Live data from Hacker News

Mathematicians urge colleagues to boycott police work in wake of killings

nature.com

151–160 of 241 posts

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#152
post #96

Earlier quoted context omitted.

It's impolite to bring that up because OF COURSE there are more violent crimes committed by a group that has been systematically discriminated against for hundreds of years. What did you expect? So, what should we take away from that? Well, we should expect everyone to receive the same treatment from government organizations (and many non-government organizations) that deliver services to people, without bias. Just b…

> The complaint of "black lives matter" is that there is a pattern of police abuse towards non-violent black people. Again I have yet to see any convincing data that supports this. There's a marketed list of names of victims, some with video footage, but I can also show you names and video footage of non-violent victims of other races being killed by police. I certainly support police reform (within the bounds of com…

Hmm. Ok, there's a march on Washington being organized by Al Sharpton for families of people that have been killed while in police custody. Such a march is possible because there are so many. What do you think the ethnic makeup of that group will be?

What pretense are you saying is false, exactly? That there is a pattern of police abuse towards non-violent black people? It's patently silly. Black people have "the talk". White people don't have "the talk". We have videos of obvious injustice and racism popping up all the time, and the one-two punch of Ahmaud Arbery and George Floyd being killed by people who are obviously, deeply, racist and nothing being done about it (indicating a severely dysfunctional police organization) catalyzed the "moral panic", but there's no such epidemic of videos showing white people being abused. Why? Because there is no such epidemic, and it is racist.

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#153
post #38

Earlier quoted context omitted.

I guess I don't really understand what is generally wrong with a police officer looking for a crime being committed in public and then stopping it?

We create what we look for

Are doctors creating cancer and high cholesterol?

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#154
post #81

The last 40 years of austerity politics in the US have starved the beast [0] so greatly that the only remaining public infrastructure that functions as it's intended to is policing. For example, in the city of LA, 53% of the city's unrestricted general fund budget went to the LAPD. For everything else, there are parallel infrastructures for the haves - private education, private healthcare, privatized transit, etc. I…

> The last 40 years of austerity politics in the US

Wikipedia[0] says that "[austerity] is a set of political-economic policies that aim to reduce government budget deficits through spending cuts, tax increases, or a combination of both." That strikes me as a correct definition of 'austerity'. But we've not had any spending cuts in real terms or in terms of GDP. Nor have we had tax increases for the past "40 years" -- sure, Presidents Bill Clinton and Barack Obama raised taxes, but Presidents Reagan, Bush, and Trump cut taxes.

Yes, reducing budget deficits has been a stated goal of most Presidents since 1980, but that's mostly been lip service. None bothered to actually try. President Clinton did manage to reduce the deficit, but not because he wanted or tried to, but because of the unexpected revenue growth caused by the unexpected economic growth caused by the unexpected IT revolution.

> For example, in the city of LA, 53% of the city's unrestricted general fund budget went to the LAPD.

What was that in 1980? Was there a step in this, say, around 1996? Did it go up in 1996 when Federal funding for 100,000 new police officers dried up? I.e., did President Bill Clinton's well-meaning 100,000 police officer initiative have unintended consequences?

> On top of that, there is some research that suggests communities rate smaller police forces better than larger ones [1].

I don't doubt this! We saw in Ferguson, MO, that the police was giving out an inordinate number of traffic fines every year! It's almost... exactly as if... they were paying themselves with that -- a terrible corruption. As I recall the number of citizens in NYC who have fines in arrears is staggering. Police department funding via fines is a form of corruption, and strongly indicative of police departments being too large.

And the invention of civil asset forfeiture (in the 80s, IIRC), and the rise in its use (in the mid-90s, IIRC), also seems like police (and prosecutors) paying themselves with poor people's assets. And it is invariably the poor who have been hardest hit by civil asset forfeiture.

The police have to be sized appropriately to:

  - investigate and close most cases
  - support investigations appropriately
  - provide a minimum of neighborhood police
    presence to deter crime and build a rapport
    with the community
Also, police need to be sized appropriately for dealing with civil disturbances, though here calling on neighboring districts when needed can help reduce the size of police force needed.

The police departments of most large U.S. cities are almost certainly oversized at present. One of the first steps in remediating some of iniquities of our system should be to end police department funding via fines and civil asset forfeiture (by simply ending civil asset forfeiture -- as clearly unconstitutional a practice as any). In some States fine revenue is shared with the State, but perhaps the amounts of the fines should go down and the percentage that goes to the issuing jurisdiction should go down significantly. At the very least the parameters of traffic violation fining need fine-tuning. Once the police can only be funded by general revenue of the local town/city, and the towns/cities cannot fund themselves with fines / civil asset forfeitures, we'll see a proper reconsideration of police funding.

I.e., the police need to be not so much defunded as right-funded, and today that would generally mean spending less on police. I'm not sure that in 1993 it was the case that the police needed to be funded more, though I'd be happy to see evidence to the contrary. Crime rates did go down significantly in the 90s, but those extra ~80,000 new police officers are not the only cause of that, and probably not even a major cause at all. The main cause of the fall in crime rates in the 90s seems to have been demography.

