zachwooddoughty

Joined 13 years ago

Karma
103
Submissions
26
Member since
Sep 3
View on Hacker News

Recent submissions

Comment · 9 years ago
Of course all models are biased, but we can try to build models that minimize the types of bias we care most about.

I think the so-called panic comes from the fact that more complicated algorithms are difficult for an untrained person to understand. For example, it's also illegal to discriminate housing/lending based on race, but if I hide that discrimination inside of a sufficiently-complicated ML model, I may be able to get away with it.

Comment · 9 years ago
>> ProPublica labelled the algorithm as biased based primarily on the fact that it (correctly) labelled blacks as more likely than whites to re-offend (without using race as part of the predictor), and that blacks and whites have different false positive rates. >> In the conception of these authors, “bias” refers to an algorithm providing correct predictions that simply fail to reflect the reality the authors wish existed.

The gist of the article is that statistical bias is not the bias that journalists are interested in. The article doesn't discuss how these biases are related or relevant to each other, but rather assumes that statistical bias is the only one that should matter. I think the article is missing a discussion of the gaps between what are the measured inputs into your statistical model, and what can be acted upon from a policy perspective.

As a thought experiment, suppose the only two inputs that determine a person's recidivism rate is their past criminal history and whether they had lead poisoning as a child, but that of these two we can only measure past criminal history as an input into our algorithm. If race is strongly associated with childhood lead poisoning (such as in real life [1]), then our algorithm might get higher classification accuracy by including race [2] as an input in its training data. This might have less statistical bias, but would bias against individuals of a race who are in truth not at higher risk of recidivism.

[1] https://scholar.harvard.edu/files/alixwinter/files/sampson_w...

[2] The actual COMPAS algorithm doesn't use race as an explicit input, but that doesn't really change the issue.

Comment · 10 years ago
I think the argument is that if people tend to "just be rude" to women more often than to men, then that is evidence of sexism. Even if there isn't a conscious reason for the rudeness.

There clearly isn't any easy way to measure "aggregate rudeness" to be able to prove that women receive a higher quantity of "just being rude", but that doesn't necessarily mean it isn't there.

Comment · 10 years ago
While I empathize, anonymity makes moderation very hard. Most people I've talked to prefer the tradeoff of reduced anonymity for improved moderation. If you email me at [email protected] I can set you up with a pre-verified account that you can use anonymously.

Quality is hard. I don't think there's any silver bullet, but you can check out the community guidelines at discoverboard.com/about which hope to curtail the thoughtlessness you're talking about.

Comment · 10 years ago
You'll have to verify your phone number to sign up, which will make moderation more effective (hard to make an additional account after getting kicked out).

If it grows and that isn't enough to keep it respectful, I will add a $1-2 signup fee for new users.

I'm also going to be spending most of my time recruiting quality people to join the discussions.

Comment · 10 years ago
Plug: I'm launching Discoverboard.com for this in a couple weeks
Comment · 10 years ago
Comment · 11 years ago
Hi Sam -- do you think YC has a role or responsibility in addressing social issues in the bay area or the US?
Comment · 11 years ago
From the article:

>The offenses that landed them in solitary? Most often, it was evidence that they were "affiliated" with a prison gang, whether or not they had broken any rules.

Solitary isn't being used as a last resort for people who have killed previous cellmates -- it's being used for all sorts of things.

I think solitary makes sense for people who are already unable to coexist with other inmates, but that doesn't seem to be the main use at the moment.

Comment · 11 years ago
Comment · 11 years ago
Comment · 11 years ago
Short-term efficacy is not zero efficacy. If the fish works to alleviate anemia for three months, it seems worthwhile to continue distributing the fish and to rally support for further research.

Sure, this fish is probably getting too much hype for how effective it is, but isn't it better than nothing?

Comment · 11 years ago
>Some people are in power, they are white

This is the root of racial inequality in our justice system. No one is claiming that all white people are evil supervillians. But many white people don't comprehend the harsher reality people of color face when being arrested.

The comment "I love when white people are shocked" is a reference to the fact that "some people are in power, they are white". If the people in power are shocked by what it's like to be arrested, perhaps it will spark positive change in the system.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton