA President Ken Howard:

“It was with great sadness that I learned of Ken Howard’s passing today. He was a tireless advocate for social justice, for his union, and for its members. It is a fitting tribute to his commitment to improving the lives of actors, broadcasters and recording artists that he was the first president of the united SAG-AFTRA. He will be missed as a leader and as a bright light in the creative community. On behalf of the WGAW I would like to send our condolences to his family, his friends, and all those whose lives he touched throughout his long and exemplary career.”

Mobile connectivity at top of Wells Fargo’s ATM strategy; ATM AI remains opaque

Intrigued by the growing power and proficiency of the Wells Fargo ATMs I frequent (they are using machine learning and predictive analytics to offer me the options I’m most likely to want, given my interactive ATM history), I reached out to the bank’s Corporate Communications office with a list of ten questions about the AI in their ATMs.  I got a prompt response from corporate spokesperson Hilary O’Bryne, who e-mailed me, saying:

“Thank you for reaching out! While we can’t comment on your questions below,  there are actually a couple of different cardless ATM features we’re planning to roll out in the next year or so that you’ll likely be interested in. First, later this spring, Wells Fargo customers will be able to use NFC [Near Field Communication] technology to conduct ATM transactions without using a debit or ATM card, instead using a mobile wallet. A customer initiates an ATM transaction by authenticating through a thumbprint or passcode on their smartphone, and holding the phone near an NFC-enabled ATM terminal. Once authenticated, the ATM will proceed to the main menu, allowing the customer to input his/her PIN and proceed to complete a transaction. Right now the wallet that we support is Android Pay, and we’ll continue to evaluate additional wallets.

”Additionally, in late 2016, Wells Fargo will pilot a feature that allows customers to withdraw cash from an ATM using a mobile app instead of an ATM card. First, customers log into the Wells Fargo mobile app, select an account and request an access code. When the customer arrives at the ATM, he or she simply types in the 8-digit code and their ATM pin, and proceeds to the main menu to perform a transaction.

“These new functionalities give customers another option to access their accounts in ways that are most convenient to them, and they exemplify Wells Fargo’s commitment to innovation that helps our customers succeed financially. Wells Fargo customers use all of our channels – everything from mobile to phone to stores to ATMs – to manage their finances. But mobile continues to be critically important, and it’s become the primary way in which millions of our customers interact with us daily and 20 or more times per month. We’re excited to provide our customers with another way to conveniently and securely make ATM transactions how and where they are – and for many that’s on a mobile device.”

Here are the unanswered questions I posed to the bank:

1.    Does the software in the ATM network that learns about a user’s habits and offers options based on their interactive history with it have a name?
2.    How long has it been in operation?
3.    Was it developed in-house by Wells Fargo or was it developed by an outside vendor?
4.    What unique deep learning and predictive analytics algorithms does it employ to reach its conclusions about what a user may want next when they access an ATM?
5.    How much time and money does the ATM’s AI save for Wells Fargo and its customers?
6.    Does Wells Fargo contemplate licensing the underlying software capabilities of its ATM systems to third parties for further modification, extension and use in other venues?
7.    Are the predictive algorithms being used to present useful options to ATM users closely related to other algorithms used by Wells Fargo to predict if a given transaction may be fraudulent?
8.    How do Wells Fargo’s predictive algorithms for ATMs compare with the systems in use by other banks?
9.    How have these enhanced capabilities been received by Wells Fargo’s ATM users? 
10.  Do you consider your ATM’s predictive capabilities to be a factor in enhancing user loyalty and retention rates?

Wednesday, March 9, 2016

Software-savvy Ron Unz seems to want Harvard to go open source with admission algorithms

It would seem that analytics pioneer and long-time political activist Ron Unz wants his alma mater (he graduated from Harvard College in 1983) to pull open its corporate kimono and show the world the algorithms it uses to select the members of its elite student body.

Unz believes that Harvard University is a hedge fund disguised as an educational charity.

On the website of “Free Harvard/Fair Harvard” he argues that:

Harvard is now one of the world’s largest hedge-funds, with its $38 BILLION portfolio tax-exempt because of the college it runs as a charity off to one side.