[0] https://en.wikipedia.org/wiki/Austerity

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#155

Earlier quoted context omitted.

I don't think that's accurate at all. These algorithms are only as accurate as the data being input into them. The current system of policing is deeply flawed and as a result the "data collected from crime statistics" will be skewed by these inherent, systemic biases. Mathematical optimization against a flawed dataset of this kind is highly likely to create a feedback loop of increasing bias with terrible human conse…

This seems like a plausible criticism, but it also seems like a plausible criticism of all data and derivates about human beings (and probably other things).

Clearly you haven't seen crime or police data. It's truly, truly awful. That, and the people running analysis on it are just as bad:

https://www.chicagoreporter.com/chicago-police-arrested-more...

Garbage in crunched by garbage is still garbage out.

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#156

People are boiling this down to "less crime = good". Like if a mathematician can design a deployment system such that the percentage of crimes where an arrest is made increases from 15% to 20%, how can that be a bad thing? What if every single one of those additional arrests is a Black person? Well that might require an additional look. It's possible that it's all Black people doing those crimes. It's also possible t…

> It's also possible that the mathematician built an algorithm anchored on years of arrest data from racist cops doing racial profiling. That possibility is explicitly excluded since the data comes from crimes reported by victims, not investigations initiated by police officers.

What about racist victims?

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#157

People are boiling this down to "less crime = good". Like if a mathematician can design a deployment system such that the percentage of crimes where an arrest is made increases from 15% to 20%, how can that be a bad thing? What if every single one of those additional arrests is a Black person? Well that might require an additional look. It's possible that it's all Black people doing those crimes. It's also possible t…

Statisticians will be the best of all the disciplines at teasing out that sort of bias. If the police are going to be racially biased they don't need statisticians to help them. The addition of statisticians is only going to increase the number of people who are uncertain about the model. There are two broad paths here:

(1) Statisticians predict where crimes are likely to happen for the police using data.

(2) Police predict where the crimes are likely to happen using highly biased guesswork.

The statisticians are being irresponsible pushing the police towards (2). Option (1) can be de-biased, can be debated, its effects can be assessed and its parameters can be tweaked over time.

This is a choice between two options and the people signing on to this letter are arguing for the worse one. They link a bunch of news articles, but the academic complaint seems to be that the model says "assume crimes happen in high crime areas". That isn't a very scary model, and if anyone has an alternative they should be pushing it towards the police, not holding it back.

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#158

Earlier quoted context omitted.

There is no austerity. Here is state-and-local spending as a percentage of GDP from 1970 to present: https://s3-us-west-2.amazonaws.com/courses-images-archive-re... And here is GDP in 2010 dollars: https://data.worldbank.org/indicator/NY.GDP.PCAP.KD?location... State and local governments went from spending 10% of $23,000 per person to 15% of $55,000 per person. Per-person state and local spending went up by a factor…

Why would you compare 2014 ridership to 2020 capital spending? The same charts you posted show a 40% increase in ridership with only a 20% increase in capital spending from 1982-2004. Obviously the capital spending boost that is happening right will only be reflected in future ridership.

I’m comparing 2014 ridership to 2015 capital spending, but with the money expressed in 2020 dollars.

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#159
post #6

Earlier quoted context omitted.

I urge you to read Minority Report, or watch the movie.

> I urge you to read Minority Report, or watch the movie. This is fiction .

FAST, the Future Attribute Screening Technology, was a federal precrime research effort from the Department of Homeland Security which sought to pick criminals out at public locations based on biometrics. [0] FAST was tested in secret on the USA population in an unknown location. [1]

PredPol is a precrime system written by a mathematician (an assistant professor of math and CS) which recommends where police officers ought to patrol. [2] PredPol was deployed in Santa Cruz and Los Angeles, CA.

This is not fiction.

[0] https://web.archive.org/web/20100603031047/http://www.dhs.go...

[1] https://www.nature.com/news/2011/110527/full/news.2011.323.h...

[2] http://math.scu.edu/~gmohler/predpol.html

Re: Mathematicians urge colleagues to boycott police work in wake of killings

#160

Earlier quoted context omitted.

> I urge you to read Minority Report, or watch the movie. This is fiction .

FAST, the Future Attribute Screening Technology, was a federal precrime research effort from the Department of Homeland Security which sought to pick criminals out at public locations based on biometrics. [0] FAST was tested in secret on the USA population in an unknown location. [1] PredPol is a precrime system written by a mathematician (an assistant professor of math and CS) which recommends where police officers…

> This is not fiction.

No the book and movie are fiction, I mean. You can't learn anything from them. They're made up. Inside someone's head. There's zero data points in them.

I'm sure you can learn something from studying the history of the two projects you mentioned!

But not from a story someone made up and that some people acted out on a film set. That's not data. Don't base you philosophy on something that didn't actually happen.

Post reply on HN