Tag: Algorithms

Good

No Comments

Finally, a court has ruled that in order for a DNA test to be admitted as evidence, the source code must be made available to the defense.

To my mind, any software used for prosecutions should be publicly available for review:

A New Jersey appeals court has ruled that a man accused of murder is entitled to review proprietary genetic testing software to challenge evidence presented against him.

Attorneys defending Corey Pickett, on trial for a fatal Jersey City shooting that occurred in 2017, have been trying to examine the source code of a software program called TrueAllele to assess its reliability. The software helped analyze a genetic sample from a weapon that was used to tie the defendant to the crime.

The maker of the software, Cybergenetics, has insisted in lower court proceedings that the program’s source code is a trade secret. The co-founder of the company, Mark Perlin, is said to have argued against source code analysis by claiming that the program, consisting of 170,000 lines of MATLAB code, is so dense it would take eight and a half years to review at a rate of ten lines an hour.

MATLAB is a pretty high level language, so if you have 170,000 lines of code in it, you are writing bloated code. 

Also, if you have 170,000 lines of code in it, I guarantee that there are bugs, and likely substantial ones, because most of those lines of code are there to handle edge (unlikely) cases, where the programmer has to make broad assumptions about the data.

………

On Wednesday, the appellate court sided with the defense [PDF] and sent the case back to a lower court directing the judge to compel Cybergenetics to make the TrueAllele code available to the defense team.

“Without scrutinizing its software’s source code – a human-made set of instructions that may contain bugs, glitches, and defects – in the context of an adversarial system, no finding that it properly implements the underlying science could realistically be made,” the ruling says.

Kit Walsh, senior staff attorney for the Electronic Frontier Foundation, hailed the appellate ruling. “No one should be imprisoned or executed based on secret evidence that cannot be fairly evaluated for its reliability, and the ruling in this case will help prevent that injustice,” she said in a blog post. If TrueAllele is found wanting, presumably that will not affect the dozen individuals said to have been exonerated by the software.

It should be noted that the studies “validating” TrueAllele have been conducted by Mark Perlin, and as such are suspect.

Also, though these are slightly different application, we already know that algorithms used for health care software, Zoom face detection, educational evaluations, and criminal sentencing are all explicitly racist.

It is no stretch to assume that an algorithm explicitly developed for police and prosecutors would be biased in their favor.

That is how you make sales.

A Feature, Not a Bug

No Comments

Admission to university in the UK is driven by the “A-Level” exams, which have been canceled due to the Covid-19 pandemic.

The Boris Johnson government came up with an algorithm driven alternative, which turned out to be so biased against students from poorer publicly funded schools that they had to withdraw it following massive protests.

This article details all of the problems with the process that was developed, but ignores the underlying issue, that it was not a good faith effort.

The Tories were explicitly looking at finding a way of favoring the inbred elites of British society while depriving less affluent, and less-white students fair access to the the UK’s elite educational institutions:

When the UK first set out to find an alternative to school leaving qualifications, the premise seemed perfectly reasonable. Covid-19 had derailed any opportunity for students to take the exams in person, but the government still wanted a way to assess them for university admission decisions.

Chief among its concerns was an issue of fairness. Teachers had already made predictions of their students’ exam scores, but previous studies had shown that these could be biased on the basis of age, gender, and ethnicity. After a series of expert panels and consultations, Ofqual, the Office of Qualifications and Examinations Regulation, turned to an algorithm. From there, things went horribly wrong.

Nearly 40% of students ended up receiving exam scores downgraded from their teachers’ predictions, threatening to cost them their university spots. Analysis of the algorithm also revealed that it had disproportionately hurt students from working-class and disadvantaged communities and inflated the scores of students from private schools. On August 16, hundreds chanted “F%$# the algorithm” in front of the UK’s Department of Education building in London to protest the results. By the next day, Ofqual had reversed its decision. Students will now be awarded either their teacher’s predicted scores or the algorithm’s—whichever is higher.

(%$# mine)

The problem was not that education authorities failed, but that they over-succeeded.

If they had hit 10% or 20%, they would have gotten away with making higher education in the UK richer and whiter, and the students complaining would have been dismissed as ungrateful (insert race or class epithets here)s.

How This Works

No Comments

Someone is Gaming the System

At The Markup, a news org created to do deep dives on technical news story, has found that there are significant differences in the ways that Gmail handles emails from different presidential campaigns.

The Buttigieg andYang campaigns are achieving disproportionate success in getting into the Gmail primary inbox.

The implication of this story is that Google could alter its algorithms to favor one candidate over another.

I do not think that this is a credible concern, at least not yet.

However, it is entirely possible that there are people inside Google who favor one candidate over another who would provide detailed information to the campaigns about how to game the filters.

IMHO, the two campaigns most likely to have a Google insider feeding them information would be those of Buttigieg and Yang, and it is their emails that have achieved the most success in reaching the primary email tab.

It’s called a man on the inside attack:

Pete Buttigieg is leading at 63 percent. Andrew Yang came in second at 46 percent. And Elizabeth Warren looks like she’s in trouble with 0 percent.

These aren’t poll numbers for the U.S. 2020 Democratic presidential contest. Instead, they reflect which candidates were able to consistently land in Gmail’s primary inbox in a simple test.

The Markup set up a new Gmail account to find out how the company filters political email from candidates, think tanks, advocacy groups, and nonprofits.

We found that few of the emails we’d signed up to receive —11 percent—made it to the primary inbox, the first one a user sees when opening Gmail and the one the company says is “for the mail you really, really want.”

Half of all emails landed in a tab called “promotions,” which Gmail says is for “deals, offers, and other marketing emails.” Gmail sent another 40 percent to spam.

For political causes and candidates, who get a significant amount of their donations through email, having their messages diverted into less-visible tabs or spam can have profound effects.

“The fact that Gmail has so much control over our democracy and what happens and who raises money is frightening,” said Kenneth Pennington, a consultant who worked on Beto O’Rourke’s digital campaign.

………

It’s well known that Facebook and Twitter curate which posts people see through the news feed, highlighting some while others are scarcely shown. What’s received less attention is how email has also become an algorithmically curated and monetized platform—essentially another feed—and the effect that can have. Some nonprofits and political causes said inbox curation is reducing donations and petition signatures.

Google communications manager Katie Wattie said in an email that the categories “help users organize their email.”

………

Google communications manager Katie Wattie said in an email that the categories “help users organize their email.” 

………

The tabs also serve another purpose: ad inventory. While Gmail does not sell ads in the primary inbox, advertisers can pay for top placement in the social and promotions tabs in free accounts.