The university’s annual investment income is twenty-five times larger than its net tuition revenue.

Meanwhile, thousands of student families are forced to spend most of their life-savings on $180,000 of total tuition, while relatively few non-affluent students even bother applying.

Paying tuition to a giant hedge-fund is unconscionable, and Harvard should immediately abolish all college tuition.

All the reforms he is calling for in his campaign to elect a Quintumvirate of insurgents to the Harvard Board of Overseers, can, viewed properly, be seen as  efforts to reprogram certain University computer systems.  Mr. Unz founded Wall Street Analytics, Inc., which has morphed into Moody’s Analytics.  Accordingly, we can view his efforts to re-program the allocation algorithms at the $38 billion endowment to cover all undergraduate tuition as an extension of his previous efforts to apply computing to solve financial problems. 

Similarly, his call for greater transparency in the undergraduate admissions process can be seen as a call to make public (open source) the code that rates and ranks applicants and optimizes for certain variables in selecting the overall class.  This could be preparatory to re-programming those algorithms to admit students and compose a class based on a revised set of criteria.

If Unz is right and making Harvard free would encourage millions more students to apply for admission, it’s certain that machine learning and predictive analytic systems would be doing most of the heavy lifting in the admissions process, and so could readily be tweaked to adjust the variables contributing to the analytic prediction that such-and-such a student is the most likely to flourish in and enhance the Harvard milieu and so deserves one of the University's Golden Tickets.

Read more about the up-coming Harvard Board of Overseers election here.

Free Harvard/Fair Harvard overseer candidate Stuart Taylor, Jr., submits questionnaire answers, questions commitment of its sponsors to open-mindedness

Lawyer and author Stuart Taylor, Jr., a member of the Quintumvirate of candidates seeking election this Spring to the Harvard Board of Overseers (is “overseer” an archaic echo of slavery?) under the banner “Free Harvard/Fair Harvard,” today submitted a detailed set of responses to a questionnaire on diversity and related-issues sent to the eight insider and five outsider candidates for seats on this second-most-powerful body in the governing apparatus of American’s oldest and, arguably, most prestigious and influential university.

He took issue with the attitude of one of the sponsors of the questionnaire, sent by a consortium of 10 groups, including the “Coalition for a Diverse Harvard,” which writes on its web site that:

“As Harvard alumni and students, we call on members of the Harvard community to join this Coalition against the petition slate — and in favor of race-conscious and holistic admissions practices that support campus diversity.

The “petition slate” refers to the fact that the insurgent Quintumvirate, which, in addition to Mr. Taylor includes Ralph Nader, Steven Hsu, Lee Cheng, and Ron Unz, the principal organizer of this effort, qualified for the Overseers ballot by submitting the required number of petition signatures from Harvard alumni by the February 1, 2016, deadline.  Mail-in ballots for the election will go out to 320,000 Harvard alumni by April 1st and need to be returned by May 20, 2016.

The Free Harvard/Fair Harvard slate is proposing the abolition of undergraduate tuition at Harvard College and more transparency in the admission process.  You can learn more about it here.

In the cover e-mail returning his answers to the questionnaire, Mr. Taylor writes:

I attach my responses to your questionnaire, which I spent several hours preparing out of respect for the integrity of the election process.

“The questionnaire offered this assurance: "Your answers to the following questions will help guide us in our recommendations to our members regarding the vote."

“I note that at least one of the ten organizations sponsoring the questionnaire (listed below) proceeded to publicize its opposition to my candidacy, and to that of my colleagues, without awaiting responses.

“This tells me all I that need to know about that organization's respect for the integrity of the election process. It also raises questions about the other nine allied organizations.”

Here is the complete questionnaire:

Dear Candidates,

Issues of diversity at Harvard and beyond have been pushed to the fore of the upcoming Overseers election. Our organizations care deeply about these matters. Your answers to the following questions will help guide us in our recommendations to our members regarding the vote. Please send your answers to [our e-mail address] by 5 p.m. on Thursday, March 10th. Responses may be edited for length when shared with our members. 

