Live data from Hacker News

AI Model Detects Mental Disorders Based on Web Posts

home.dartmouth.edu

71–74 of 74 posts

Re: AI Model Detects Mental Disorders Based on Web Posts

#71

I'd say very little good can come from this. What would be an example of an "appropriate" use of this technology? Basically where it could excel is giving probabilistic prediction of "mental disorders" across a high volume of users. What is a good thing you would do with that? Advertising to suggest someone to get help, maybe?

I worked on a shared task (CLPsych [1]) that estimates mental distress severity. So the classification task was to determine how acutely a user message/post on a mental health forum would require moderator intervention. Some people announce their suicides live on the forum for instance. You want to point the moderators to those people asap. Our research group applied this tech to Belgium's suicide prevention center c…

Did you open source your solution?

Re: AI Model Detects Mental Disorders Based on Web Posts

#72
post #49

Earlier quoted context omitted.

if the concern is abuse, i wouldn’t give it to gov

The concern is directing public resources in ways that are most effective. I'm thinking about populations and not individuals.

i like your idea of improving effectiveness

population level data of social media is vulnerable to abuse by gov. social media is a poor input and gov will exploit ambiguity to their benefit.

Re: AI Model Detects Mental Disorders Based on Web Posts

#73
post #70
post #65

Earlier quoted context omitted.

If that was your original intent, you phrased it really badly. You didn't say "trust no one". People were asking what to do with it, and you said "i wouldn’t give it to gov". That isn't "this is dangerous and shouldn't exist".

> I'd say very little good can come from this. > if the concern is abuse not my intent, OP. the comment i replied to suggested gov instead. i made a comment about gov there was as much reason to assume what i meant as much as what you assumed

I see what you're saying now. Sorry to misinterpret you.

Re: AI Model Detects Mental Disorders Based on Web Posts

#74

Earlier quoted context omitted.

"chemical imbalances", is that a thing?

My understanding is that it has gone out of favor as a theory. It wasn’t ever developed into much of a robust model and is sorely lacking in many ways. But this is a layman understanding so I could be missing something.

I've been reading a little about epigenetics, the systems by which your body's genetic code responds to your environment.

For instance, people with scalp psoriasis can change the epigenetic structure of the skin on their scalp by using anti-psoriatic shampoos, so that the body itself, being freed of the irritance response of psoriasis stops acting like it is psoriatic.

Of course, as soon as the medication stops, the psoriasis returns, so it's not a permanent fix, nothing in epigenetics is. The fact that you can alter your genetic representation in real time is an amazing thing to me nevertheless.

I said all of that to say that there's a chance that any "chemical imbalances in the brain" might be the reaction to the underlying mental illness, and not the cause of or solution to the mental illness in the first place, and that treating the mental imbalance may give the person time away from the effects of their illness.

That, combined with therapy and validation, can conceivably result in "curing" the illness in people that had a chance to recover in the first place by giving them the boost they needed to get over the hump.

Not a one-size-fits-all solution, of course, we don't know enough about the brain to know what would be a permanent solution, but in your case, finding your own path and having uplifting experiences thanks to psychedelics might have been all that you needed to put that chapter of your life behind you.

This is all armchair conjecture but it's a thought I've been having so I figured I would share.

Post reply on HN