How A.I. Helped Improve Crowd Counting in Hong Kong Protests | Hacker News Reader

How A.I. Helped Improve Crowd Counting in Hong Kong Protests

nytimes.com

68 pointspseudolus8 comments

The article's demonstration of a counting model is horribly inaccurate to the point where I'm not sure why it was included. Most people see "AI" as being either good at something or not. There's little nuance such as some models being better than others. This kind of demonstration just weakens the reader's confidence in more fully developed results or different approaches. I'm not sure this brief statement offsets the prominent visuals:

> On the day of the protest, Mr. Yip and the A.I. team used technology that is much more advanced. They spent weeks training their program to improve its accuracy in analyzing crowd imagery.

Setting aside the presentation, from the photos the researchers appear to be using object detection rather than density estimation. This choice is problematic given the quantities involved and the need for temporal consistency.

I'm also skeptical of using human volunteers and surveys to calibrate the model. Humans are terrible at counting large numbers of people in real-time. That's a central point of the article with different groups of people providing wildly different counts.

You can count from a still picture or a video segment and use that to test or calibrate. So it may just be inaccurate reporting to give the impression that the calibration depends on humans surveying live action, which is exactly what they are working to replace. If the researchers publish their results the exact methodology used will tell but I assume they are competent.

Why do not need to count in the first place? whilst scale of protests important, who has counted the crowd. And given the authority ignored it why bother? The counting is silly the walk is not.

Anyway walk the three walk and given I do not pass the bridge, my legs hut until 3 days later. I have not been counted but I know I have walked.

Oh for *'s sake. Companies and institutions are desperate to justify their AI investment any way they can.

They would have gotten a perfectly good estimate by taking some photos of the crowd at different density areas over the day and doing some good old-fashioned sampling and extrapolation. Something that could be done by primary school students (and is done, remember counting cars or sampling insects on field days?).

Just because you can, doesn't mean you should.

Seems the NYT labelling for the bottom chart is broken/offset.

The problem is they cannot back test the final result.

Or back test the previous protests that has always been highly divergent results.

Even if you take this AI approach at face value, it is evident that the police's count is not that far off, compared to the organizer's estimates.

But hey, the po-po is always bad, no?

Po-po definitely not always bad, but there is damning evidence of their false-flag interference in this particular protest.

Reply on news.ycombinator.com