1. How important should student diversity be at Harvard? What strategies should the University pursue regarding this? 

2. Please state your views on affirmative action.

3. Please state your views on race-conscious college admissions (if not specified in your answer to question #2). 

4. Please state your views on whether Harvard should be more transparent about its college admissions process, particularly about how the mix of students is created. 

5. What steps have you taken to bring diversity to your workplace or to an organization that you have participated in? 

Thank you.

Here’s what Stuart Taylor had to say in response to these questions:

Harvard University
2016 Board of Overseers Election
Candidate Questionnaire:  Stuart Taylor, Jr., J.D. 1977

1. How important should student diversity be at Harvard? What strategies should the University pursue regarding this?

There are many diversities, and the most critical of them in my view are diversity of socioeconomic background, intellectual (including ideological) orientation, and racial (including ethnic) background.

Socioeconomic diversity. In this era of persistent and growing economic inequality, it is critical that Harvard to be a beacon of opportunity for promising people who were not born rich, privileged, or even upper-middle class. So socioeconomic diversity is critical. It also fosters racial diversity because the traditionally disadvantaged minorities are disproportionately of modest means. Indeed, affirmative action was originally -- and still should be -- designed to increase socioeconomic as well as racial diversity.

The promising children of schoolteachers, grocers, construction workers, laundry workers, factory workers, cops, firefighters, taxi drivers, janitors, and structurally unemployed parents deserve opportunities to improve their lives. And many could do well at Harvard. They also have much to teach their more prosperous classmates, and vice versa. In addition, experience shows that students of modest means may well gain ground on or even surpass wealthier classmates, from better schools, who enter college with somewhat higher test scores and high school grades.

Of course, it would a student of modest means -- or any student -- no favor for Harvard to to admit her, in pursuit of a numerical target, if she lacks the academic preparation to do well. This would and thus to set her up for academic struggle or failure by. And most entering students want to do well enough to pursue aspirations for careers that require challenging undergraduate courses and good grades. Harvard must strike a careful balance to extend opportunity broadly without using academically underprepared students as pawns, while misleading them about their academic prospects.

My impression is that Harvard has done far too little to attract and admit promising students of modest means. If elected, I will have access to much more information to test this impression. I will hope to be proved wrong. If proved right, I will do my utmost to change that.

Both the "Free Harvard" (abolishing tuition) and the "Fair Harvard" (admissions transparency) planks of our five-candidate slate would greatly foster socioeconomic diversity. Making Harvard tuition-free would attract countless well-qualified applicants from families of modest to average means who would now have great difficulty paying for a Harvard education, even with relatively generous financial aid. Free tuition would also attract countless more people, especially from small-town America, who now assume that Harvard is for the rich. Admissions transparency would also help bring in more students of modest and average means, for the reasons given in my response to question #4.

Intellectual diversity. It is vital for any university to welcome outstanding students and faculty with a wide range of intellectual orientations. Intellectual dialectic that challenges one's views is vital to clear thinking and progress. We all learn much from civil discourse with people whose intellectual orientations and opinions are unlike our own. Intellectual homogeneity makes potentially smart people stupid by fostering groupthink, herd behavior, and complacency. I rarely feel that I am learning much of value about important public issues unless I am in touch with, or at least reading, thoughtful conservatives, liberals, and moderates, ideally ranging from end to end of the ideological spectrum of opinion.

As a moderate -- with views to the left of most of the general public and to the right of most academics and journalists -- I have long believed that America's top universities, including Harvard, are sorely lacking in intellectual diversity and getting worse. I was especially shocked when coauthoring (with KC Johnson) my first book, Until Proven Innocent. It focused on (among other things) the unthinking, disgraceful, mob-like rush to judgment in 2006 and 2007 by more than activist Duke professors from the far Left -- and the university's leadership, which feared far left professors and their media allies -- against Duke lacrosse players who were falsely accused of gang rape. I was even more horrified that virtually all of the more fair-minded Duke professors were too intimidated by the far-left mob to speak up for due process and the presumption of innocence.

Worse, I have been told that it would be almost impossible today -- in part because of thinly veiled discrimination -- for an academic with conservative or even moderate ideological views to get tenure at Harvard in any field outside the STEM disciplines and perhaps economics. (I would be very glad to see this proved wrong.) Students would learn much more from an intellectually diverse faculty.

