Hacker News Reader

Comment by karpierz | Hacker News Reader

This isn't a study that shows that people's faces indicate their political leanings.

It's a study which shows that pictures that people select to represent themselves publicly have features that indicate political leaning.

View on HN

Exactly. There was some fuss a while back about a similar classifier for sexuality. It turned out to be guessing mostly based on head tilt, personal hygiene and whether the person was wearing glasses. The physiognomy component was ~nonexistent even though it was publicized as though it weren’t. People intentionally if at times subconsciously present themselves in a way that signals information to kindred spirits. You’d need to bring in hundreds of people, wash them and basically take mugshots to control for that.

I think that conclusion is at least as interesting as physiognomy. It's remarkable that a computer could be more sensitive to it than people.

People don't get confirmation of those details normally unlike the ML algorithm. Of course we don't lock people in boxes with a stack of training set photos for a period of time equivalent to ML training.

Agree. Even the few pixels bordering the face in the sample image can show she's outside. She chose a smiling picture, she's wearing makeup, etc...

That's a really good observation to note. The prior embedded in their image data is their own bias of what is a "good" representation of themselves.

>images were tightly cropped around the face and resized to 224 × 224 pixels

One can only wonder what the result would be by using the equivalent of government ID photos (neutral expression, no smile or make-up, solid backgrounds).

Yep, I'd look very closely for training data bias.

Profile photo with a person wearing a baseball cap and Oakleys? Yup, that’s a republican.

If it's that easy then why was human guessing only 55 percent accurate?

At least partly due to lack of feedback on accuracy. I don’t know about you but I don’t necessarily ask everyone I meet their political leanings, so it’s hard to train yourself other than through stereotype.

Worth noting the human guessing was not on the same data set, but I believe the machines are going to beat us at this in general.

My guess would be sampling bias. Most people base their model on small geographically restricted samples, and are heavily biased by media. It is possible that some people can repeatedly perform better than 55%.

Reply on news.ycombinator.com