"Free Harvard" would indirectly increase intellectual diversity by attracting applicants from a broader range of backgrounds. "Fair Harvard" could indirectly increase intellectual diversity by exposing any ideology-based discrimination against applicants. It could also shed light on whether admissions preferences based on race, legacy status, or athletic talent foster undue ideological homogeneity. And although intellectual diversity is not part of the "Freer Harvard, Fair Harvard" platform, it will be a personal priority of mine if I am elected.

Racial diversity. This is important both to the (limited) extent that it still fosters socioeconomic diversity and because, ideally, Harvard should look more like America and the world. For this reason, I have long supported affirmative action in the original and most healthy sense: energetic outreach, recruitment, and talent-development efforts to extend opportunity to traditionally underserved groups and to break down historical hierarchies. I support such programs in walks of life ranging from the construction trades to Harvard admissions. Harvard should, to the extent possible, work directly with inner-city schools to improve their students' academic preparation and competitiveness.

2. Please state your views on affirmative action.

I incorporate by reference my support for the kinds of affirmative action discussed in the last paragraph of my response to question #1.

At the same time, as detailed below, I have grave concerns about the excesses associated with large racial preferences. By that I mean preferences that lead to large racial gaps in the entering academic credentials, and consequently in the academic performance, of students at Harvard or any other school. Such racial preferences should not be made a permanent feature of American life, a goal of many racial-preference supporters.

I also oppose have grave concerns about the use of "affirmative action" -- and of "diversity" -- as euphemisms for excessive use of large racial preferences and quotas. This is unfortunate and misleading because a great many Americans do not equate "affirmative action" with racial preferences. The evidence is that while polls have for decades consistently shown overwhelming public opposition to "racial preferences" and even to "consideration of race in admissions," the polls are much more closely divided on "affirmative action."

Describing racial preferences as "affirmative action" is also misleading as a historical matter. The Kennedy and Johnson executive orders that popularized the phrase did not endorse -- indeed, they forbade -- preferences for, as well as discrimination against, any racial group. 

Please see also my responses to questions #1 and #3.

3. Please state your views on race-conscious college admissions (if not specified in your answer to question #2).

I again incorporate by reference my support for the kinds of affirmative action discussed in the last paragraph of my response to question #1.

But excessively large racial admissions preferences for favored minorities have very large social costs. The most obvious are unfairness to well-qualified Asian American and white applicants, especially the former, and the resentments that unfairness inevitably creates.

Indeed, in most selective universities -- including Harvard, I suspect -- the academically strong child of two Asian American or white cab drivers is almost certain to be passed over to admit the academically less strong child of two wealthy black or Hispanic lawyers. Again, as I said in my response to question #1, I will hope to be proved wrong. And again, if proved right, I will do my utmost to change that.

Less well-known but perhaps even more important is the powerful evidence that large racial preferences have backfired both against their supposed beneficiaries and against socioeconomic diversity.

Many of the most severely damaged victims of large racial preferences are, ironically, black and Hispanic students who have been misled about their academic prospects and thereby set up for struggle or even failure. Much of this evidence is detailed in Mismatch, a 2012 book that Richard Sander and I coauthored. It cited copious empirical research, and a number of individual stories, showing that large admissions preferences often have dire effects on recipients' academic performance, self-confidence, and career options.

Many very good African American, Hispanic American, and Native American students who would do well at most prestigious colleges are not academically prepared to compete at Harvard against some of the world's most brilliant and best-prepared students. This "mismatch," and the low grades that it causes, often forces such students to abandon challenging courses, especially in the STEM disciplines, that are gateways to the professions that many students of all races want to pursue. Mismatch also fosters self-doubt, self-segregation, understandable (if incorrect) suspicions that discrimination must be at work, unhappiness, and a sense of betrayal among students who were promised that they would do well. It also reinforces racial stereotypes about intellectual capacity.

A second irony of large racial preferences is that they ultimately undermine the beneficial effects of racial diversity itself. They do this by causing preferred minorities to cluster in easily graded courses, to sit at separate tables in dining halls, to live and socialize primarily with other academically mismatched minorities, and to demand more "safe spaces" where they can minimize contact with members of other racial groups . For these reasons, studies have shown, students tend to become friends mainly with academic peers.

A third irony has been the drift away from the original affirmative-action goal of advancing poor and blue-collar students whose families have been held back in life by racial discrimination or other forces. In order to avoid even larger racial gaps in academic performance than now, the black and Hispanic that students whom Harvard chooses to admit come very disproportionately from families that are either wealthy or foreign-born or both. The reason is that these students tend to have stronger academic credentials than the disproportionately poor and badly educated descendants of slaves.

"Free Harvard" would improve these problems by increasing the pool of well-qualified, socioeconomically diverse minority applicants and therefore reducing the pressure to use large racial preferences to reach numeral targets. Meanwhile, "Fair Harvard" would expose the costs as well as the benefits of large racial preferences, as detailed in my response to question #4.

4. Please state your views on whether Harvard should be more transparent about its college admissions process, particularly about how the mix of students is created.

Harvard should become much more transparent about its college admissions process, and about how the mix of students is created. This might be the single most healthy policy change that Harvard could adopt, especially in light of the deep suspicions among Asian Americans (and others, including me) that Harvard may systematically discriminate against them.

If transparency shows this suspicion to be unfounded, it will be good news for everyone. If transparency shows it to be true, it should lead immediately to ending an indefensible violation of Harvard's professed principles. Sunlight, in the wise words of Justice Brandeis, is the best of disinfectants.

Perhaps equally important, admissions transparency would serve a vital consumer-protection purpose for students from preferred racial minority groups, as well as the recruited athletes, children of large donors, and legacy applicants who also receive preferences. Harvard does not now give such preferred applicants an honest appraisal of what they are getting into. It does not tell them how their academic credentials compare with those of their median classmates. It does not tell them how well or poorly Harvard's own admissions people predict they will do, based on the academic indices -- a mix of high school grades and test scores -- that virtually all selective universities use to rank applicants by academic potential. It does not tell those who hope to take tough pre-med courses the success rates of past students with similar academic preparation.

Preference recipients, and everyone else, could figure out such questions for themselves if there were adequate transparency. That could mean public disclosure of all admissions policy documents and data that can be disclosed without violating the privacy of individual students. The data could include the average and median academic index scores all admitted students; of recruited athletes; of legacies; of those whose parents or grandparents have made large gifts or pledges during the three years preceding the admissions decisions; and of those in each identified racial group. The data could also include the average Harvard grades, the STEM-discipline retention rates, the most popular majors, and the graduation rates of students whose admissions-office academic index scores were in the 10th, 20th, 30th, and so on up to 90th percentiles.

Admitted applicants should privately be told their own academic index scores so that they can check how well past students with similar scores have done, on average. This would also allow those with relatively low academic index scores to focus on whether, in the long rung, they would be better off at Harvard, perhaps ending up near the bottom of the class, or at another, prestigious but less competitive school, where they could be academic stars.

Such transparency would also confirm or dispel myths about some closely guarded secrets that should not be secret: the extent and size of all forms of preferences -- athletic, legacy, and large-donor as well as racial -- and the related gaps in academic achievement. Transparency would thus enable informed debate over proposed policy changes.

Some people fear that transparency would create pressure to reduce somewhat the size of the racial gaps in the academic preparation of admitted students. They may prove to be wrong, at least at Harvard, because the racial gaps behind the current veil of secrecy may prove to be smaller than those at any other selective college. That is the inference to be drawn both from the best available data (which are dated and sketchy) and from Harvard's well-known ability to attract a lion's share of the top students in all racial groups. This will be all the more true when Harvard becomes both free and fair.

5. What steps have you taken to bring diversity to your workplace or to an organization that you have participated in?

As a career journalist and author, I have never been involved in hiring decisions and thus have never been in a position to bring diversity to my workplaces. In addition, all of my employers have been so sensitive to the benefits of diversity that they appeared to be taking already whatever steps I might recommend. I am not (yet) an active enough participant in any organization to make policy proposals. If that changes -- and certainly if I am elected to be a Harvard overseer -- I will make special efforts to push for as much socioeconomic, intellectual, and racial diversity as can reasonably be achieved.